Viticultural zoning in Europe
advertisement
Universidade de Trás-os-Montes e Alto Douro Viticultural zoning in Europe: Climate scenarios and adaptation measures Zonagem vitícola na Europa: cenários climáticos e medidas de adaptação Tese de Doutoramento em Ciências Agronómicas e Florestais Helder José Chaves Fraga João Carlos Andrade dos Santos Aureliano Natálio Coelho Malheiro José Manuel Moutinho Pereira Vila Real, Fevereiro 2014 Universidade de Trás-os-Montes e Alto Douro Viticultural zoning in Europe: Climate scenarios and adaptation measures Zonagem vitícola na Europa: cenários climáticos e medidas de adaptação Tese de Doutoramento em Ciências Agronómicas e Florestais Helder José Chaves Fraga João Carlos Andrade dos Santos Aureliano Natálio Coelho Malheiro José Manuel Moutinho Pereira Composição do Júri: Prof. Dr. José Paulo Mourão de Melo e Abreu (Instituto Superior de Agronomia) Prof. Dr. Joaquim Ginete Pinto (Universidade de Reading/Universidade de Colónia) Prof. Dr. João Carlos Andrade dos Santos (Universidade de Trás-os-Montes e Alto Douro) Prof. Dr. José Manuel Couto Silvestre (Instituto Nacional de Investigação Agrária e Veterinária, I.P.) Prof. Dr. Aureliano Natálio Coelho Malheiro (Universidade de Trás-os-Montes e Alto Douro) Prof.ª Dr.ª Maria Celeste Pereira Dias (Universidade de Aveiro) Vila Real, Fevereiro 2014 Este trabalho foi expressamente elaborado como tese original para a obtenção do grau de Doutor em Ciências Agronómicas e Florestais, de acordo com o disposto no Decreto-Lei n.º 216/92, de 13 de Outubro alterado pelo Decreto-Lei n.º 115/2013 de 7 de Agosto. III IV Ao meu pai José Fraga, À minha irmã Carmen Fraga, À minha mãe Leonida Fraga (1940-2000). V VI Agradecimentos À Universidade de Trás-os-Montes e Alto Douro. Aos meus orientadores, Prof. Doutor Aureliano Natálio Coelho Malheiro, Prof. Doutor José Moutinho-Pereira e Prof. Doutor João Carlos Andrade dos Santos, por todo o apoio prestado, pelos valiosos conhecimentos transmitidos e pela importante orientação prestada. Ao Prof. Doutor Joaquim Werner Pinto da Universidade de Colónia/Universidade de Reading (Alemanha/Inglaterra) por ter gentilmente cedido dados essenciais para elaboração desta dissertação. Ao Doutor José Eiras-Dias e ao Doutor José Silvestre do Instituto Nacional de Investigação Agrária e Veterinária por terem gentilmente cedido dados para a elaboração desta tese. Ao Engenheiro Fernando Alves da Associação para o Desenvolvimento da Viticultura Duriense pelo valioso conhecimento fornecido ligado ao sector vitivinícola nacional. Aos Professores José Tadeu Aranha, Manuel Teles Oliveira, José Martinho Lourenço e Ana Alexandra Oliveira pelos conhecimentos transmitidos em unidades curriculares, essenciais para a elaboração desta tese. Ao Professor Doutor António Nazaré Pereira, coordenador do curso de doutoramento; Um agradecimento à FCT pelo apoio financeiro do projeto ClimVineSafe “Short-term climate change mitigation strategies for Mediterranean vineyards” (PTDC/AGR-ALI/110877/2009), no qual esta tese se enquadra. Ao CITAB (FCOMP-01-0124-FEDER-022692) por ter fornecido as condições para a realização desta tese. Aos projetos/grupos (http://www.worldclim.org), ENSEMBLES ECA&D (GOCE-CT-2003-505539), (http://eca.knmi.nl), MODIS WorldClim (http:// http://modis.gsfc.nasa.gov) e COSMO-CLM Consortium, cujos dados fornecidos foram usados para a elaboração desta tese. A todos que de alguma forma contribuíram para concretização desta tese de doutoramento. VII VIII Summary IX Summary X Summary Viticulture and winemaking are strongly governed by weather and climate and so are highly vulnerable to climate change. For the most renowned viticultural regions in Europe, establishing the role of climate on grapevine development, quality attributes and wine production, is of key importance for the economic stability and growth of the sector. This is of particular interest in some region presenting Mediterranean-like climatic conditions, such as Portugal, given the existing climatic variability and its projected changes in the future. In the present study, the existing bioclimatic zoning over Europe is assessed through the computation of specialized viticultural indices, using a state-of-the art observational dataset. By using an ensemble of climate models, future projection and associated uncertainties are assessed on a regional scale. The resulting bioclimatic indices allow the isolation of the most suitable zones for grapevine growth in Europe, and enable to infer on the regional potential for wine quality, risk of contamination by pests and diseases and water demands. It is clear that many countries in southern Europe currently present a high thermal accumulation and moderate to high levels of dryness, with potential implications in grapevine maturity and wine quality. Future projections, show increased thermal accumulation throughout Europe. In southern Europe, projections point to future drying and warming, resulting in additional threats to the winemaking sector. Conversely, in northern and central Europe, warm and moist climates may result in higher risks for pests and diseases. Interannual variability is also expected to increase in the future, resulting in a higher fluctuation in both yield and quality attributes. Although, the obtained projections are quite robust, they present a certain level of uncertainty, which is higher in precipitation-based indices than in temperature-based indices, especially in the Mediterranean-like climatic regions. The study continued with a more detailed analysis of the implications of climate change on Portuguese viticulture. The study focused on bioclimatic zoning assessment under present and future scenarios for the Portuguese mainland viticultural regions. Furthermore, due to the typical Mediterranean-like climate of the Portuguese winemaking regions, extreme bioclimatic metrics were also analysed. These metrics have a high importance for the Portuguese viticulture, already affected by high water and heat stress. Results show that significant increases in thermal accumulation and dryness levels are expected to occur in the future, particularly over the southern and innermost regions. Additionally, an intensification of extreme weather events could result in an increased inter-annual variability in yield and quality attributes. Thus, a reshaping of the Portuguese winemaking regions is likely to occur XI Summary in the next decades, emphasizing the need for the development of appropriate climate change adaptation measures, in order to preserve wine styles. Due to the large array of conditions in which grapevines are grown in Portugal (e.g. elevation, orientation), it became necessary to enhance the spatial resolution of our results. Therefore, a regional assessment at a very-high spatial resolution (approximately 1 km) of the bioclimatic parameters is performed. For this purpose, a statistical downscaling method is applied to a set of 13 regional climate models on the 12 winemaking regions over mainland Portugal. A categorized bioclimatic index is then developed taking into account grapevine thermal demands, water availability and ripening conditions, in each region. This index also allows direct comparison between each winemaking region in Portugal, representing a reference tool for viticultural zoning. Results depict the current spatial distribution of the Portuguese vineyards and the Atlantic/Mediterranean climatic contrast over Portugal. Nonetheless under future scenarios, increased dryness and thermal accumulation may result in loss of viticultural suitability, lower bioclimatic diversity and earlier phenological events. These projections may lead to changes in varietal selection and wine characteristics of each region. These outcomes are particularly relevant for the winemaking sector, as this very-high resolution highlights the small scale changes in the spatial variability of each winemaking region. The study progressed with the analysis of the influence of climate on wine production. These relationships are analysed using a relatively long wine production series in the Minho wine region, in Portugal. Through the analysis of this series it is possible to isolate the main atmospheric components that influence wine production. Several climatic elements have shown have great impact on yield and wine production. Overall, a moderate water stress during the growing season, high production 3-yrs before, cool weather in February-March, settled-warm weather in May, warm moist weather in June and relatively cool conditions preceding harvest are generally favourable to a high wine production. The linkage between large scale atmospheric circulation/patterns and wine production was also demonstrated. Additionally, a modelling approach is undertaken taking into account these factors, providing a high skill, which could prove to be of great value for the winemaking sector. In the next stage, the grapevine phenological development is studied, and its relationship with atmospheric parameters is established. For this purpose, a phenological time series of four grapevine varieties in the Lisbon winemaking region is used. Subsequently, through multivariate linear regressions, it is possible to determine which climatic factors have XII Summary contributed the most for these development stages. Results show that air temperatures preceding grapevine phenological stages have great influence on the onset and duration of these stages. The developed statistical models show that this climatic forcing can explain most of the variability in phenological development, confirming the existence of significant links between climate and grapevine development stages. This study also allows a better understanding of which of these varieties will be more adapted to future climates. Despite the key role played by climate, other factors such as soil and topography also acquire an important role for grapevine development and growth. Therefore, the viticultural zoning study is extended by analysing not only the climatic conditions but also the characteristics of soil, topography and vegetative growth, of the winemaking regions in the Iberian Peninsula. This study allowed us to integrate all these factors towards a comparison between various regions, while assessing the relative importance of each factor. Results show climate plays a leading role for grapevine vegetative growth, when compared to other grapevine-influencing elements such as soil type and topography. The established climate-grapevine relationships and climate change projections can prove to be of major importance to the regional viticultural sector. These projections may enable the implementation of strategies to cope with the inter-annual variability in both grapevine phenology and wine production. Furthermore, the awareness to the climate change impacts in viticulture may also allow reducing the sector vulnerability. These results are particularly important for the winemaking sector in Europe and especially in Portugal. XIII XIV Sumário XV Sumário XVI Sumário A vitivinicultura é fortemente influenciada pelos elementos climáticos, sendo altamente vulnerável às alterações climáticas. Nas mais famosas regiões vitivinícolas da Europa, estabelecer o papel do clima no desenvolvimento da videira, atributos qualitativos e produção de vinho, adquire uma importância fundamental para a sustentabilidade económica e social do setor. Este fator é de particular interesse em algumas regiões com climas tipicamente Mediterrânicos, como é o caso Portugal, dada a variabilidade climática existente e as alterações projetadas para o futuro. No presente estudo, a atual zonagem bioclimática na Europa é analisada através do cálculo de índices vitivinícolas especializados, utilizando uma base de dados observacional. Através de um conjunto de modelos climáticos, as projeções futuras e incertezas associadas são posteriormente avaliadas à escala regional. Os índices bioclimáticos permitem sinalizar as zonas com melhor aptidão para o cultivo da vinha na Europa, e também inferir sobre a qualidade do vinho, o risco de contaminação por doenças e pragas, bem como a avaliação das necessidades hídricas da videira em cada região. De notar que muitas regiões do sul da Europa apresentam atualmente uma elevada acumulação térmica e moderados níveis de secura, com implicações na maturação e qualidade do vinho. As projeções futuras mostram um aumento da acumulação térmica um pouco por toda a Europa. No sul da Europa, as projeções apontam para um futuro muito mais quente e seco, resultando em riscos adicionais para a cultura da vinha. Por outro lado, na Europa central e do norte, climas mais quentes e húmidos podem aumentar a vulnerabilidade das videiras a determinadas doenças como por exemplo o míldio. A variabilidade inter-anual também deverá aumentar no futuro, o que resultará numa maior oscilação na produtividade e nos atributos qualitativos do vinho. Embora, as projeções obtidas sejam bastante robustas, elas apresentam um certo grau de incerteza, o qual é mais elevado em índices baseados em precipitação do que em temperatura, em especial nas regiões que apresentam climas tipicamente Mediterrânicos. O estudo prosseguiu com uma análise mais detalhada sobre as implicações das alterações climáticas na viticultura em Portugal. O estudo focou-se na avaliação de zonagem bioclimática em cenários atuais e futuros. Além disso, devido ao clima tipicamente Mediterrânico existente em Portugal, foram desenvolvidas métricas de extremos bioclimáticos. Estas métricas têm uma grande importância para a viticultura em Portugal, já de si bastante afetada pelo elevado stresse hídrico e térmico. Os resultados mostram que são esperados aumentos significativos na acumulação térmica e níveis de secura, especialmente sobre as regiões do interior sul do país. Além disso, a intensificação de eventos climáticos XVII Sumário extremos pode resultar num aumento da variabilidade inter-anual desta cultura. Assim, é expectável que nas próximas décadas ocorra uma reformulação das regiões vitivinícolas em Portugal, enfatizando a necessidade para o desenvolvimento de medidas adequadas de adaptação às alterações climáticas, a fim de preservar a tipicidade enológica em cada região vitícola. Devido à grande variabilidade de condições em que as videiras são cultivadas em Portugal (e.g. elevação, orientação), tornou-se necessário melhorar a resolução espacial dos resultados obtidos. Para tal, foi efetuada uma zonagem bioclimática à escala regional com uma resolução espacial muito elevada (cerca de 1 km). Nesse sentido, foi desenvolvido um método de downscaling estatístico (redução de escala), aplicado a um conjunto de 13 modelos climáticos regionais nas 12 regiões vitivinícolas em Portugal Continental. De seguida foi desenvolvido um índice bioclimático categorizado tendo em conta as exigências térmicas da videira, a disponibilidade de água e as condições de maturação, em cada região. Este índice permite também a comparação direta entre cada região vitivinícola em Portugal, representando uma ferramenta de referência para a zonagem vitícola. Os resultados permitem capturar a atual distribuição espacial da cultura da vinha e mostram o contraste climático existente em todo o território. Em cenários futuros, o aumento da secura e acumulação térmica pode resultar em perda de aptidão vitícola, diminuição da diversidade bioclimática e antecipação de eventos fenológicos. Estas projeções poderão levar a mudanças na seleção de castas e porta-enxertos e, por conseguinte, nas características do vinho de cada região. Estes resultados são particularmente relevantes para o setor vitivinícola, uma vez que a elevada resolução dos resultados permite identificar alterações na viticultura Portuguesa à escala local. O estudo prosseguiu com a análise da influência do clima na produção de vinho. Esta relação é analisada utilizando uma série de produção de vinho, relativamente longa, da região vitivinícola do Minho, em Portugal. Através da análise desta série, foi possível isolar os principais fatores climáticos com influência para a produção de vinho. Vários elementos climáticos mostraram ter um grande impacto na produtividade e na produção de vinho. No geral, stresse hídrico moderado durante a época de crescimento, elevada produção há 3 anos atrás, tempo fresco em Fevereiro-Março, tempo estável em Maio, tempo quente e húmido em Junho e temperaturas baixas durante a colheita são geralmente favoráveis a uma elevada produção. A ligação entre a circulação atmosférica de larga escala e a produção de vinho também foi demonstrada. De seguida, foi desenvolvido um modelo estatístico, tendo em conta XVIII Sumário estes fatores, oferecendo uma elevada fiabilidade, podendo vir a ser de grande valor para o setor vitivinícola em Portugal. Na fase seguinte, o desenvolvimento fenológico da videira foi analisado, e a sua relação com elementos climáticos devidamente estabelecida. Para tal, foi usada uma série temporal da fenologia de quatro castas de videira, da região vitivinícola de Lisboa. Posteriormente, através de regressões lineares multivariadas, foi possível determinar quais os fatores climáticos que mais contribuíram para a evolução dos estados fenológicos. Os resultados mostraram que a temperatura do ar nos meses precedentes à ocorrência dos estados fenológicos tem uma grande influência sobre o início e a duração destes. Os modelos estatísticos desenvolvidos mostraram ainda que o forçamento climático pode explicar a maior parte da variabilidade no desenvolvimento fenológico da videira, confirmando a existência de ligações significativas entre o clima e os diferentes estados. Este estudo também permitiu compreender quais destas castas estarão melhor adaptadas ao clima futuro. Apesar do papel fundamental desempenhado pelo clima, outros fatores como o solo e a topografia também possuem um papel importante para o desenvolvimento da videira. Neste sentido, o estudo da zonagem vitícola é aperfeiçoado incluindo não só as condições climáticas, mas também as características de solo, topografia e crescimento vegetativo, nas regiões vitivinícolas da Península Ibérica. Este estudo permitiu integrar todos estes fatores para uma comparação entre as várias regiões, e também avaliar a importância relativa de cada fator. Os resultados mostram que o clima desempenha um papel fundamental para o crescimento vegetativo da videira, quando comparado a outros elementos, como tipo de solo e topografia. As relações clima-viticultura e as projeções das alterações climáticas aqui estabelecidas revelam ser de grande importância. Estas projeções permitem a implementação de estratégias para lidar com a variabilidade inter-anual, tanto da fenologia da videira como da produção de vinho. Além disso, a perceção dos impactos das alterações climáticas na viticultura pode também permitir a redução da vulnerabilidade do setor. Estes resultados são particularmente pertinentes para o setor vitivinícola na Europa e especialmente em Portugal. XIX XX Index Agradecimentos ............................................................................................................................................. VII Summary .................................................................................................................................................IX Sumário ................................................................................................................................................XV Introduction ...........................................................................................................................................XXXIII Chapter 1. An overview of climate change impacts on European viticulture ................................................ 1 1.1. Viticulture worldwide........................................................................................................................ 3 1.2. Vine physiology and climate influences ............................................................................................ 6 1.3. Climate change projections in agriculture ........................................................................................ 9 1.4. Climate change impacts on viticulture ........................................................................................... 12 1.5. Adaptation and mitigation measures ............................................................................................. 14 1.5.1. Varietal and rootstock decisions .................................................................................................... 15 1.5.2. Irrigation management ................................................................................................................. 16 1.5.3. Tillage treatments ......................................................................................................................... 17 1.5.4. Vineyard microclimate .................................................................................................................. 18 1.6. Conclusions ..................................................................................................................................... 18 Chapter 2. Future scenarios for viticultural zoning in Europe: ensemble projections and uncertainties .... 21 2.1. Introduction .................................................................................................................................... 23 2.2. 2.2.1. 2.2.2. 2.2.3. 2.2.4. Materials and Methods .................................................................................................................. 25 Bioclimatic indices ......................................................................................................................... 25 Model data .................................................................................................................................... 27 Model output statistics and model uncertainties ......................................................................... 29 Inter-annual variability .................................................................................................................. 31 2.3.1. 2.3.2. 2.3.3. 2.3.4. 2.3.5. 2.3.6. Results ............................................................................................................................................ 31 Recent-past viticultural zoning ...................................................................................................... 31 Composite index vs. viticultural regions ....................................................................................... 32 Future viticultural zoning .............................................................................................................. 33 Climate change signals and uncertainties ..................................................................................... 34 Categorization of CompI................................................................................................................ 38 Inter-annual variability .................................................................................................................. 40 2.3. 2.4. Summary and Discussion ................................................................................................................ 41 Chapter 3. Climate change impacts on the Portuguese viticulture from a multi-model ensemble ............. 45 3.1. Introduction .................................................................................................................................... 47 3.2. Methods.......................................................................................................................................... 49 3.3. Results ............................................................................................................................................ 52 3.4. Summary and Conclusions .............................................................................................................. 56 Chapter 4. Very high resolution bioclimatic zoning of Portuguese wine regions: present and future scenarios .................................................................................................................................. 59 4.1. Introduction .................................................................................................................................... 61 XXI 4.2. 4.2.1. 4.2.2. 4.2.3. Materials and methods................................................................................................................... 63 Bioclimatic zoning ......................................................................................................................... 63 Spatial pattern downscaling .......................................................................................................... 64 Categorized index .......................................................................................................................... 65 4.3.1. 4.3.2. 4.3.3. Results ............................................................................................................................................ 66 Regional climatic zoning ................................................................................................................ 66 Projected category transitions ...................................................................................................... 72 Ensemble variability ...................................................................................................................... 72 4.3. 4.4. Discussion and conclusions ............................................................................................................. 73 Chapter 5. Climate factors driving wine production in the Portuguese Minho region ................................ 77 5.1. Introduction .................................................................................................................................... 79 5.2. 5.2.1. 5.2.2. 5.2.3. 5.2.4. 5.2.5. 5.2.6. Materials and methods................................................................................................................... 82 Wine production data ................................................................................................................... 82 Climatic data.................................................................................................................................. 82 Production cycles .......................................................................................................................... 82 Weather regimes ........................................................................................................................... 83 Bioclimatic indices ......................................................................................................................... 84 Logistic model ............................................................................................................................... 85 5.3.1. 5.3.2. 5.3.3. 5.3.4. 5.3.5. Results and discussion .................................................................................................................... 85 Wine production series ................................................................................................................. 85 Candidate regressors ..................................................................................................................... 88 Weather regimes ........................................................................................................................... 90 Significant regressors .................................................................................................................... 92 Logistic model ............................................................................................................................... 95 5.3. 5.4. Discussion and conclusions ............................................................................................................. 97 Chapter 6. Winegrape phenology and temperature relationships in the Lisbon Wine Region, Portugal ... 101 6.1. Introduction .................................................................................................................................. 103 6.2. 6.2.1. 6.2.2. 6.2.3. Material and methods .................................................................................................................. 105 Location and climate data ........................................................................................................... 105 Plant material and phenology ..................................................................................................... 107 Statistical analysis ....................................................................................................................... 108 6.3.1. 6.3.2. 6.3.3. Results and discussion .................................................................................................................. 109 Climatic conditions ...................................................................................................................... 109 Characteristics and relationships in phenology .......................................................................... 110 Temperature and phenology: effects and trends ....................................................................... 114 6.3. 6.4. Conclusions ................................................................................................................................... 119 Chapter 7. Integrated analysis of climate, soil, topography and vegetative growth in Iberian viticultural areas ....................................................................................................................................... 121 7.1. Introduction .................................................................................................................................. 123 7.2. 7.2.1. 7.2.2. 7.2.3. 7.2.4. 7.2.5. 7.3. Material and methods .................................................................................................................. 125 Viticultural regions and vineyard area ........................................................................................ 125 Topography ................................................................................................................................. 126 Climate ........................................................................................................................................ 128 Soils ............................................................................................................................................. 129 Vegetative growth ....................................................................................................................... 130 Results .......................................................................................................................................... 131 XXII 7.3.1. 7.3.2. Mesoscale patterns ..................................................................................................................... 131 DO analysis .................................................................................................................................. 136 7.4. Discussion and conclusions ........................................................................................................... 140 Chapter 8. Synthesis and final conclusions ............................................................................................... 143 8.1. Introduction .................................................................................................................................. 145 8.2. Summary of the research findings ................................................................................................ 146 8.3. Adaptation measures: considerations at variety level.................................................................. 149 8.4. Final remarks and future work ..................................................................................................... 152 References ……………………………………………………………………………………………………………………………………....…...155 Supplementary material 1 ............................................................................................................................ 177 Supplementary material 2 ............................................................................................................................ 181 Supplementary material 3 ............................................................................................................................ 192 XXIII XXIV List of figures Fig. 1.1 - a) Total vine surface area and percentage of this area in Europe, Asia, USA + southern hemisphere regions, and others, from 2000 to 2011. b) Total grape production, wine production and wine consumption for 2000 to 2011. Adapted from (OIV 2012) .................................................................. 3 Fig. 1.2 - World Distribution of the viticultural regions (darker areas). .................................................................. 5 Fig. 1.3 - Vegetative and reproductive cycles and vine phenological stages. Adapted from Eichorn and Lorenz (1977) and Magalhães (2008). ............................................................................................................. 6 Fig. 1.4 - Maturity groupings based on the growing season mean temperature, for a set of grapevine varieties. Adapted from Jones (2006).................................................................................................................. 7 Fig. 1.5 - a) Schematic view of the components of the climate system, their processes and interactions. b) Multimodel averages and assessed ranges for surface warming. Solid lines are multi-model global averages of surface warming (relative to 1980–1999) for the scenarios A2, A1B and B1, shown as continuations of the 20th century simulations. Shading denotes the ±1 standard deviation range of individual model annual averages. The orange line is for the experiment where concentrations were held constant at year 2000 values. The grey bars at right indicate the best estimate (solid line within each bar) and the likely range assessed for the six SRES marker scenarios. The assessment of the best estimate and likely ranges in the grey bars includes the AOGCMs in the left part of the figure, as well as results from a hierarchy of independent models and observational constraints. (a) Adapted from Le Treut et al. (2007), their FAQ 1.2, (b) Adapted from (Alley et al. 2007), their Fig. SMP.5 ................................................................................................................................................. 10 Fig. 2.1 - (a) Huglin Index, (b) Dryness Index, (c) Hydrothermal Index and (d) Composite Index for the baseline period (1961-2000). (e) European wine regions (circles) with the corresponding Composite Index value (cf. scale) for the period of 1980-2009 using the E-OBS data. Source of the wine regions locations: “Wine Regions of the World Version 1.31” (URL: http://geocommons.com/overlays/3547). ........................................................................................ 32 Fig. 2.2 - As in Fig. 2.1 but for the future period of 2041-2070 under the A1B IPCC-SRES scenario. .................... 34 Fig. 2.3 - (a) Composite index of the more severe model (HC RCA3 (C4I)). (b) Composite index showing the less severe model (HC RCA (SMHI)), for the time period of 2041-2070. .................................................. 35 Fig. 2.4 - Left Panel: Differences in the mean patterns (2041-2070 minus 1961-2000) of the (a) Huglin Index, (b) Dryness Index, (c) Hydrothermal Index and (d) Composite Index. Right Panel: Normalized Interquartile Range (third quartile minus first quartile divided by the mean at each grid point) showing variability in the 16-member ensemble for the same Indices as in left panel. Differences not statistically significant at the 99% confidence level are grey shaded (NS). ................................. 37 Fig. 2.5 - Leading categories in the Composite Index (HI classes 1, 2 and 3) for the baseline period (a) and the two future periods (b, c). ................................................................................................................... 38 Fig. 2.6 - Latitudinal differences (2041-2070 minus 1961-2000) in the number of grid cells equal to or above 0.50 in the Composite Index. ............................................................................................................. 39 Fig. 2.7 - Ratio between the averages of the 16 inter-annual standard-deviations (calculated for each of the 16 ensemble members separately) in 2041-2070 and 1961-2000 of the (a) Huglin Index, (c) Dryness Index. Normalized Interquartile Range showing variability in the 16-member ensemble for the ratios of the (b) Huglin Index, (d) Dryness Index. No statistically significant ratios are plotted in grey shading (NS). ...................................................................................................................................... 41 Fig. 3.1 - The wine regions of mainland Portugal (IGP- Protected geographical indication). ............................... 48 Fig. 3.2 - Huglin Index (HI; in ºC) for a) the baseline period (1961-2000), b) the future time period (2041-2070), and c) overall differences in 2041-2070 minus 1961-2000 (under the A1B IPCC-SRES scenario). Black areas represent the current vineyard land cover. ............................................................................. 52 XXV Fig. 3.3 - Cool night Index (CI; in ºC) for a) the baseline period (1961-2000), b) the future time period (20412070), and c) overall differences in 2041-2070 minus 1961-2000 (under the A1B IPCC-SRES scenario). Black areas represent the current vineyard land cover. ................................................... 53 Fig. 3.4 - a) Differences in the number of days with maximum daily temperature equal to or above 40ºC (20412070 minus 1961-2000). b) Differences in the precipitation totals (in mm) during the growing season (2041-2070 minus 1961-2000)............................................................................................... 53 Fig. 3.5 - Dryness Index (DI; in mm) for a) the baseline period (1961-2000), b) the future time period (20412070), and c) overall differences in 2041-2070 minus 1961-2000 (under the A1B IPCC-SRES scenario). Black areas represent the current vineyard land cover. ................................................... 54 Fig. 3.6 - Hydrothermal Index (HyI; in ºC.mm) for a) the baseline period (1961-2000), b) the future time period (2041-2070), and c) overall differences in 2041-2070 minus 1961-2000 (under the A1B IPCC-SRES scenario). Black areas represent the current vineyard land cover. ................................................... 55 Fig. 3.7 - Composite Index (CompI) for a) the baseline period (1961-2000), b) the future time period (20412070), and c) overall differences in 2041-2070 minus 1961-2000 (under the A1B IPCC-SRES scenario). Black areas represent the current vineyard land cover. Not statistically significant differences at the 99% confidence level are grey shaded. ................................................................ 55 Fig. 3.8 - Latitudinal differences between the periods 2041-2070 and 1961-2000 in the number of grid cells equal to or above 0.5 in the composite index (CompI)...................................................................... 56 Fig. 4.1 - (a) Vineyard land cover over mainland Portugal using Corine Land Cover (EEA 2002). (b) Winemaking regions in Portugal (IVV 2011). (c) Elevation in meters using the GTOPO30 digital elevation model (http://eros.usgs.gov/#/Find_Data/Products_and_Data_Available/gtopo30_info), the red line differentiates the Atlantic/Mediterranean climatic zones. ............................................................... 63 Fig. 4.2 - CatI for the winemaking regions of (a, b) Minho, (c, d) Trás-os-Montes and (e, f) Douro/Porto for (left panel) 1950-2000 and (right panel) 2041-2070 under the A1B scenario. The current vineyard land cover is shown in dark stipple. See Table 4.1 for class structure for the CatI.................................... 67 Fig. 4.3 - CatI for the winemaking regions of (a, b) Beira-Atlântico, (c, d) Terras-do-Dão and (e, f) Terras-deCister for (left panel) 1950-2000 and (right panel) 2041-2070 under the A1B scenario. The current vineyard land cover is shown in dark stipple. See Table 4.1 for class structure for the CatI. ............ 68 Fig. 4.4 - CatI for the winemaking regions of (a, b) Terras-da-Beira, (c, d) Lisboa and (e, f) Tejo for (left panel) 1950-2000 and (right panel) 2041-2070 under the A1B scenario. The current vineyard land cover is shown in dark stipple. See Table 4.1 for class structure for the CatI. ................................................ 70 Fig. 4.5 - CatI for the winemaking regions of (a, b) Península-de-Setúbal, (c, d) Alentejo and (e, f) Algarve for (left panel) 1950-2000 and (right panel) 2041-2070 under the A1B scenario. The current vineyard land cover is shown in dark stipple. See Table 4.1 for class structure for the CatI. ........................... 71 Fig. 5.1 - Right panel: map showing the geographical location of the Minho Wine Region (MWR) in northwestern Portugal, along with the other Portuguese wine regions. Left panel: magnified map of the MWR. Grid boxes represent the E-OBS original resolution of 0.25º latitude × 0.25º longitude. Highlighted grid boxes show where data was extracted for area-mean computations. The main current vineyard land cover is also shown in black stipple. ............................................................... 81 3 Fig. 5.2 - (a) On the left: chronogram of the wine production in the MWR (in 10 hl) in 1950-2010 (grey line), along with the corresponding best-fit second order polynomial and the 11-yr moving average. On the right: corresponding box-plot, where the median is the horizontal line within the box, the 75th and 25th percentiles are the upper and lower box limits, and the whiskers limits are the maximum and minimum. (b, c) The same as in (a), but for the transformed wine production series (b) and for the percentiles of the transformed series (c). The horizontal dashed lines correspond to the 25th and 75th percentiles, the thresholds used in the class definitions for logistic regression. ............... 87 Fig. 5.3 - Normalized power spectral density (dimensionless) of the percentile-transformed wine production series in the MWR for 1950-2010. ..................................................................................................... 88 XXVI Fig. 5.4 - (a) Chronogram of the HyI from 1950 to 2010, using a 9 grid-box area-mean over the MWR and the respective linear trend. (b) The same as in (a), but for the DI. .......................................................... 89 Fig. 5.5 - (a-f) Composites of the daily mean sea level pressure fields (in hPa) over the Euro-Atlantic sector (2565ºN; 60ºW-20ºE) for the six outlined weather regimes (AA, E, NW, C, R and A) and for all days in 1950-2010. ......................................................................................................................................... 91 Fig. 5.6 - Relative monthly frequencies of the occurrence (in %) of each weather regime for the period 19502010 and across the year. .................................................................................................................. 92 Fig. 5.7 - Relative monthly frequencies of the occurrence (in %) of each weather regime for the period 19502010 and across the year. .................................................................................................................. 95 Fig. 5.8 - (a) Chornogram of the observed (modeled) classes represented by black (white) circles. (b) The number of matches (light grey bars), the number of errors (dark grey bars) and the percentage of errors in each class are pointed out................................................................................................... 96 Fig. 6.1 – a) Average annual (left panel) and April-September (right panel) precipitation (in mm) for the 19502000 period, Lisbon Wine Region, Portugal. b): the same as a) but for the air temperature (in ºC). ......................................................................................................................................................... 105 Fig. 6.2 - Yearly evolution (1990-2011) of maximum (TX), mean (TG) and minimum (TN) temperatures (reconstructed series) for the (a) winter (Jan-Feb-Mar), (b) spring (Apr-May-Jun), (c) summer (JulAug-Sep) and (d) autumn (Oct-Nov-Dec) in Dois Portos, Portugal. Straight lines represent linear 2 regressions for significant trends, LT their trends, and R their coefficients of determination. pvalues < 0.05 are shown................................................................................................................... 107 Fig. 6.3 - Annual values of maximum (TX), mean (TG) and minimum (TN) temperatures (reconstructed series) for the 1990-2011 period in Dois Portos, Portugal. Enclosed symbols of mean temperatures correspond to values that fell below first (squares) and above third (circles) quartiles and straight 2 lines represent linear regressions for significant trends, LT their trends, and R their coefficients of determination. p-values < 0.05 are shown. ..................................................................................... 110 Fig. 6.4 - On the left: Box plots of budburst, flowering, véraison and harvest dates of Aragonez, Castelão, Chasselas and Fernão Pires varieties from 1990 to 2011 in Dois Portos, Portugal. Medians correspond to horizontal black lines within boxes, lower (upper) box limits to the first (third) quartiles and whiskers to the non-outlier maxima and minima. First (second) order outliers are indicated by circles (asterisks) and represent events above/below the box upper/lower limit by at least 1.5 (3.0) times the respective box height (interquartile range). On the right: Time series of the same phenophases and varieties for the outlined period. The dashed lines represent the linear 2 regression for the budburst of Aragonez and harvest of Fernão Pires, LT represents its trend, and R represents the coefficient of determination. .................................................................................. 111 Fig. 7.1 - - a) Location of the viticultural regions in Iberia, along with their denomination. b) Spatial distribution of the vineyard land cover over Iberia (dark-red), assessed using the Corine Land Cover, version 132012, along with the viticultural regions (light-grey)....................................................................... 127 Fig. 7.2 - a) Elevation (m) in the Iberian Peninsula, calculated using the GTOPO30 dataset. b) As in (a) but for the aspect. c) Solar radiation over Iberia, mean growing season values in 1989-2012 calculated using MERRA data at a 0.6º spatial resolution. ......................................................................................... 132 Fig. 7.3 - a) CatI over Iberia calculated according to Table 1, for the period of 1989-2012 using WRF simulations. b) SoilT according to Table 2 using HWSD data. c) Mean EVI and EVIc for the grapevine growth period (April-October) in 2012, using MODIS data. The spatial-average of the vineyard areas corresponds to 0.23. Below this value the EVIc equals 1 (transparent overlay), above this value EVIc equals 2 (hatched overlay). .............................................................................................................. 133 Fig. 7.4 - a) Circular accumulated EVIc (1 – red, 2 – green) as a function of the SoilT and CatI for all vineyards in Iberia. b) Circular accumulated EVIc as a function of the elevation and aspect for all vineyards in Iberia. The size of each circular chart depicts the accumulated vineyard area belonging to that EVI class, and the inner (outer) circular class depicts the largest (smallest) EVI class. .......................... 135 XXVII Fig. 7.5 - Elevation (m) of the vineyards in each DO/DOCa in Iberia. The inner circle represents the mean elevation and the horizontal bars represent the minimum and maximum, of the locations of the vineyards inside the DO. .................................................................................................................. 137 Fig. 7.6 - a) Geographical aspect (orientation) vineyards in Iberia according to the mean elevation of each DO. b) as in (a) but normalized using the vineyard area......................................................................... 137 Fig. 8.1 – Schematic representation of the main topics that contributed to the thesis. .................................... 145 Fig. 8.2 - Growing degree day (ºC) in which some grapevine varieties are grown in Portugal, calculated using the WorldClim dataset using the varietal spatial distribution. Medians correspond to horizontal black lines within boxes, lower (upper) box limits to the minimum and maximum of the GDD value. .... 151 XXVIII List of tables Table 1.1 - Top 16 wine producing countries in 2011(with respective growth rate from 2007). Total vineyard area in 2011 is also shown (with respective growth rate from 2007). Table is sorted by a descending order based on wine production. (OIV 2012) ...................................................................................... 4 Table 2.1 - List of the bioclimatic indices used for climatic viticultural zoning in Europe, along with their corresponding definitions and references. The calculations (summations) are carried out using daily climate variables. ............................................................................................................................... 26 Table 2.2 - Summary table of all GCM / RCM model chains used in this study. The corresponding acronyms, institutions and original grid resolutions are also listed. In all simulations the period 2011-2070 was used under the IPCC-SRES A1B scenario. Relevant references to each chain are also indicated. ..... 28 Table 2.3 Ensemble means, medians, minima, maxima, normalized interquartile ranges (NIQR) and total ranges (TR) of the Composite Index in 2041-2070 and for a selection of European famous winemaking regions (the respective grid box coordinates are also listed). Regions are ranked in ascending order with respect to their TR values. ......................................................................................................... 36 Table 3.1 - List of all the bioclimatic indices used in this study, their definitions and references. ....................... 50 Table 3.2 - Summary of all GCM / RCM model chains, original grid resolutions, institutions and references used in this study. ....................................................................................................................................... 51 Table 4.1. Categories of CatI, along with the corresponding classes of the combined indices: Huglin, Dryness and Cool Night indices. A short description of the each category is also provided. ................................. 65 Table 4.2 - Wine regions with the corresponding dominant category (with the highest spatial coverage) in the ensemble-mean CatI for the future period (2041-2070; cf. Figs. 4.2-5). The respective single-model dominant categories are also listed, along with the percentages of occurrence across the 13-model ensemble (ensemble-mean dominant category in bold). .................................................................. 73 2 Table 6.1 - Means, standard deviations and linear trend parameters (R , p-value and annual trend) of main phenophase dates and corresponding intervals for the four varieties in Dois Portos, Portugal (19902011). ............................................................................................................................................... 112 Table 6.2 - Pearson correlations coefficients between main phenophases of the four varieties in Dois Portos, Portugal (1990-2011). Values in bold are significant (p < 0.05). ...................................................... 114 Table 6.3 - Summary of most significant and not inter-correlated (99% confidence level) monthly (or multimonthly) temperature variables (TX, TN and TG combined and analysed separately) and 2 corresponding coefficients of determination after cross-validation (Rcv ) for key phenophases of the four varieties in Dois Portos, Portugal (1990-2011). TX, TN and TG: maximum, minimum and mean temperatures, respectively. nd: not determined. Note that all temperatures are negatively correlated with phenophase dates. ................................................................................................. 115 Table 7.1 - Categorized Index (CatI), along with the corresponding classes of the combined indices: Huglin, Dryness and Cool Night indices, according to Fraga et al. (2014a) .................................................. 128 Table 7.2 - Soil texture categories, along with the respective percentages of Clay, Silt and Sand, according to USDA soil textural classification (USDA 2006). ................................................................................ 130 Table 7.3 - CatI, SoilT and EVI class for each viticultural region in Iberia. Only the predominant categories are shown............................................................................................................................................... 139 XXIX XXX List of abbreviations asl – above sea level; BI – Branas Heliothermic Index; CatI – Categorized Index; CCS – Climate Change Signal; CI – Cool Night Index; CompI – Composite Index; DI – Dryness Index; DJF – December-January-February; ECA&D – European Climate Assessment & Dataset; EVI - Enhanced Vegetation Index; EVIc - Enhanced Vegetation Index Class; GCM – Global Climate Models; GDD – Growing Degree Day; GSP – Growing Season Precipitation; GSS – Growing Season Suitability; GST – Growing Season Temperature; HI – Huglin Index; HyI – Hydrothermal Index; IBP – Index Baseline Pattern; IFP – Index Future Pattern; IPCC – Intergovernmental Panel on Climate Change; JJA – June-July-August; LWR – Lisbon Wine Region; MODIS - Moderate Resolution Imaging Spectroradiometer; MOS – Model Output Statistics; MWR – Minho Wine Region; NIQR – Normalized interquartile range; OIV – International Organization of Vine and Wine; P – Precipitation; RCM – Regional Climate Models; SI – Selianinov Index; SoilT – Soil Textural Class SRES – Special Report on Emission Scenarios; Tavr - Mean Temperature; Tmax - Maximum Temperature; Tmin – Minimum Temperature; TR – Total range; WP – Wine Production XXXI XXXII Introduction Viticulture and winemaking represent a sector with a high socio-economic importance worldwide, particularly for the so called “old world” viticulture, Europe. Grapevines are a climate sensitive crop, since its development and growth are strongly influenced by the prevailing atmospheric conditions in each region. As such, the optimum climatic conditions for grapevine growth are limited geographically, and very specific conditions are needed to obtain consistent yields and balanced fruit composition. Given this climatic forcing, grapevines are highly vulnerable to climate change. The amounting evidence for significant climate change in the upcoming decades is expected to bring new challenges to the European viticultural sector. Moreover, in most Mediterranean-like climatic regions, such as Portugal, vineyards are already subjected to high levels of heat and water stress, and should face even more challenges due in the future. Therefore, given the importance of the winemaking sector in Europe, and especially in Portugal, the study of climate change impacts in viticulture is of utmost importance for the future sustainability of this crop. In the past, some studies approached the subject of climate change impacts on viticulture (e.g. Duchene and Schneider 2005; Jones et al. 2005a; Moriondo and Bindi 2007; Malheiro et al. 2010; Webb et al. 2007). Nonetheless, the current thesis provides several innovative approaches. By using a multi-model ensemble for future scenarios over Europe, and by quantifying the inherent uncertainties, it is possible to obtain robust climate change projections. Additionally, the very-high resolution climatic zoning and future scenarios of the Portuguese viticulture regions attained in this study, was never before achieved. The integration of several bioclimatic indices into a reference categorized index allowed integrating the most important bioclimatic characteristics of each winemaking region, which is a step forward in this field of research. Furthermore, the thesis provides evidence on the linkage between grapevine yield/quality attributes and climate, soil, topography and other grapevine influencing elements. These relationships are extremely valuable not only for the current scientific knowledge but also for the winemaking sector as a whole. Therefore, the current thesis attempts to fill a gap in current state of knowledge regarding the climate change impacts in European viticulture, and more specifically in Portugal. XXXIII The aim and objectives of the study The current thesis focuses on the viticultural zoning and climate change impacts on European viticulture, as well as establishing links between climatic factors and grapevine development and growth, and wine production and quality attributes. Therefore, the current thesis aims to i) to study the impact of climate change in European viticulture and, in particular, in Portugal; and ii) to develop viticultural zoning for present and future scenarios, in order to draw general guidelines for suitable adaptation and mitigation measures. The specific objectives were to: 1. provide a complete review of the current state of knowledge regarding climate change effects on European viticulture; 2. model and compute a set of the bioclimatic indices using observational data (recent-past conditions) in Europe. Validate/calibrate regional climate models in simulating the observed indices and develop future projections based on these parameters. Analyse the corresponding model uncertainties. Assess climate change impacts in viticultural zoning and in the inter-annual variability of this crop; 3. optimize and adjust the bioclimatic indices for the Portuguese viticultural regions. Analyse future scenarios for viticultural zoning with emphasis on the Portuguese winemaking regions. Develop bioclimatic extreme metrics that can be adapted to Portuguese climate. Evaluate climate change impacts on these metrics. Analyse possible suitable adaptation measures to future climates; 4. develop very-high resolution spatial pattern downscaling methodologies for viticultural parameters in Portugal. Select the most suitable bioclimatic indices for viticultural zoning and establish a reference index for assessing viticultural suitability of a given region. Provide a regional assessment of viticultural parameter and analyse climate change impacts, focusing on possible adaptation/mitigation measures; 5. establish the forcing of climate in wine production, by selecting the main climatic factors that influence the inter-annual variability of grapevine yield. Quantify of cycles in wine production and analyse trends. Identify the large-scale atmospheric patterns/regimes associated with the variability wine production. Develop a high skill statistical model for wine production near term prediction; XXXIV 6. evaluate the influence of temperature variability on grapevine phenology. Establish relationships between the onset and duration of the main phenological stages and their trends; 7. enhance the viticultural zoning methodologies by assessing the prevailing conditions in terms of climate, soil and topography in the Iberian viticultural regions. Develop an integrated analysis of the previous three elements and their impact on vegetative growth; 8. assess the implications of the results in the definition of mitigation/adaptation measures and sustainability in the viticultural sector. Structure of the thesis This thesis is organized into eight chapters, six of which (chapters 2-7), include full original scientific research published or submitted to refereed international journals indexed in the Thomson Reuters® Journal Citation Reports®. In Chapter 1, the current state-of-the-art scientific knowledge regarding the impacts of climate change on European viticulture is described. This chapter, also published in an international scientific journal as a review article, presents an overview of the most relevant research concerning the climatic conditions and its relationships with viticulture in Europe and the potential climate change impacts. Chapter 2 main objectives are to establish a viable bioclimatic zoning in Europe and to develop future scenarios for viticulture. In this section, an historical/observational dataset and an ensemble of regional climate models are used to calculate specialized bioclimatic indices. Additionally, by using a 16 member ensemble of climate models, the uncertainties associated with climate modelling were also assessed on a regional scale, enhancing the robustness of the results. The resulting bioclimatic indices allow assessing the impacts of climate change on European viticulture. Chapter 3 provides a more a more detailed study of the implications of climate change on Portuguese viticulture. This section constituted a gap in the current scientific knowledge. Due to the typical Mediterranean-like climate of the Portuguese winemaking regions, the study focused on future projections of extreme bioclimatic metrics. These metrics have a high importance for the Portuguese viticulture, very affected by high water and heat stress during the grapevine growing season. Additionally, potential adaptation/mitigation measures to future climates are also discussed. XXXV The assessments provided in the latter chapter are extremely valuable for the understanding of climate change impacts in Portuguese viticulture on a mesoscale level. Nevertheless, it became necessary to enhance the spatial resolution of our results, as vineyards depend on many grapevine-influencing elements of the landscape (e.g. elevation, orientation). Therefore, in Chapter 4, a regional assessment at a very-high spatial resolution (approximately 1 km) of bioclimatic parameters is undertaken. For this purpose, a statistical downscaling method is applied to a set of 13 regional climate models on the 12 winemaking regions in Portugal. This very high viticultural zoning was never before achieved over the Portuguese mainland. Subsequently, a categorized bioclimatic index is developed taking into account grapevine thermal demands, water availability and ripening conditions, in each region. This index also allows direct comparison between each winemaking region in Portugal, representing a reference tool for viticultural categorization. In Chapter 5, wine production and its relationships with bioclimatic factors are analysed. Establishing these relationships is extremely valuable for the winemaking sector. These relationships are analysed using a relatively long wine production regions demarcated series in the Minho wine regions in Portugal. Through the analysis of this series it is possible to determine the main atmospheric components that influence wine production. Additionally, a modelling approach is undertaken taking into account these factors, providing a high skill. These assessments are also very useful taking into account the climate change projections, established in previous chapters. In Chapter 6, the phenological development of grapevines and its relationship with atmospheric parameters are established. For this purpose, a phenological time series of four grapevine varieties in the Lisbon winemaking region is used. Subsequently, through multivariate linear regressions, it is possible to determine which climatic factors have contributed the most for these development stages. This study also allows a better understanding of which of these varieties will be more adapted to future climates. Through the previous chapters, the importance of climate is established for a suitable viticultural zoning. Despite the key role played by climate, soil and topography also acquire an important role for grapevine development and growth. Therefore, in Chapter 7, the viticultural zoning study is extended by analysing not only the climatic conditions but also the characteristics of soil, topography and vegetative growth, of the winemaking regions in the Iberian Peninsula. This study allowed us to integrate all these factors towards a comparison between various regions, while assessing the relative importance of each factor. XXXVI Finally, Chapter 8 provides a synthesis of the study, together with the main conclusions of the thesis and a work plan for future research. XXXVII XXXVIII Chapter 1. An overview of climate change impacts on European viticulture Food and Energy Security, 2012, 1(2): 94-11 Impact Factor (5-yr): n/a Helder Fraga, Aureliano C. Malheiro, José Moutinho-Pereira, João A. Santos *This manuscript was submitted by invitation of the Journal Editor: Prof. Ricardo Azevedo and published on the first issue of the journal. 1 Chapter 1. An overview of climate change impacts on European viticulture Abstract The importance of viticulture and of the winemaking socio-economic sector in Europe is largely acknowledged. The most famous winemaking regions in Europe commonly present very specific environmental characteristics, where climate often plays a central role. Furthermore, given the strong influence of the atmospheric factors on this crop, climate change can significantly affect yield and wine quality under future conditions. Recent-past trends recorded on many viticultural regions in Europe hint at an already pronounced increase in the growing season mean temperatures. Furthermore, climate change projections give evidence for significant changes in both the growing season temperatures and precipitations in the next decades. Although grapevines have several survival strategies, the amounting evidence for significant climate change in the upcoming decades urges adaptation and mitigation measures to be taken by the whole winemaking sector. Short term adaptation measures can be considered as a first protection strategy and should be focused at specific threats, mostly changes in crop management practices (e.g. irrigation, sunscreens for leaf protection). At long term, however, a wide range of adaptation measures should be considered (e.g. varietal and land allocation changes). An overview of the current scientific knowledge, mostly concerning the European viticulture, the potential climate change impacts and feasible adaptation measures is provided herein. 2 Chapter 1. An overview of climate change impacts on European viticulture 1.1. Viticulture worldwide Both viticulture and winemaking are important practices that represent a key economic activity in many regions worldwide. In the International Organization of Vine and Wine (OIV) most recent report (OIV 2012), it is estimated that the world vineyards reached an area of 7.59 Mha in 2011, despite the downward trend in the last years (Fig. 1.1a). This trend is clearly associated with the decreasing of the vineyard area over Europe. Although Europe has lost some of its dominancy to Asia, USA and some southern hemisphere areas (Argentina, Australia, Chile, South Africa), Europe still encompasses the largest vineyard area in the world (38%; Fig. 1.1a). Based on the same report, global wine production stood at 265 Mhl in 2011 (Fig. 1.1b), while wine consumption reached 244 Mhl. In spite of the downward trend in vine surface area, grape production underwent an upward trend over the last years (Fig. 1.1b). The world’s top wine producing countries (Table 1.1) are Fig. 1.1 - a) Total vine surface area and percentage of this France, Italy and Spain, while it is worth area in Europe, Asia, USA + southern hemisphere regions, noticing that China, Chile and New and others, from 2000 to 2011. b) Total grape production, wine production and wine consumption for 2000 to 2011. Adapted from (OIV 2012) Zealand recorded the largest increases in production over the last years. 3 Chapter 1. An overview of climate change impacts on European viticulture Table 1.1 - Top 16 wine producing countries in 2011(with respective growth rate from 2007). Total vineyard area in 2011 is also shown (with respective growth rate from 2007). Table is sorted by a descending order based on wine production. (OIV 2012) Country Wine production (Mhl) / Growth rate Vine area (mha) / Growth rate France 49.6 9% 807 -7% Italy 41.6 -10% 776 -6% Spain 34.3 -1% 1032 -4% USA 18.7 -6% 405 2% Argentina 15.5 3% 218 -4% China 13.2 6% 560 4% Australia 11.0 14% 174 0% Chile 10.6 29% 202 3% South Africa 9.3 -1% 131 -12% Portugal 5.9 -2% 240 -3% Romania 4.7 -11% 204 0% Brazil 3.5 -1% 92 0% Greece 2.6 -26% 111 -6% Hungary 2.4 -24% 65 -13% New Zealand 2.4 59% 37 21% Bulgaria 1.3 -29% 73 -22% Figure 1.2 depicts the current viticultural regions of the world, including some world renowned winemaking regions, such as Bordeaux, Burgundy, California, Cape/South Africa, Champagne, La Mancha, La Rioja, Mendoza, Mosel, Porto/Douro, South Australia, Tuscany, among others. The current viticultural regions are usually within areas with Recognized Appellation of Origin (RAO; Resolution OIV/ECO 2/92) or Designation of Origin (DO; European Community 479/2008 Art. 34 1a), which ensures the typicity of wines. In Europe, more specifically, these famous winemaking regions are traditionally distinguishable by their prevailing environmental characteristics, such as climate, soils and grown varieties. While most of the regions in southern Europe have typical Mediterranean climates, with long warm and dry summers, central and northern Europe are often characterized by more continental and/or humid climates, with mild and rainy summers (Kottek et al. 2006; Peel et al. 2007). According to Carbonneau (2003), the microclimatic and mesoclimatic characteristics of a given viticultural region are key factors in understanding its varietal suitability and wine types. These climatic aspects have been valorized and properly taken into account in all 4 Chapter 1. An overview of climate change impacts on European viticulture ancestral wine regions, where vine growers have gradually adapted to best suite their regional/local environmental conditions (Jones 2012). Soil and cultural practices are also important controlling factors in the development of viticulture. It is widely known that certain winegrape varieties produce the best results in soils with specific characteristics. Soil chemistry and structure may indeed influence wine grape composition (Mackenzie and Christy 2005). Moreover, management practices such as crop Fig. 1.2 - World Distribution of the viticultural regions (darker areas). load, girdling, pinching, pruning, rootstock and scion, thinning and topping, may also influence winegrape growth and quality (Winkler 1974). Regarding the enological practices, Unwin (1996) stated that they have remained nearly unchanged until recent decades, when some technological breakthroughs have been carried out. As such, climate, soil and management practices form an highly complex and interactive system called Terroir (Magalhães 2008). According to the OIV (Resolution OIV/VITI 333/2010), “Terroir is a concept which refers to an area in which collective knowledge of the interactions between the identifiable physical and biological environment and applied vitivinicultural practices develops, providing distinctive characteristics for the products originating from this area. Terroir includes specific soil, topography, climate, landscape characteristics and biodiversity features”. This system significantly affects vine development and berry composition and has been accepted as a central aspect in determining wine quality and typicity (van Leeuwen et al. 2004). 5 Chapter 1. An overview of climate change impacts on European viticulture Nowadays viticulture faces new challenges and threats, some of the most important being related to climate change. Therefore, a discussion on the interconnections between vine physiology and climate is presented in section 2. The climate change projections and their impacts on viticulture are presented in sections 3 and 4. Section 5 is devoted to the adaptation and mitigation strategies. Lastly, section 6 outlines the main conclusions. 1.2. Vine physiology and climate influences The vine undergoes morphological and physiological changes resulting from different stages of its vegetative and reproductive cycles (Fig. 1.3). The duration of each phenological stage differs according to each grapevine variety, which is generally tied to the thermal conditions of each region (Mandelli et al. 2005). The prediction of stage evolution is of utmost importance in planning viticultural activities and winemaking decisions (Lopes et al. 2008). Jones (2006) showed that the length of the growing season, for each variety, is directly related to the growing season mean temperature (Fig. 1.4). Additionally, Webb et al. (2012) concluded that the length of the growing season could also be linked to soil moisture, air temperature and crop management practices. In fact, climate strongly influences the development of this crop, by requiring suitable temperatures, radiation intensities/duration and water availability during its growth cycle, which ultimately influence yield and wine quality (Magalhães 2008; Makra et al. 2009). Fig. 1.3 - Vegetative and reproductive cycles and vine phenological stages. Adapted from Eichorn and Lorenz (1977) and Magalhães (2008). 6 Chapter 1. An overview of climate change impacts on European viticulture Air temperature is considered the most important factor in the overall growth and productivity of winegrapes (Jones and Alves 2012a). In effect, grapevine physiology metabolism/composition and are fruit highly influenced by the mean temperature along the growing season (Coombe 1987). Even though this crop has a good adaptation to environmental stresses, enduring extremely low temperatures in short time periods during winter (Hidalgo 2002), negative temperatures during spring may severely damage the developing buds and leaves/shoots (Branas 1974). This crop is also very Fig. 1.4 - Maturity groupings based on the growing season mean temperature, for a set of grapevine varieties. Adapted sensitive to late frost and hail events (e.g. from Jones (2006). Spellman 1999). However, winter chill is an important aspect in its growth development, since cold promotes bud dormancy (Kliewer and Soleiman 1972), initiating carbohydrate reserves for the following year (Bates et al. 2002; Field et al. 2009). In the same way, a 10ºC basal temperature is required for the vine to break this dormancy and initiate its growing cycle (Amerine and Winkler 1944; Winkler 1974). Extreme heat or heat weaves may also permanently affect vine physiology and yield attributes (Kliewer 1977; Mullins et al. 1992), though some varieties may be more tolerant than others (Schaffer and Andersen 1994; Moutinho-Pereira et al. 2007). Grapevines growing under severe heat stress experience a significant decline in productivity, due to stomatal and mesophyll limitations in photosynthesis (Moutinho-Pereira et al. 2004), as well as injures under other physiological processes (Berry and Bjorkman 1980). Some studies argue that cool night temperatures in the period preceding the harvest (maturation/ripening), combined with high diurnal temperatures, stimulate the synthesis of anthocyanins and other phenolic compounds, being thus beneficial for high quality wines (Kliewer and Torres 1972; Mori et al. 2005). 7 Chapter 1. An overview of climate change impacts on European viticulture Annual precipitation and its seasonality are also critical factors influencing viticulture, as water stress can lead to a wide range of effects, yet largely dependent on the stage of development (Austin and Bondari 1988). For instance, proper soil moisture during budburst and shoot/inflorescence development is of foremost importance for vine growth development (Paranychianakis et al. 2004; Hardie and Martin 2000). Water stress at this stage may also cause small shoot growth, poor flower-cluster and berry set development (Hardie and Considine 1976). On the contrary, excessive humidity during these early stages overstimulates the vegetative growth, which leads to denser canopies and to more likely disease problems in leaves and in inflorescences. From flowering to berry ripening, severe water stress results in low leaf area, limiting photosynthesis, flower abortion and cluster abscission (During 1986). During this development stage, moderately dry and stable atmospheric conditions are considered favorable for high quality wines (Jones and Davis 2000b; Nemani et al. 2001; Ramos et al. 2008). Further, slower leaf canopy development may lead to higher transpiration efficiency (Porter and Semenov 2005). During ripening, excessive humidity is unfavorable to maturation (Tonietto 1999), due to the promotion of sugar dilution (Reynolds and Naylor 1994). On the contrary, moderate dryness, at this stage, seems to enhance quality (Storchi et al. 2005). Solar radiation is also a key factor affecting viticulture. Adequate radiant energy is required, especially during ripening (Manica and Pommer 2006). During maturation, sugar and phenolic contents are favored by the occurrence of sunny days (Riou et al. 1994). Regions with less sunlight tend to surmount this limitation by adjusting training systems, optimizing solar exposure and canopy density. With more exposed leaves and grape clusters, stomatal conductance and photosynthesis are favored, but increasing water demands (Archer and Strauss 1990) and boosting other problems, namely sunburns in leaves and clusters. In opposition, less exposed grape clusters result in lower berry temperatures, but at the expense of reducing sugar and anthocyanin concentrations (Smart et al. 1985; Sparks and Larsen 1966). High canopy density can also reduce bud fertility (Morgan et al. 1985), which is important for the next years. To better characterize the relationship between winegrapes and climate, several bioclimatic indices have been developed. Using the growing degree sums for a basal temperature of 10°C (degree-days), Amerine and Winkler (1944) developed one of the earliest indices (Winkler Index). Still using this concept, the Huglin index (Huglin 1978) was developed to also account for maximum temperatures during the growing season and include 8 Chapter 1. An overview of climate change impacts on European viticulture a radiation component. The cool night index (Tonietto 1999; Tonietto and Carbonneau 2004) , which accounts for minimum temperatures preceding the harvest, is also often used as a bioclimatic index. The dryness index (Tonietto and Carbonneau 2004; Riou et al. 1994) was developed to estimate potential soil water availability. More recently, Malheiro et al. (2010) developed a composite index based on the limiting thresholds of 3 bioclimatic indices, allowing a more comprehensive viticultural zoning and the assessment of the suitability of a given region to grapevine growth and wine production. In most Mediterranean-like climatic regions, vineyards are subject to high radiation levels interacting with high temperatures and strong atmospheric and soil water deficits, which largely constrain grapevine productivity. Frequently, leaves display permanent photoinhibition and chlorosis, followed by necrosis, exposing the grape clusters, thus leading to low intrinsic water use efficiency (WUE; Moutinho-Pereira et al. 2004). Hence, low vigor tends to be associated with reduced berry weight, sugar content and yield. Other berry organoleptic properties, such as color, flavor and aroma components are inhibited by excessive solar radiance and severe dryness too. This results in unbalanced wines, with high alcoholic content and excessively low acidity (Jones 2004). In this context, the Mediterranean viticulture and winemaking may be significantly challenged by climate change (Jones 2006). 1.3. Climate change projections in agriculture Climate change is an inevitable challenge that society will have to cope with in the upcoming decades. During the 20th century, most of Europe endured changes in numerous climatic factors with great regional heterogeneity (IPCC 2007). Significant changes in temperatures were found during the 20th century (Santos and Leite 2009), including increases of 2.3-5.3ºC in northern Europe and 2.2-5.1ºC in southern Europe (Christensen et al. 2007). Moreover, decreases/increases in the annual precipitations over southern/northern Europe (Christensen et al. 2007) are also expected to under future higher anthropogenic greenhouse gas (GHG) forcing. These changes were shown to occur not only in the normal values but also in the rate of occurrence of extremes (Hanson et al. 2007). In fact, changes in the frequency of temperature and precipitation extremes, in Europe, were related to certain atmospheric features, such as the North Atlantic Oscillation (NAO; Santos and Corte-Real 2006). 9 Chapter 1. An overview of climate change impacts on European viticulture The first climate theories considered climate as the mean state of the weather conditions over a long period (Hann 1883). The modern perspective of climatic analysis may be perceived as a system-analytic approach (Peixoto and Oort 1992; von Storch and Flöser 1999), where climate dynamics is a result of the combined interactions among the different components of the global climate system (atmosphere, hydrosphere, biosphere, cryosphere and lithosphere). Due to the complexity and high non-linearity of the climate system (Fig. 1.5a), the isolation of the forcing factors underlying climate variability is indeed a rather difficult task. In fact, the differentiation of the human-induced from the natural climate change is only possible when the physical processes in the climate system are fully understood. These processes can be described by balance equations and their numerical integration allows climate simulation. The study of climate change impacts on environmental systems usually employs data from atmospheric models. Due to different aspects of climate variability in paleontological and present day conditions, different approaches of climate modeling are carried out, induced by varying Fig. 1.5 - a) Schematic view of the components of the CO2 climate system, their processes and interactions. b) concentrations. The International Panel on Multi-model averages and assessed ranges for surface Climate Change (IPCC) developed different climate projections, referring to different future scenarios of divergent CO2 emission pathways (Fig. 1.5b), such as the A2, A1B and B1 (Nakićenović et al. 2000). These scenarios aim at objectively representing likely pathways of human development until the end of the 21st century, covering a feasible level of uncertainty (Nakićenović et warming. Solid lines are multi-model global averages of surface warming (relative to 1980–1999) for the scenarios A2, A1B and B1, shown as continuations of the 20th century simulations. Shading denotes the ±1 standard deviation range of individual model annual averages. The orange line is for the experiment where concentrations were held constant at year 2000 values. The grey bars at right indicate the best estimate (solid line within each bar) and the likely range assessed for the six SRES marker scenarios. The assessment of the best estimate and likely ranges in the grey bars includes the AOGCMs in the left part of the figure, as well as results from a hierarchy of independent models and observational constraints. (a) Adapted from Le Treut et al. (2007), their FAQ 1.2, (b) Adapted from (Alley et al. 2007), their Fig. SMP.5 10 Chapter 1. An overview of climate change impacts on European viticulture al. 2000). The IPCC 4th Assessment Report (4AR) provides evidence on the general agreement among observations from different measurement techniques; all documenting a clear upward trend in the global mean surface temperature in recent decades (Trenberth and Josey 2007). Besides the significant changes detected in the means, the frequencies of occurrence and strength of some extremes have also increased, including precipitation (droughts and heavy rain events) and temperature (heat waves, hot days/nights) extremes (Trenberth and Josey 2007; Andrade et al. 2012). Further, the recent-past observed trends can only be reproduced by climate modeling when anthropogenic forcing is taken into account (Hegerl et al. 2007). For future decades, ensemble projections reveal that global climate change tends to be mostly consistent with the trends already recorded during the twentieth century, though their magnitude is highly dependent on the emission scenario (Meehl et al. 2007). The global mean surface temperature is expected to increase at about 0.2ºC per decade, reaching values between 1-6ºC at the end of the 21st century for the full range of SRES scenarios (Fig. 1.5b). These values imply remarkably different impacts on the Earth’s system, as the environmental and socio-economic systems present non-linear responses to a certain change in temperature, besides their limited ability to adapt to new external conditions. Moreover, the regional impacts may be stronger/weaker than the global mean signal, which highlights the need for regional assessment studies using regional dynamical downscaling (Christensen et al. 2007). Lastly, it is worth mentioning that all these projections still encompass high uncertainties (Fraga et al. 2013), not only due to model limitations and to emission scenario uncertainties, but also due to uncertainties inherent to the climate-carbon cycle (Denman et al. 2007; Meehl et al. 2007). A new generation of emission scenarios (Moss et al. 2010) will enable the analysis of the interactions within this cycle in more detail. The projected changes in atmospheric parameters are of key importance for agricultural practices. Regarding perennial crops (such as vineyards), under future climate change, Lobell et al. (2006) states that actual yield changes will reflect the combined influence of climatic factors and the potentially positive effects of management, technology, and increased atmospheric CO2 contents. For estimating impacts of climate on agricultural systems it is needed, in addition to the first order impact of changed climate parameters on physiological and yield indicators, knowledge about the second order effect of CO2 influence on primary plant production. Furthermore, understanding the interactions of elevated CO2 concentrations with changes in climatic parameters, including extremes, soil water availability, pests, 11 Chapter 1. An overview of climate change impacts on European viticulture diseases and physiological interactions remain a key aspect for assessing climate change impacts on agriculture (Tubiello and Fischer 2007). Although a global assessment of climate change impacts on agriculture highlight negative impacts due to more frequent extreme weather events (Porter and Semenov 2005), increased irrigation demands (Doll 2002) and increased risk of pests and diseases (Alig et al. 2002), this should not occur evenly throughout the European continent. Olesen and Bindi (2002) stated two different climate change consequences for European agriculture. For northern Europe, positive effects can be expected from the introduction of new crop species/varieties, higher crop production and new suitable agricultural areas, but also experiencing disadvantages through the greater need for plant protection and increased risk of nutrient leaching. In southern Europe, on the other hand, water scarcity and extreme climate conditions may cause lower yields, higher yield variability and a reduction in suitable areas for traditional crops. Moreover, Moriondo and Bindi (2007) found that the Mediterranean crops may have earlier development stages and may experience a reduction in the length of growing season under future climates. As such, the development of future climate projections based on feasible future socioeconomic storylines is of great value, as they provide objective information that can be used in developing suitable adaptation/mitigation managements to minimize climate change impacts on environment and on human activities. 1.4. Climate change impacts on viticulture Climate change can potentially influence vine yield and quality (Kenny and Harrison 1992; Jones 2005). Recent-past temperature trends, focusing on viticultural regions, show that the growing season mean temperatures have increased globally about 1.3ºC in 1950-1999 and 1.7ºC from 1950 to 2004 in Europe (Jones et al. 2005b; Jones et al. 2005a). For some European viticultural regions, in Italy, Germany and France, studies already reported shortenings of the growing season and earlier phenological events (Bock et al. 2011; Daux et al. 2011; Chuine et al. 2004; Dalla Marta et al. 2010; Jones et al. 2005a). Similar changes were also reported in some Australian viticultural regions (Webb et al. 2011; Sadras and Petrie 2011). Furthermore, advances in the phenological events resulting in ripening during a warmer period can have negative impacts on wine quality (Webb et al. 2008a). Vrsic and Vodovnik (2012) showed that higher temperatures during the growing season in North East 12 Chapter 1. An overview of climate change impacts on European viticulture Slovenia promoted a significant decrease in the total acidity content of early ripening varieties. Climate change projections for the 21th century are expected to have important impacts on viticulture, as changes in the temperature and precipitation patterns (Meehl et al. 2007) may significantly modify the current viticultural zoning in Europe (Malheiro et al. 2010). Recent climate change studies by Fraga et al. (2012b) for Portugal, Neumann and Matzarakis (2011) for Germany, and Duchene and Schneider (2005) for Alsace, France, hint at an increase in the growing season temperature. Santos et al. (2013), using a multi-model ensemble for the Douro Valley (Portugal), demonstrated that springtime warming may lead to earlier budburst under a future warmer climate, which may affect wine quality. Additionally, future projections for this same region, suggest higher grapevine yields (Santos et al. 2011) and wine productions (Gouveia et al. 2011), but also suggest increased risks of pests and diseases. Orduna (2010) argues that winemaking regions under extremely hot temperatures may lead to a significant increase in the risk of organoleptic degradation and wine spoilage. Under a future warmer climate, higher temperatures may inhibit the formation of anthocyanin (Buttrose et al. 1971), thus reducing grape color (Downey et al. 2006) and increasing volatilization of aroma compounds (Bureau et al. 2000). Future changes in minimum temperatures during ripening in the Iberian Peninsula were also reported (Malheiro et al. 2012; Fraga et al. 2012b), suggesting a decrease in wine quality. In the future scenarios a decrease in the suitability of the current winemaking regions in southern Europe might also be expected (Jones et al. 2005b; Stock et al. 2005; Fraga et al. 2012b). Southern European winegrapes are also expected to face adverse conditions due to severe dryness (Santos et al. 2003; Malheiro et al. 2010). In fact, these regions may become excessively dry for high quality winemaking (Kenny and Harrison 1992), or even unsuitable for grapevine growth without sufficient irrigation (Koundouras et al. 1999). Malheiro et al. (2010) stated that regions like Alentejo, Andalucía, Mancha, Sicily, Puglia, Campania will suffer from water deficits. Santos et al. (2012) also showed increased summer dryness in southern Europe. As an illustration, Alonso and O'Neill (2011) highlighted the negative impacts of climate change in the Spanish viticulture, which may result in increased water demand due to irrigation. Camps and Ramos (2012) found a decrease in winegrape yield for northeastern Spain that can be attributed to water deficits. Ruml et al. (2012), in a study for Serbia, identified changes that may require additional vineyard irrigation. In addition to a lowering of wine quality, expected in the future for some southern European winemaking 13 Chapter 1. An overview of climate change impacts on European viticulture regions, changes in the inter-annual variability and extremes may increase the irregularity of the yields (Schultz 2000; Jones et al. 2005a), with detrimental effects on the whole winemaking sector. In contrast to southern Europe, future warmer climates may be beneficial for many regions in central and western Europe, such as Alsace, Champagne, Bordeaux, Bourgogne, Loire Valley, Mosel and Rheingau (Malheiro et al. 2010; Neethling et al. 2012; Stock et al. 2005). Despite the projected increases in precipitation, which can be favorable to pests and diseases (e.g. downy mildew), the warming will enable the growth of a wider range of varieties (Malheiro et al. 2010). As an example, Eitzinger et al. (2009) project a doubling of the potential winegrape growing areas in Austria by 2050s. Hungarian southern wine regions are also expected to expand according to Gaal et al. (2012). Furthermore, the projected warming in central and northern European regions (e.g. Mosel) will result in prolonged frostfree periods and growing seasons (Bertin 2009), which will favor wine quality (Ashenfelter and Storchmann 2010). The enhanced concentrations of CO2 in the future, per se, are expected to have positive impacts on the grapevine development cycle and yield attributes (Bindi et al. 1996; Moutinho-Pereira et al. 2009; Goncalves et al. 2009). Higher CO2 will promote a decrease in plant transpiration, which will tend to overcompensate for the increased soil evaporation (Rabbinge et al. 1993), resulting in a reduced evapotranspiration in the future climate (Wramneby et al. 2010). This indirect effect of CO2 increase will be combined with the direct effect of an increase in carbon compound accumulation (Drake et al. 1997), which may thus provide a positive response to climate change. 1.5. Adaptation and mitigation measures While mitigation refers to measures that require human intervention, usually over long temporal periods, in reducing the sources or enhance the sinks of GHG (IPCC 2001), adaptation can be either a human or a natural response to the actual or expected climate change effects (IPCC 2001). Although the complex inter-relationship between adaptation and mitigation measures are noteworthy, mitigation measures are mainly determined by international agreements and national public policies, while adaptation measures involve local entities and private actions (Klein et al. 2007). 14 Chapter 1. An overview of climate change impacts on European viticulture Mitigation measures are crucial, since long term stabilization of CO2 concentrations may reduce damage to yield and quality (Easterling et al. 2007). Fischer et al. (2007) found that mitigation strategies, resulting in lower GHG concentrations, may reduce agricultural water requirements by about 40% when compared with unmitigated climate. As for agricultural mitigation measures, tillage systems are of key importance, as they may slightly compensate for GHG emissions (Ugalde et al. 2007). No-till systems and minimum tillage are considered the best for this purpose, as no disturbance of the soil surface promotes carbon retention/sequestration (Kroodsma and Field 2006). In regions with very steep slope (e.g. Douro Valley), no-till systems may also significantly contribute to reduce soil erosion. Short term adaptation measures may be considered as the first protection strategy against climate change and should be focused at specific threats, aiming at optimizing production. These measures mostly imply changes in management practices (e.g. irrigation, sunscreens for leaf protection), while changes in the enological practices, through technological advances (Lobell et al. 2006), may also have positive effects on wine quality. Long term adaptation measures mainly include varietal and land allocation changes, as some regions may become excessively warm and dry, while others consistently show high winemaking suitability (Malheiro et al. 2010). Changes to cooler sites, to higher altitudes or coastal areas, may also prove beneficial for future vineyards (changes in the vineyard microclimatic and mesoclimatic conditions). These regions may however struggle against an increasing risk of pests and diseases, requiring more intense plant protection. In this case, biological control agents should be preferably used, thus reducing the environmental impacts (Butt and Copping 2000). Adequate and timely planning of the adaptation measures need to be adopted by the winemaking sector. The readiness to implement adaptation measures is highly correlated with the degree of changes planned, independently of climate change (Battaglini et al. 2009). A deeper insight into some adaptation measures is discussed in the next subsections. 1.5.1.Varietal and rootstock decisions The optimum climate for a given variety produces consistent yields, balanced fruit composition and acceptable vintage variation (Jones 2006). A key factor in adapting to climate change may include the growing of varieties with different thermal requirements and higher summer stress resistance. Currently, despite the large number (thousands) of existing 15 Chapter 1. An overview of climate change impacts on European viticulture varieties, the global wine market is dominated by only a few of them (e.g. Airén, Cabernet Sauvignon, Chardonnay, Merlot, Pinot Noir, Tempranillo, Touriga Nacional, Riesling), owing to current trading practices (This et al. 2006). In the future, some northern European regions may benefit from a wide range of varieties for winegrape growth (Stock et al. 2005), while southern Europe will need to adapt to varieties more suitable to warmer and dryer climates. Jones (2006) gives clues to some varieties more adapted to warmer climates, such as Cabernet Franc, Cabernet Sauvignon, Malbec, Merlot, Syrah, Tempranillo (Fig. 1.4). White et al. (2006) proposes that grapevine breeding programs should now focus on the development of heat-resistant varieties. Still with respect to this issue, Duchene et al. (2012) developed a framework for genetic breeding of new varieties more adapted to future climatic conditions, while maintaining some key aspects to existing varieties. Furthermore, owing to the large number of existing grapevine varieties, the maintenance of the natural biodiversity is essential for the better adaptation to climate change (Tello et al. 2012). Rootstock is another factor that may affect yield, quality and other vine physiological parameters (Hedberg et al. 1986; Pavlousek 2011). Its effects under warm and dry climates should be taken into account, since rootstocks show complex interaction with soil water availability (Romero et al. 2006). Several studies have been undertaken to assess rootstock effects under different water conditions (Harbertson and Keller 2012; Pavlousek 2011; Ozden et al. 2010). As an example, Koundouras et al. (2008) used the Cabernet Sauvignon variety in Greece and compared the effect of two different rootstocks, concluding that the 1103P is better for winegrape growth under semi-arid conditions, while S04 is preferable where no water limitation exists. As such, accessing the more adapted rootstock for each case can improve WUE, thus improving yield and quality. 1.5.2. Irrigation management Regions under conditions of higher water scarcity will need to improve grapevine WUE (Flexas et al. 2010), thus lowering water usage for irrigation purposes (Chaves et al. 2010). Deficit irrigation strategies (e.g. regulated deficit irrigation – RDI; partial root-drying – PRD; sustained deficit irrigation – SDI) can be used to improve WUE, allowing an optimal grape maturity and wine quality. Different deficit irrigation techniques are usually achieved by assessing soil (water potential and moisture) and physiological parameters (leaf or stem water 16 Chapter 1. An overview of climate change impacts on European viticulture potential, relative transpiration using sap flow techniques, trunk growth variations, canopy temperature and chlorophyll fluorescence) as water status indicators (Centeno et al. 2010; Pellegrino et al. 2004; Cifre et al. 2005; Sousa et al. 2006). Romero et al. (2010), using RDI by applying 60% crop evapotranspiration (ETc) water for irrigation during the growing season, found significant increases in WUE on Monastrell grapevines. Junquera et al. (2012) found that irrigating at 45% of the reference evapotranspiration (ET0) shows the best results for Cabernet-Sauvignon. Conversely, a recent study by Basile et al. (2012) reported that RDI may also affect wine sensory attributes. PRD is another deficit irrigation strategy that has been adapted for viticulture to improve WUE (Chaves et al. 2007; Romero and Martinez-Cutillas 2012). In this system, half of the plant root system is slowly dehydrating, whereas the other half is irrigated, decreasing water usage by 50%. Using this technique, Santos et al. (2005b) reported no reduction in yield for the Castelão grapevines. Improvements in quality were also reported using this irrigation strategy on other varieties, resulting from increased anthocyanin concentration (Du et al. 2008; Poni et al. 2007; Bindon et al. 2008). Chalmers et al. (2010) implemented a SDI strategy by applying a lower volume of water at each irrigation event. These authors state that this may increase soil water tension by not replenishing the entire root-zone. In this study, wine anthocyanin and other phenolic compounds (for Cabernet Sauvignon and Shiraz) showed significant increases using this irrigation strategy. Along the previous lines, it can be stated that a sustainable water management (together with available water) may be a profitable economic strategy for the grape grower (Garcia et al. 2012), providing a compromise solution between environmental costs and plant water requirements, which is highly pertinent under increasingly dryer and water demanding southern European climates (Bruinsma 2009). 1.5.3. Tillage treatments Different tillage treatments can also affect yield and quality (Bahar and Yasasin 2010). These treatments may differ in both the start and duration of the tillage process. In the no-till or conventional tillage (CVT), natural vegetation is usually allowed to grow spontaneously. In a minimum tillage (MIT) treatment, periodic tillage is applied from the fruit-set to veraison. In the conservative tillage (CST), some vegetation is always allowed to grow, although 17 Chapter 1. An overview of climate change impacts on European viticulture periodic tillage is made from the beginning of the growing season to the end of veraison. While most of these treatments have been used to stimulate plant competition, aiming to increase vine performance and wine quality (Bahar and Yasasin 2010), they may be required in order to account for future climate change too. Monteiro and Lopes (2007) found that although intensive soil tillage increases soil moisture during spring, cover cropping increases upper layer soil moisture from veraison to harvest, indicating lower soil evaporation caused by the cover cropping. Moreover, Xi et al. (2010) reported that a cover crop increased total phenols in berry and wine, thereby improving wine quality. Judit et al. (2011) for the Tokaj region (Hungary), reported that covering the soil with straw mulch had a positive effect on the soil water content, while using a cover crop implied higher water demand. Celette et al. (2008) demonstrated that the use of cover cropping is increasing throughout European vineyards except in the Mediterranean regions, due to the possibility of competition for water resources. These authors stated that cover crops can modify grapevine root systems in order to explore other soil zones, increasing water intake. Also, the use of a cover crop allows better wintertime soil water renewal. 1.5.4. Vineyard microclimate In order to account for the joint effect of increased temperature, water stress and high solar radiation under future climates, adjustments in the traditionally settled training systems can also be applied by optimizing canopy management (Pieri and Gaudillere 2003). Row orientation should also be considered (if feasible), since it is one of the main factors influencing solar radiation interception (Intrieri et al. 1998; Grifoni et al. 2008). The use of shading nets, often used to protect agricultural crops from excessive solar radiation, can be applied to viticulture (Shahak et al. 2008). In fact, covering the vines with a shading material can help reducing heat stress at the cost of reduced vine biomass (Greer et al. 2011). Furthermore, the use of mineral chemically inert sunscreens for leaf protection against sunburns may prove to be an important alternative (Glenn et al. 2010; Pelaez et al. 2000). 1.6. Conclusions Climate change is expected to bring new challenges to the European viticultural sector (Malheiro et al. 2010). Although grapevines have several survival strategies (e.g. deep root 18 Chapter 1. An overview of climate change impacts on European viticulture system, efficient stomatal control), viticulture is strongly dependent on climate. Hence, the amounting evidence for significant climate change in the upcoming decades urges adaptation measures to be taken. A lowering in the suitability of some important winemaking regions was indeed reported (Hall and Jones 2009). Therefore, appropriate measures need to be adopted by the viticultural sector to face climate change impacts, mainly by developing suitable adaptation and mitigation strategies at regional scales (Metzger et al. 2008). Winegrape growers are becoming progressively more aware of this problem (Battaglini et al. 2009), as strategic planning will provide them a competitive advantage over competitors. Nevertheless, to effectively cope with the projected changes, short and long term strategies deserve much greater attention in future research (Metzger and Rounsevell 2011). Even though the full extent of the contribution of the adaptive strategies in reducing climate change impacts is still unclear (Lobell et al. 2006), adaptation strategies to climate change can be highly beneficial for the agricultural sector (Tubiello and Fischer 2007). As an illustration, Reidsma et al. (2010) concluded that adaptation measures can largely reduce the impacts of climate change and climate variability on crop yields. A selection of alternative crop species, grapevine cultivars, rootstocks and changes in the management practices are amongst the measures already being taken by European viticulturists to adapt to climate change (Olesen et al. 2011). 19 Chapter 7. Integrated analysis of climate, soil, topography and vegetative growth in Iberian viticultural areas 20 Chapter 2. Future scenarios for viticultural zoning in Europe: ensemble projections and uncertainties International Journal of Biometeorology, 2013, 57: 909-925 Impact Factor (5-yr): 2.785 Helder Fraga, Aureliano C. Malheiro, José Moutinho-Pereira, João A. Santos 21 Chapter 2. Future scenarios for viticultural zoning in Europe: ensemble projections and uncertainties Abstract Optimum climate conditions for grapevine growth are limited geographically and may be further challenged by a changing climate. Due to the importance of the winemaking sector in Europe, the assessment of future scenarios for European viticulture is of foremost relevance. A 16-member ensemble of model transient experiments (generated by the ENSEMBLES project) under a greenhouse gas emission scenario and for two future periods (2011-2040 and 2041-2070) is used in assessing climate change projections for six viticultural zoning indices. After model data calibration/validation using an observational gridded daily dataset, changes in their ensemble means and inter-annual variability are discussed, also taking into account the model uncertainties. Over southern Europe, the projected warming combined with severe dryness in the growing season is expected to have detrimental impacts on the grapevine development and wine quality, requiring measures to cope with heat and water stress. Furthermore, the expected warming and the maintenance of moderately wet growing seasons over most of the central European winemaking regions may require a selection of new grapevine varieties, as well as an enhancement of pest/disease control. New winemaking regions may arise over northern Europe and high altitude areas, when considering climatic factors only. An enhanced inter-annual variability is also projected over most of Europe. All these future changes will thereby impose new challenges for the European winemaking sector. 22 Chapter 2. Future scenarios for viticultural zoning in Europe: ensemble projections and uncertainties 2.1. Introduction Climate has a predominant role on grapevine growth (e.g. van Leeuwen et al. 2004; Santos et al. 2011), as vine physiology and its development phases are mostly determined by specific environmental conditions (e.g. Magalhães 2008). In fact, monthly mean temperatures and precipitation totals in the growing season present significant correlations with grapevine yield in many regions (e.g. Makra et al. 2009; Santos et al. 2013). Grapevine is a heat demanding crop, requiring a 10ºC basal temperature for its growing cycle onset and development (Amerine and Winkler 1944; Winkler 1974), and relatively high solar radiation intensities (e.g. Magalhães 2008). However, prolonged exposure to excessive heat (e.g. temperatures above 40ºC) may have detrimental impacts on some physiological processes (Berry and Bjorkman 1980; Osorio et al. 1995), resulting in poor yields and quality (Kliewer 1977; Mullins et al. 1992). Although grapevines are also resistant to relatively low temperatures (lower thermal lethal limit of approximately -17ºC) during early stages (Hidalgo 2002), frost occurrences during spring can severely damage crop production (e.g. Spellman 1999). Further, excessive humidity in spring can trigger pests and diseases, such as downy mildew (Carbonneau 2003), while severe dryness during the growing season can also lead to harmful water stress (Koundouras et al. 1999) thus leading to reductions in grapevine productivity (Moutinho-Pereira et al. 2004). Along the previous lines, climate change brings new and major challenges for winegrape growers. In fact, projected future changes over Europe under the A1B International Panel on Climate Change (IPCC) – Synthesis Report on Emission Scenarios (SRES) scenario (Nakićenović et al. 2000) include a temperature increase of 2.3-5.3ºC in northern Europe and 2.2-5.1ºC in southern Europe until the end of the twenty-first century (Christensen et al. 2007). In addition, for the same scenario, it has been shown that temperature extremes are also expected to increase throughout Europe (Andrade et al. 2012). Changes in the inter-annual variability and extremes in the climatic factors result in shifts in grapevine phenology, disease and pest patterns, lower predictability and regularity of the yields and wine quality (Schultz 2000; Jones et al. 2005a), being thus an additional pitfall for the winemaking sector. Conversely, benefits coming from the increase in CO2 concentrations under future atmospheric conditions may also play a key role in grapevine physiology and yield attributes (Moutinho-Pereira et al. 2009), though this forcing is out of the scope of the present study. 23 Chapter 2. Future scenarios for viticultural zoning in Europe: ensemble projections and uncertainties Taking this climatic forcing into account, current grapevine global geographical distribution is largely limited to regions where the growing season mean temperature (AprilSeptember in the Northern Hemisphere) is within the 12–22°C range (Jones 2006). Taking the aforementioned delicate balance between environmental conditions and viticultural zoning into account, climate change is likely to have considerable regional impacts not only on grapevine attributes, but also on wine quality (e.g. Jones and Davis 2000a; Jones et al. 2005b). Due to the high socioeconomic relevance of the winemaking sector in many European regions, the assessment of the impacts of climate change on viticulture is of utmost importance, particularly for the most renowned winemaking regions spread over the continent. Winegrape growers are indeed becoming increasingly aware of these changes and their resulting impacts (Battaglini et al. 2009). Future climatic zones are then expected to shift polewards, leading to changes in the current regions suitable for grapevine growing and to new potential winemaking regions. In addition, due to the projected decrease in the annual precipitations over southern Europe (Christensen et al. 2007), grapevines are also expected to be negatively affected be severe dryness (Koundouras et al. 1999; Santos et al. 2003). Several studies in different areas of research used ensembles of model runs in assessing climate change projections under anthropogenic radiative forcing. As simple illustrations, Lobell et al. (2006) used a set of regional climate models for assessing the impacts of climate change on perennial crop yields in California, while Heinrich and Gobiet (2011) used an 8member ensemble for future projections of dry/wet spells in Europe. The use of multi-model ensemble projections enables to quantify numerical model uncertainties, arising from differences in physics and modelling approaches, model parameterizations and initializations, among others. This is important because, model uncertainty may lead to different outcomes (Deser et al. 2012). Another advantage of this approach is that multi-model ensembles commonly outperform studies based on projections from a single model (Knutti et al. 2010). Besides using statistical methodologies for model validation/calibration, enabling the correction of model bias, the quantification of the uncertainty associated to climate change projections may be as important to the winemaking sector as the climate change signal itself. In the present study, potential changes in the suitability of a given region for winegrape growth in Europe under human-driven climate change are assessed. For this purpose, several bioclimatic indices, specifically developed for viticultural zoning, are applied using datasets from a 16-member multi-model ensemble under the A1B emission scenario for 2011-2040 24 Chapter 2. Future scenarios for viticultural zoning in Europe: ensemble projections and uncertainties and 2041-2070 (future periods). Hence, this study is structured as follows: 1) a multi-model ensemble is used for assessing climate change projections in a set of viticultural bioclimatic indices, 2) a validation with a state-of-the-art observational gridded daily dataset is conducted and calibration techniques are then applied to the model outputs, 3) the ensemble projections and the corresponding model uncertainties are analysed, 4) an analysis based on a categorized Composite Index (CompI) is carried out and 5) the projected changes in the inter-annual variability are assessed. In the next section, the bioclimatic indices are defined, model data is presented, model output statistics (MOS) and model uncertainty are discussed. In section 3, after an initial model skill inter-comparison, the bioclimatic indices are presented, along with their inter-annual variability. An analysis of each individual model of the 16-member ensemble is also performed, depicting the cost of using single models versus the use of model ensembles. In the last section, the main outcomes of this study will be summarized and discussed. 2.2. Materials and Methods 2.2.1. Bioclimatic indices The implications of climate change on viticultural zoning in Europe were assessed on the basis of projections for the following bioclimatic indices: Huglin Index (HI; Huglin 1978), Dryness Index (DI; Riou et al. 1994), Hydrothermal Index (HyI; Branas et al. 1946), Growing Season Suitability (GSS; Jackson 2001) and Growing Season Precipitation (GSP; BlancoWard et al. 2007). The mathematical definitions of all these indices, as well as their main references, are listed in Table 2.1, and will not be further detailed here. Following two previous studies (Malheiro et al. 2010; Santos et al. 2012), an improved Composite Index (CompI) is also computed in the current study. The CompI at a given location is the ratio of “optimal years” for winegrape growth over a given time period to the total number of years in the period. An “optimal year” must simultaneously fulfil the following criteria: 1) HI ≥ 900ºC; 2) DI ≥ -100 mm; 3) HyI ≤ 7500ºC mm; and 4) total absence of days with minimum temperature below -17ºC. This new CompI differs from the previous definitions in the first and third thresholds (cf. Malheiro et al. 2010; Santos et al. 2012). A lower HI threshold (900 instead of 1200ºC) is considered herein in order to include viticultural regions in northern Europe with marginally suitable winemaking conditions. In 25 Chapter 2. Future scenarios for viticultural zoning in Europe: ensemble projections and uncertainties fact, in two previous studies by Jones et al. (2010) and Hall and Jones (2010), a Winkler Index (WI; Winkler 1974) above 850ºC was found to be already suitable for winegrape growth in western USA and Australia, respectively. Additionally, (Santos et al. 2012) showed a clear correspondence between the WI and HI patterns in Europe using 850ºC and 900ºC as lower limits, respectively. The threshold in the third criterion of the CompI (HyI ≤ 7500ºC mm) was chosen so that only climatic conditions noticeably favourable to pests/diseases in the vineyards are excluded. Furthermore, when considering a lower threshold (e.g. 5100ºC mm), many winemaking regions in central Europe become misleadingly unsuitable (Santos et al. 2012). Table 2.1 - List of the bioclimatic indices used for climatic viticultural zoning in Europe, along with their corresponding definitions and references. The calculations (summations) are carried out using daily climate variables. Bioclimatic Definition Units Ratio of days with T ≥ 10°C ºC Index Growing Season Suitability (GSS) Growing Season Precipitation (GSP) Suitable Threshold - References (Jackson 2001) Sept. (P) April mm - (Blanco-Ward et al. 2007) ºC > 900 (Huglin 1978) ºC mm <7500 (Branas et al. 1946) P – Precipitation (mm) (T 10) (Tmax 10) d 2 April Sept . Huglin index (HI) T –Mean temperature (ºC); Tmax –Maximum temperature (ºC) d – Length of day coefficient Aug . (T * P) Hydrothermal index (HyI) April T – Mean temperature (ºC);P – Precipitation (mm) Sept. (Wo P Tv Es ) April (Riou et al. Wo – Soil water reserve (mm)a; mm >-100 1994) P – Precipitation (mm); Tv - the potential transpiration in the vineyard (mm); Es - Direct evaporation from the soil (mm) a Wo should be equal to 200 mm in the beginning of the growing season. For the following months Wo takes the value of the DI in the previous month. Dryness index (DI) 26 Chapter 2. Future scenarios for viticultural zoning in Europe: ensemble projections and uncertainties In order to better discriminate the optimal climatic requirements of different winegrape varieties (Jones et al. 2005a), an innovative category analysis of the CompI is also carried out. At each site, the CompI is first classified as a function of three pre-defined HI classes , namely: 900 ≤ HI < 1500ºC; 1500 ≤ HI < 2100ºC; 2100 ≤ HI < 3000ºC. Then the leading HI category is identified an assigned to the site. HI values below 900ºC are considered unsuitable for winegrape growth while values above 3000ºC are relatively rare in Europe (not shown). This analysis allows a variety-dependent assessment of the climatic change impacts on the viticultural sector. 2.2.2. Model data Climate models remain a valuable tool in climate change assessments (Solomon et al. 2011), despite their widely reported limitations (e.g. Knutti et al. 2008; Hulme and Mahony 2010), and are used here as a source of information about future climate change, namely in the temperature and precipitation fields in Europe. Regarding the definitions of the selected bioclimatic indices (Table 2.1), daily mean, maximum and minimum 2-meter air temperatures and daily precipitation totals are used in their calculations. These four daily atmospheric variables were extracted from datasets of 16 state-of-the-art transient experiments generated by the EU-FP6 project ENSEMBLES (http://ensembles-eu.metoffice.com; van der Linden and Mitchell 2009), in a total of 15 different Global Climate Model / Regional Climate Model (GCM/RCM) chains (for the ECHAM5/COSMO-CLM combination, two ensemble simulations are used). Table 2.2 lists the 16 model experiments, along with their acronym and main references. Only these 16 transient experiments provide records without data gaps in the two selected future periods (2011-2040 and 2041-2070). However, they provide a sufficiently representative sample of models, with different parameterizations, spin ups, physics and modeling approaches, covering a large amount of the uncertainty inherent to numerical model simulations. 27 Chapter 2. Future scenarios for viticultural zoning in Europe: ensemble projections and uncertainties Table 2.2 - Summary table of all GCM / RCM model chains used in this study. The corresponding acronyms, institutions and original grid resolutions are also listed. In all simulations the period 2011-2070 was used under the IPCC-SRES A1B scenario. Relevant references to each chain are also indicated. Acronym GCM RCM Original grid Institution References ARP Aladin(CNRM) ARPEGE-RM5.1 CNRM-Aladin 25 km CNRM (Gibelin and Deque 2003) ARP HIRHAM(DMI) ARPEGE DMI-HIRHAM 25 km DMI (Christensen et al. 1996) BCM RCA(SMHI) BCM SMHI-RCA 25 km SMHI ECHAM5-r3 KNMI-RACMO2 25 km KNMI EH5 CLM1(MPI) ECHAM5-r1 COSMO-CLM-1 18 km MPI-M EH5 CLM2(MPI) ECHAM5-r2 COSMO-CLM-2 18 km MPI-M EH5 HIRHAM(DMI) ECHAM5-r3 DMI-HIRHAM 25 km DMI EH5 RCA(SMHI) ECHAM5-r3 SMHI-RCA 25 km SMHI EH5 RegCM(ICTP) ECHAM5-r3 ICTP-RegCM3 25 km ICTP EH5 RACMO(KNMI) EH5 REMO(MPI) HC CLM(ETHZ) HC HadRM3Q0(HC) ECHAM5-r3 HadCM3Q0 (normal sens) HadCM3Q0 (normal sens) HC HadCM3Q16 HadRM3Q16(HC) (high sens) HC HadRM3Q3(HC) HC RCA(SMHI) HC RCA3(C4I) HadCM3Q3 (low sens) HadCM3Q3 (low sens) HadCM3Q16 (high sens) MPI-REMO 25 km MPI-M (Kjellström et al. 2005) (Samuelsson et al. 2011) (Lenderink et al. 2003) (Böhm et al. 2006) (Steppeler et al. 2003) (Böhm et al. 2006) (Steppeler et al. 2003) (Christensen et al. 1996) (Kjellström et al. 2005) (Samuelsson et al. 2011) (Elguindi et al. 2007) (Pal et al. 2007) (Jacob and Podzun 1997) (Jacob 2001) (Steppeler et al. 2003) ETHZ-CLM 25 km ETHZ HC-HadRM3Q0 25 km HC (Collins et al. 2011) HC-HadRM3Q16 25 km HC (Collins et al. 2011) HC-HadRM3Q3 25 km HC (Collins et al. 2011) SMHI-RCA 25 km SMHI C4I-RCA3 25 km C4I (Jaeger et al. 2008) (Kjellström et al. 2005) (Samuelsson et al. 2011) (Kjellström et al. 2005) (Samuelsson et al. 2011) The A1B IPCC-SRES scenario (2001-2100) was used, which corresponds to a moderate anthropogenic radiative forcing (Nakićenović et al. 2000). Other SRES emission scenarios, such as A2, are not available for all ensemble members and are therefore not considered for the present analysis. However, the A1B and A2 scenarios only start to clearly diverge in their emission pathways from 2070 onwards (Nakićenović et al. 2000), later than the aforementioned two future periods. Daily model data was also extracted for the 40-year baseline period (C20; 1961-2000) so as to validate/calibrate it using an observational dataset. The daily station-based gridded dataset (E-OBS, version-4) from the EU-FP6 project ENSEMBLES (http://ensembleseu.metoffice.com), provided by the European Climate Assessment & Dataset (ECA&D) project (http://eca.knmi.nl) in the baseline period, was used for this aim. The original gridded 28 Chapter 2. Future scenarios for viticultural zoning in Europe: ensemble projections and uncertainties data is defined over land areas at 201 grid boxes along latitude and 464 grid boxes along longitude (0.25º latitude × 0.25º longitude grid). This dataset, despite some limitations, represents a valuable resource for climate research in Europe (Hofstra et al. 2009). Detailed information about the E-OBS dataset can be found in (Haylock et al. 2008). The baseline period (1961-2000) corresponds to the longest available common time-series amongst the different datasets (observations and simulations). All model datasets were bi-linearly interpolated (in latitude and longitude) from the original rotated grids (cf. their original spatial resolution in Table 2.2) to a regular grid of 0.25º latitude × 0.25º longitude to enable model validation/calibration using exactly the same grid as in E-OBS. Only the geographical sector [35-60ºN; 12ºW-36ºE] is considered herein (102 grid boxes along latitude × 190 grid boxes along longitude), where viticultural zoning is relevant. The selected bioclimatic indices were then computed for all grid points with available data (excluding blank/missing data) in the E-OBS (11 506 grid cells). Statistically significant differences at a 99% confidence level (using the two-sample Student’s t-test) between the future period of 2041-2070 and the baseline period for HI, DI, HyI and CompI are also computed and plotted. 2.2.3. Model output statistics and model uncertainties Several studies demonstrated that equal weights are generally a better approach than performance-based weights for computing the ensemble mean statistics (Christensen et al. 2010; Weigel et al. 2010). For this reason, only equally-weighted ensemble mean patterns for 2041-2070 (higher greenhouse gas forcing) are presented henceforth. However, an intercomparison of the model performances in reproducing the observed mean patterns of the selected bioclimatic indices in the baseline period (1961-2000) was undertaken in order to summarize and highlight the most important deviations (bias) between simulated and E-OBS mean patterns. Scatterplots representing the different models as a function of their spatial mean bias (MB) and absolute spatial mean bias (AMB; less sensitive to extremes then the root mean square deviation) in the HI, DI and HyI patterns are shown as supplementary material (Fig. S1.1). Due to the model bias referred above, MOS are used to fit raw model data to observations, as the statistical distributions of the simulated data present biases with respect to observations (Wilks 2006). This is also a common procedure when assessing climate 29 Chapter 2. Future scenarios for viticultural zoning in Europe: ensemble projections and uncertainties projections (Mearns et al. 2001). Linear transfer-functions are applied to obtain transformed (adjusted) data with the same mean patterns as the observational data. Taking into account the large number of grid boxes (11 506), higher order polynomial, exponential or logarithmic transformations were not tested. This is not a clear shortcoming, mainly because the bioclimatic indices are defined on a yearly basis and most of them are normally distributed, according to the Lilliefors test applied to all indices and simulations (not shown); only a few exceptions were identified over some mountainous areas (e.g. the Alps). The linear transformations are then applied to the yearly bioclimatic indices, at each grid box, and for each model dataset individually. The transfer-functions between indices calculated from E-OBS and indices calculated from the models were estimated for the 40-year baseline period (1961-2000). This scaling procedure has been used in previous studies (Alexandrov and Hoogenboom 2000; Santos et al. 2013; Fraga et al. 2012b; Jakob Themeßl et al. 2011) and enables the correction of biases in the model simulations, also enabling a comparison between the different RCMs. As such, for 1961-2000, the corrected data of each index are equal amongst the different simulations and equal to the corresponding E-OBS data. Therefore, for the baseline period, only results obtained using E-OBS are presented. The same transfer-functions (coefficients) were applied to all datasets in the two future periods (20112040 and 2041-2070), leading to corrected patterns for future conditions. In this procedure, the temporal invariance of the transfer-function between recent-past and future is a basic underlying assumption. In any case, it must be kept in mind that the patterns of the mean differences between a future period and the baseline period (climate change signal) are independent of these corrections. Also, applying transfer-functions to the indices and not to the original climatological data is a new methodology that may require further research. Measuring model uncertainty is also crucial when assessing future impacts of climate change. For this purpose, the 16-member normalized interquartile range (third quartile minus first quartile divided by the mean at each grid point; NIQR hereafter) of the HI, DI, HyI and CompI is also assessed and plotted. The spatial correlations of the CompI between the 16 simulations and the ensemble mean are also presented so as to clarify the spatial coherence amongst them. 30 Chapter 2. Future scenarios for viticultural zoning in Europe: ensemble projections and uncertainties 2.2.4. Inter-annual variability As previously stated, irregularity in the yields and wine quality is an additional pitfall for the winemaking sector. Significant alterations in the inter-annual variability could lead to large economic impacts for this perennial crop. The ratios between the averages of the 16 inter-annual standard-deviations (calculated for each of the 16 ensemble members separately) in 2041-2070 and 1961-2000 of the bioclimatic indices (HI, DI and HyI) are computed to assess changes in inter-annual variability. 2.3. Results 2.3.1. Recent-past viticultural zoning The HI has shown to be an effective tool for viticultural zoning and has been thus widely applied (Jones et al. 2010). Its ensemble mean pattern (Fig. 2.1a) is highly coherent with the GSS pattern (not shown) and highlights the fact that large areas of southern and central Europe are suitable for winegrape growth, whereas regions northwards of the 53ºN parallel are generally unsuitable. In this context, it should also be emphasized that different classes of the HI are related to grapevine varieties with different thermal requirements (Huglin 1978). Hence, lower values do not necessarily mean lower suitability, but rather conditions that might be optimal for specific varieties (e.g. white grapevine varieties are generally favoured by cooler climates; Duchene and Schneider 2005). The mean patterns for the DI and for the HyI (Fig. 2.1b and Fig. 2.1c.) give opposite perspectives of the humidity requirements for optimal grapevine development. Severe dryness (assessed by the DI) and excessive humidity (assessed by the HyI) during the growing season commonly have detrimental impacts on the different stages of the grapevine development (Branas et al. 1946). Overall, the DI pattern depicts a remarkable contrast between southern Europe and central and northern Europe, and shows that dryness can only be a limitation over small areas of southern Europe and in certain years (values below -100 mm), whilst excessive humidity levels (HyI above 5100ºC mm) are generally restricted to Atlantic coastal areas or to mountainous areas, such as the Alps or the Carpathian range. The baseline period mean pattern for the CompI (Fig. 2.1d) demonstrates this index usefulness in viticultural zoning, since it is coherent with the spatial distribution of well-known traditional viticultural regions, showing higher suitability in the southern European areas. 31 Chapter 2. Future scenarios for viticultural zoning in Europe: ensemble projections and uncertainties Fig. 2.1 - (a) Huglin Index, (b) Dryness Index, (c) Hydrothermal Index and (d) Composite Index for the baseline period (1961-2000). (e) European wine regions (circles) with the corresponding Composite Index value (cf. scale) for the period of 1980-2009 using the E-OBS data. Source of the wine regions locations: “Wine Regions of the World - Version 1.31” (URL: http://geocommons.com/overlays/3547). 2.3.2. Composite index vs. viticultural regions As mentioned in section 2.1, the CompI is an attempt to characterize the most relevant atmospheric requirements for winegrape growth with a single index. According to our 32 Chapter 2. Future scenarios for viticultural zoning in Europe: ensemble projections and uncertainties analysis, areas characterized by values of the CompI in excess of 0.5 (50% of optimal years) in the period 1980-2009 encloses the most worldwide famous winemaking regions in Europe (circles in Fig. 2.1e), which are mostly located over southern and central Europe, particularly in countries such as France, Italy, Spain, Portugal and Germany). This attest the utility of this index for the European viticultural zoning. The more recent period 1980-2009 was chosen because it more closely reveals the present time atmospheric conditions reflected in the current winemaking regions. As E-OBS data, used for model calibration, is available for 697 out of a total of 754 winemaking regions, only the former regions are taken into account (most of the missing regions are islands or coastal areas). From these 697 regions, about 93% present a CompI equal to or above 0.5, i.e., are in agreement with the CompI pattern. 2.3.3. Future viticultural zoning The HI patterns for the future period (Fig. 2.2a) shows a northward displacement and a stronger increase over southern and Western Europe. This reveals an apparent northward extension of the high suitability areas for 2041-2070. In fact, new suitable regions for grapevine development within the latitude belt 50-55ºN are projected to arise, which is in line with the results obtained for GSS (not shown). Furthermore, important changes over southern Europe should also be expected under anthropogenic radiative forcing. For the DI in the future period (Fig. 2.2b) there is a significant drying over most of the southern half of Europe, which is in agreement with the projected changes in the GSP (not shown). These changes are likely to yield severe dryness (DI < -100 mm) in areas such as southern Iberia, Greece and Turkey. On the other hand, the projected changes in the HyI (Fig. 2.2c) reveal an enhancement of the humidity levels over central and eastern Europe, largely explained by the joint effect of warmer and moister conditions in the future (increase in the GSP and in the growing season temperature). Therefore, while dryness may represent a threat/challenge for winegrape growth in southern Europe (e.g. harmful water-stress), excessive humidity in central and eastern Europe can potentially trigger pests and diseases in the vineyards (e.g. outbreaks of downy mildew disease). 33 Chapter 2. Future scenarios for viticultural zoning in Europe: ensemble projections and uncertainties The ensemble mean pattern for the CompI (Fig. 2.2d) reveal a general decrease in suitability in southern European regions, especially due to the increased dryness in these areas. Conversely, large areas of central and Western Europe are projected to become more suitable for viticulture, due to the more favorable thermal conditions. Fig. 2.2 - As in Fig. 2.1 but for the future period of 2041-2070 under the A1B IPCC-SRES scenario. 2.3.4. Climate change signals and uncertainties When assessing climate change projections, it is not only important to analyse the climate change signal itself, but also the respective model uncertainties (Deser et al. 2012). The spatial correlations for the CompI pattern among the 16-members and their mean (Table S1.1) shows that most models are clearly inter-correlated, as well as with their ensemble mean (correlation coefficients above 0.7). The models showing higher correlations with the ensemble mean are the ARP HIRHAM (DMI), EH5 REMO (MPI), HC CLM (ETHZ) and HC HadRM3Q0 (HC). On the other hand, the models showing lower correlations are BCM RCA (SMHI) and EH5 RegCM (ICTP). 34 Chapter 2. Future scenarios for viticultural zoning in Europe: ensemble projections and uncertainties CompI maps for the models with the lowest (HC RCA3 (C4I); Fig. 2.3a) and highest (HC RCA (SMHI); Fig. 2.3b) spatial means of this index are also presented. Although their spatial patterns may show important differences, they are considered as equally probable in the present study (equal weights in the statistical measures). Fig. 2.3 - (a) Composite index of the more severe model (HC RCA3 (C4I)). (b) Composite index showing the less severe model (HC RCA (SMHI)), for the time period of 2041-2070. To assess the ensemble variability in the future projections (uncertainty) for the CompI in 2041-2070, some statistical measures (means, medians, minima, maxima, NIQR and total ranges – TR) are provided in Table 2.3 for a selection of well-known grapevine growing regions throughout Europe. The means reveal CompI values always higher than 0.80 but for two regions (Alentejo-Borba, Tokaj-Hegyalja; cf. Fig. 2.2d). The medians are consistently higher than the means, which reflects the negative skewness of the distributions (CompI upper limit of 1.0). The minima show a large variability, with values ranging from 0.00 (TokajHegyalja) up to 0.99 (Champagne), while the maxima are always 1.00, with only two exceptions (Alentejo-Borba, Tokaj-Hegyalja). Regarding the ranges of variability, since TR is more affected by extremes than NIQR, some discrepancies between them are apparent; a higher TR does not necessarily imply a higher NIQR or vice-versa. Considering TR as an uncertainty measure, some regions reveal more pronounced uncertainties (Rheinhessen, Ribera del Duero, Chianti, Porto/Douro, Barolo, La Mancha, Alentejo-Borba, TokajHegyalja), whereas others present relatively low uncertainties and CompI minima above 0.70 (Champagne, Coteaux du Loire, Bordeaux, Mosel, Rioja, Rheingau, Vinhos Verdes, Alsace). 35 Chapter 2. Future scenarios for viticultural zoning in Europe: ensemble projections and uncertainties Table 2.3 Ensemble means, medians, minima, maxima, normalized interquartile ranges (NIQR) and total ranges (TR) of the Composite Index in 2041-2070 and for a selection of European famous winemaking regions (the respective grid box coordinates are also listed). Regions are ranked in ascending order with respect to their TR values. Country Region France Champagne France Coteaux du Loire France Bordeaux Germany Mosel Spain Rioja Germany Rheingau Portugal Vinhos Verdes France Alsace Germany Rheinhessen Spain Ribera del Duero Italy Chianti Portugal Porto/Douro Italy Barolo Spain La Mancha Portugal Alentejo-Borba Hungary Tokaj-Hegyalja Longitude/ Latitude (º) 4.003/ 49.155 -0.846/ 47.368 -0.055/ 45.012 6.734/ 49.765 2.402/ 42.493 7.944/ 49.988 -8.545/ 41.570 7.6592/ 48.660 8.254/ 49.936 -4.465/ 41.609 11.040/ 43.640 -7.555/ 41.170 8.545/ 41.570 2.698/ 39.653 -7.424/ 38.797 21.349/ 48.187 Mean Median Minimum Maximum NIQR TR 0.98 1.00 0.93 1.00 0.03 0.07 0.99 1.00 0.90 1.00 0.02 0.10 0.96 0.97 0.90 1.00 0.10 0.10 0.95 0.97 0.87 1.00 0.08 0.13 0.96 1.00 0.78 1.00 0.05 0.22 0.96 1.00 0.77 1.00 0.03 0.23 0.94 1.00 0.73 1.00 0.07 0.27 0.93 0.97 0.70 1.00 0.09 0.30 0.94 0.97 0.63 1.00 0.05 0.37 0.94 1.00 0.55 1.00 0.05 0.45 0.90 0.97 0.52 1.00 0.12 0.48 0.92 1.00 0.47 1.00 0.05 0.53 0.86 0.93 0.47 1.00 0.15 0.53 0.82 0.90 0.30 1.00 0.2 0.70 0.55 0.57 0.17 0.93 0.41 0.76 0.71 0.87 0.00 0.97 0.13 0.97 The HI, DI, HyI and CompI climate change signals (difference between the ensemble means for 2041-2070 and 1961-2000) are now presented, along with the corresponding NIQR (Fig. 2.4). The latter, is adopted as a measure of the model uncertainties. The significant increases in the HI values over most of Europe (left panel Fig. 2.4a) is associated with a relatively low uncertainty (right panel Fig. 2.4a), with the exception of northern Great-Britain and the Alps. Conversely, the changes in the DI pattern (left panel Fig. 2.4b) show large uncertainties, particularly in the regions where changes are most pronounced (the Mediterranean basin; right panel Fig. 2.4b). The HyI shows significant increases in eastern, northern and central Europe, while important decreases are found over southern and western Europe (left panel Fig. 2.4c). 36 Chapter 2. Future scenarios for viticultural zoning in Europe: ensemble projections and uncertainties Fig. 2.4 - Left Panel: Differences in the mean patterns (2041-2070 minus 1961-2000) of the (a) Huglin Index, (b) Dryness Index, (c) Hydrothermal Index and (d) Composite Index. Right Panel: Normalized Interquartile Range (third quartile minus first quartile divided by the mean at each grid point) showing variability in the 16-member ensemble for the same Indices as in left panel. Differences not statistically significant at the 99% confidence level are grey shaded (NS). For this pattern, low uncertainty levels are displayed all across Europe (right panel Fig. 2.4c). Contrary to the DI, this index combines temperature with precipitation and its lower uncertainty is thereby also influenced by temperature. The CompI undergoes significant 37 Chapter 2. Future scenarios for viticultural zoning in Europe: ensemble projections and uncertainties increases over large areas of northern, eastern and central Europe, whereas decreases can be found in some areas of southern Europe, mainly southern Iberia, southern Italy and Greece (left panel Fig. 2.4d). This index reveals high uncertainties in many eastern European regions (right panel Fig. 2.4d), where the strongest climate change signal is found, changing from recent-past climate conditions mostly unsuitable for viticulture to much higher suitability in the future. In these circumstances, slight differences in the climate conditions and/or projections tend to become very significant, in relative terms. 2.3.5. Categorization of CompI The CompI is now analyzed with respect to three relevant classes of the HI over Europe, namely: 900 ≤ HI < 1500ºC; 1500 ≤ HI < 2100ºC; 2100 ≤ HI < 3000ºC, as was already explained in section 2. Hence, the CompI is split into three nonoverlapping meaningful categories for viticultural zoning in Europe. Each year at a given location is thus keyed to one of these categories. The leading category at each location and for each selected time period is then detected and plotted. Only grid cells with at least 50% of optimal years according to the total CompI (values equal to or higher then 0.5) were categorized. For the baseline period (Fig. 2.5a), this pattern clearly highlights that approximately all potential winemaking regions northwards of the 45ºN parallel are Fig. 2.5 - Leading categories in the Composite Index (HI keyed to the first category (associated with classes 1, 2 and 3) for the baseline period (a) and the two future periods (b, c). 38 Chapter 2. Future scenarios for viticultural zoning in Europe: ensemble projections and uncertainties the HI first class, Fig. 2.5b). At lower latitudes, the most important exception is verified over the northern Iberian Peninsula, where large areas are also keyed to the first category; other minor exceptions occur over high altitude regions in southern Europe. The second and third categories are widespread over southern Europe (Equatorward of 45ºN), with low altitude (warmer) areas in the Mediterranean Basin depicting a preponderance of the third category, such as in southwestern Iberia, northern Italy (Po valley) and many coastal areas. For both future periods (Fig. 2.5b, c) there is strong evidence for new regions suitable for viticulture in different parts of northern Europe (e.g. southern British Isles, the Netherlands, Denmark, northern Germany and Poland). In fact, the projected changes in the number of suitable grid boxes (CompI greater than 0.5) for 2041-2070 shows and increase poleward of the 50ºN parallel (Fig. 2.6), though it gradually weakens towards 60ºN (upper latitude limit in the present study). Conversely, many regions in southern Europe (equatorward of 41ºN), such as southern Iberia, shift to unsuitable conditions, which can be largely explained by the lack of precipitation and dryness (Fig. 2.2b). Within the latitude belt of 41-50ºN there is no remarkable change in the number of suitable grid boxes, though there are some very significant shifts amongst the three categories. Fig. 2.6 - Latitudinal differences (2041-2070 minus 1961-2000) in the number of grid cells equal to or above 0.50 in the Composite Index. On the whole, the new suitable regions northwards of 50ºN are mainly keyed, as expected, to the first category (Figs. 2.5 and 2.6). Further, most regions from 46ºN up to 50ºN tend to change from the first to the second category, while many regions within 39-46ºN are projected to change from the first or second category to the third category. Therefore, the first 39 Chapter 2. Future scenarios for viticultural zoning in Europe: ensemble projections and uncertainties category (the least heat demanding) is projected to be essentially located northwards of 50ºN, as no new regions of this category are expected to arise at lower latitudes. The second and third categories undergo a northward displacement, leaving many current winemaking regions in southern Europe with unsuitable conditions, primarily because of the severe dryness that is expected to prevail in their future climates. 2.3.6. Inter-annual variability Due to the relevance of the inter-annual variability in viticultural zoning, corresponding climate change projections are also analyzed. As stated above, apart from some few exceptions, the selected bioclimatic indices are normally-distributed and their inter-annual variability can thus be assessed by their sample standard-deviation. The map of the ratios between the ensemble mean inter-annual standard-deviations of the HI for 2041-2070 and 1961-2000 reveals values significantly above one (higher variability) over large areas of Europe, particularly over central Europe, the British Isles and over some mountainous regions, such as the Alps (Fig. 2.7a). These results also depict low model uncertainties, with the exceptions of some regions of Iberia and some regions of central and northeastern Europe (Fig. 2.7b). The DI inter-annual variability (Fig. 2.7c) also shows an upward trend, especially in central and eastern European regions, with larger values over Great-Britain. The projected increases in precipitation over these regions, combined with increased inter-annual variability, may become harmful to viticulture, with enhanced risks of pests and diseases. These results also reveal low model uncertainty (Fig. 2.7d). In addition to the increase in the inter-annual variability of the HI and DI, the HyI shows significant increases in its inter-annual variability over northeastern Europe (not shown). As such, there is evidence for an overall increase in the inter-annual variability under future climates, which may also comprise an increase in the occurrence of extremes, though this analysis is leaved for a forthcoming study. 40 Chapter 2. Future scenarios for viticultural zoning in Europe: ensemble projections and uncertainties Fig. 2.7 - Ratio between the averages of the 16 inter-annual standard-deviations (calculated for each of the 16 ensemble members separately) in 2041-2070 and 1961-2000 of the (a) Huglin Index, (c) Dryness Index. Normalized Interquartile Range showing variability in the 16-member ensemble for the ratios of the (b) Huglin Index, (d) Dryness Index. No statistically significant ratios are plotted in grey shading (NS). 2.4. Summary and Discussion The projected and well-documented warming over Europe (cf. Christensen et al. 2007) leads to changes in the thermal indices (GSS and HI; Fig. 2.2a), with significant increases, particularly over southern and western Europe (over 400ºC increase in HI for 2041-2070). Further, dryness conditions in the growing season (GSP and DI; Fig. 2.2b) are projected to undergo an enhancement over southern Europe, while humidity levels (GSP and HyI; Fig. 2.2c) are expected to remain relatively high over central and northern Europe. All these changes are also corroborated by the CompI, not only by its aggregate values (Fig. 2.2d), but also in relation to the different classes of the HI considered for category analysis (Figs. 2.5-6). These projections are consistent with other recent studies by Neumann and Matzarakis (2011), for Germany, and by Duchene and Schneider (2005), for Alsace, France. Analogous 41 Chapter 2. Future scenarios for viticultural zoning in Europe: ensemble projections and uncertainties changes in the mean patterns of HI (Fig. 2.2a) have already been reported, including an increase of about 300 units for six winegrowing European regions over the last 30-50 years (Jones et al. 2005a). Additionally, under a future warmer climate, higher temperatures (above 30ºC) may often inhibit the formation of anthocyanin (Buttrose et al. 1971) and thus reduce grape color (Downey et al. 2006). All these changes may imply the selection of new winegrape varieties and a reshaping of the European viticulture. The analysis of the GSP and DI gives evidence for very dry climates over southern Europe (Fig. 2.2b). Although grape quality is generally favored by moderate water stress conditions during berry ripening (Koundouras et al. 1999; Santos et al. 2003), severe dryness (DI below -100 mm) in many Mediterranean-like climate regions may be damaging (Chaves et al. 2010); some of these regions are actually projected to have a GSP below 200 mm (southern Iberia, southern Italy, Turkey and Greece). Changes in winemaking practices are thereby expected to arise in these regions, including crop irrigation or water stress mitigation methods (Flexas et al. 2010). In the wetter regions in central and northern Europe, however, pests and diseases can also be a drawback to wine production, since changing climate conditions may modify the complex interrelationships between vine, pest and disease development (Stock et al. 2005). Regarding downy mildew, patterns of change in the HyI suggest at first sight, low risks of contamination in southern Europe, but moderate to high risk at higher latitudes the latter as a result of warm and wet conditions. Nonetheless, implications of a future increase in the HyI at the continental scale may not be too dramatic, mainly because in the emerging winemaking areas, where thermal conditions will gradually become favorable to wine production, the HyI will remain below the maximum threshold of 7500ºC mm in most years. Further, a projected strengthening of the inter-annual variability in the HI and DI may result in additional constraints for this perennial crop and for the winemaking sector as a whole, particularly in view of an increasing irregularity and unpredictability of yields and wine quality. A reshaping of the European regions suitable for grapevine growing is likely to occur taking into account the projections for the CompI. They show many regions throughout Europe undergoing a change to a higher CompI category. As an illustration, some regions in France and Germany will in the future present similar values to those seen today in the Mediterranean Basin. A general decrease in the number of suitable regions below 39ºN might also be expected. This outcome is also supported by previous findings (e.g. Kenny and 42 Chapter 2. Future scenarios for viticultural zoning in Europe: ensemble projections and uncertainties Harrison 1992; Jones et al. 2005b; Stock et al. 2005; Fraga et al. 2012b). Overall, shifts in the CompI classifications throughout Europe suggest modifications in the suitability of a given region to a specific winegrape variety and are expected to influence wine yield/quality attributes. The assessment of the uncertainty associated with the climate change projections is of great importance for informing decision-makers. Following the recommendations in Christensen et al. (2010) and Weigel et al. (2010) in our study, model uncertainty for viticultural zoning was assessed by considering all members of the ensembles on an equal weight. Although the ensemble means shown in the current study are in overall agreement with a previous study that employed a single model for similar purposes (Malheiro et al. 2010), thus generally corroborating the previous findings, future climate conditions can largely depend on the model experiment at local/regional scales (Dessai and Hulme 2007). In fact, at these scales, remarkable differences were found between the single-model projections and the ensemble projections presented here, which plainly substantiate the current study. The increase in the CO2 concentration is expected to be beneficial for winegrape growth (Moutinho-Pereira et al. 2004; Goncalves et al. 2009; Bindi et al. 1996). Reflecting the settings of the ENSEMBLES project (van der Linden and Mitchell 2009) in this study we only considered the A1B emission pathway. A more pronounced increase in atmospheric CO2 concentrations is prospected by the A2 emission scenarios, especially relatively to the second half of the century, with implications that need to be examined in future studies. The impacts of the aforementioned changes in climate suitability for the winemaking sector in Europe, can be summarized as follows: 1) some southern regions (e.g. Portugal, Spain and Italy) may face detrimental impacts owing to both severe dryness and unsuitably high temperatures; 2) regions in western and central Europe (e.g. southern Britain, northern France and Germany), some of which are world-renowned winemaking regions, might benefit from future climate conditions (higher suitability for grapevine growth and higher wine quality); and 3) new potential winegrape growth areas are expected to arise over northern and central Europe, where conditions are currently either marginally suitable or too cold for this crop. It is still worth mentioning that the higher inter-annual variability (climate irregularity) in the future over most of Europe may lead to additional threats to this sector. Furthermore, the shortening of the growing season resulting in earlier phenological events, reported by several studies (Webb et al. 2012; Bock et al. 2011; Daux et al. 2011; 43 Chapter 2. Future scenarios for viticultural zoning in Europe: ensemble projections and uncertainties Chuine et al. 2004), may indicate the need to adapt the time periods and classes in which the bioclimatic indices are commonly calculated. Since grapevine phenology and wine quality were shown to be correlated with HI (Jones et al. 2005a; Orlandini et al. 2005), this limitation can be, to some extent, overcome by adding new classes to the existing classical bioclimatic indices, instead of modifying the time periods (difficult to achieve due to the extent of the regions studied). A similar approach was indeed carried out by Santos et al. (2012) using a new low-limit class for the HI (900-1200ºC). Future changes in the viticultural zoning in Europe impose new challenges for the winemaking sector. These modifications give clues for the development of appropriate strategies to be taken by the winemaking sector to face climate change impacts. There have already been reports of acclaimed winemaking regions that may become unsuitable for premium wine production in the current century (Hall and Jones 2009). As the patterns discussed in this study are available at a relatively high spatial resolution (25-30 km), regional climate change assessments are permitted, enabling the development of local adaptation and mitigation measures, specifically drawn for the winemaking sector. Changes in soil and crop managements, as well as genetic breeding and oenological practices, are crucial for a better adjustment between viticulture and future environment. These measures need to be adequately and timely planned by stakeholders, policymakers, and by the different socioeconomic sectors that are directly or indirectly influenced by the vineyard and winemaking activities. 44 Chapter 3. Climate change impacts on the Portuguese viticulture from a multi-model ensemble Ciência e Técnica Vitivinícola, 2012, 27(1): 39 – 48 Impact Factor (5-yr): 0.479 Helder Fraga, João A. Santos, Aureliano C. Malheiro, José Moutinho-Pereira 45 Chapter 3. Climate change impacts on the Portuguese viticulture from a multi-model ensemble Abstract Portuguese vitiviniculture represents an extremely important economic activity for the agricultural sector, particularly for some renowned winemaking regions, such as Alentejo, Dão, Douro, and Minho. Viticultural zoning allows tying the suitability of a given grapevine variety to the local soil and climatic conditions. Given the existing climate variability in the Portuguese territory and its likely changes in the future, this zoning is thus of utmost interest. In this study, the current viticultural zoning in Portugal is discussed, as well as changes induced by climate change in the period 2011-2070. For this purpose, daily temperatures and precipitation rates were used to calculate the Huglin, cool night, dryness and hydrothermal indices. A composite index based on the previous indices was also calculated. For the assessment of the recent past conditions (1961-2000), an observational dataset (E-OBS) was used, while for future climate projections, a dataset comprising 16 simulations of regional climate models (produced by the ENSEMBLES project) was considered. In the future climate, statistically significant increases in the thermal indices are projected to occur in the next decades, while for the precipitation-based indices decreases might be expected, particularly over the south and innermost regions of Portugal. A reshaping of the main Portuguese winemaking regions is likely to occur in the upcoming decades, therefore emphasizing the need for the development of appropriate measures for the adaptation to or mitigation of these climatic changes at the level of varieties and rootstocks used, as well as at the implemented cultural practices, keeping the typicity and wine styles. 46 Chapter 3. Climate change impacts on the Portuguese viticulture from a multi-model ensemble 3.1. Introduction Climate is widely acknowledged as one of the major factors affecting vine physiology, phenology and wine parameters (Santos et al. 2011; Jones and Davis 2000a). In fact, the most worldwide renowned wine regions are located within relatively narrow latitude belts that provide very specific climatic conditions for high-quality wine production (Jones 2006; Spellman 1999). Although other factors, such as soils, winegrape varieties, agricultural and oenological practices might also play a crucial role on the entire winemaking process, climate and weather represent the most challenging factors (van Leeuwen et al. 2004), as they cannot be directly controlled by producers (we can only predict them and take measures to adapt and/or mitigate their effects) and vary significantly on relatively short time scales. Hence, viticultural zoning based on climatic factors has been applied as a first approach to delineate areas where climate is (nearly) optimal to winegrape growing, thus allowing the development of a sustainable winemaking sector, providing that the other factors are also reasonably suitable. Specialized bioclimatic indices can be used not only in evaluating the climatic suitability of a specified region to winegrape growth and wine production (Malheiro et al. 2010), but also in assessing some wine quality parameters, such as the balance between acidity and alcoholic content (Huglin 1978; Magalhães 2008). The Huglin index (HI; Huglin 1978) is a degree day index that also accounts for day length during the vine growing season. It is used for assessing the basic thermal and radiative demands of the grapevine so as to complete its phenological stages, including full and adequate grape maturation. In fact, the HI classes show strong correlations with some grapevine phenological events (Jones et al. 2005a), also linking a specific grapevine variety to a given climatic region. The Cool night index (Tonietto and Carbonneau 2004), another thermal index, accounts for night temperatures (minimum temperatures) during the maturation stage (September). Some studies argue that, at this stage, moderately low nocturnal temperatures combined with diurnal high temperatures tend to be advantageous for the production of high quality wines, promoting the synthesis of anthocyanins and other phenolic compounds (Kliewer and Torres 1972; Mori et al. 2005). Besides assessing the thermal conditions for grapevine development, it is also important to take into account the soil and atmospheric water conditions. The Hydrothermal index (HyI; Branas et al. 1946) combines the effect of air humidity (using precipitation) and temperature 47 Chapter 3. Climate change impacts on the Portuguese viticulture from a multi-model ensemble during the growing season to assess the risk of grapevine exposure to certain diseases, such as downy mildew. On the other hand, the Dryness index (DI; Riou et al. 1994) accounts for the soil water availability, thus providing information about the water stress conditions. The Composite index (CompI; Malheiro et al. 2010) is useful in depicting regions with suitable climatic conditions for winegrape growth by combining critical thresholds in the previous indices. In this context, by allowing the assessment of climate suitability for winegrape growth, the bioclimatic indices are a widely used tool in viticultural zoning. The awareness of a potential climate change is fundamental in order to raise adaptive capacity (Metzger et al. 2008). Therefore, the assessment of regional climate projections is of high pertinence for the wine industry by enabling the development of adequate measures for both mitigating their impacts and adapting to the new climatic conditions. Under the A1B International Panel on Climate Change (IPCC) – Synthesis Report on Emission Scenarios (SRES) scenario include a global temperature rise within the range 2.2-5.1ºC (Nakićenović et al. 2000). More specifically, climate in Portugal (typically Mediterranean, with temperature increasing and precipitation decreasing southwards and inwards) is expected to undergo some significant changes under anthropogenic forcing, including changes in temperature and precipitation (Meehl et al. 2007), as well as in their extremes (Costa et al. 2012). Vineyards in Portugal are virtually grown over almost all of the country (globally about 238.000 ha for aprox. 6 millions hl of wine production; IVV 2011), which is divided in large wine regions throughout the country (Fig. 3.1). Fig. 3.1 - The wine regions of mainland Portugal (IGPHowever, the most important Protected geographical indication). winemaking regions are localized within legally bounded controlled appellations (e.g. from north to south: Vinhos Verdes, Douro, Dão and Alentejo). These winemaking areas are quite diverse in their climates, geomorphological features, soil characteristics, and grown grapevine varieties (Magalhães (2008). As an illustration, while Alentejo (southeast) is mostly flatland 48 Chapter 3. Climate change impacts on the Portuguese viticulture from a multi-model ensemble with a relatively homogenous climate, the Douro Valley (northeast) is very mountainous and presents a large diversity of mesoclimates. Taking into account the important incomes the winemaking sector brings to the Portuguese economy, amounting nearly 2% of the total national exportation revenue (IVV 2011), the present study is devoted to the understanding of the climatic viticultural zoning in Portugal and the future implication of the climate change in this sector. The current climatic zoning and its likely changes under future climates are discussed using the aforementioned five bioclimatic indices. As such, this study aims at contributing to a better planning of the measures that need to be taken by producers (grapegrowers and winemakers), associations and organizations across the Portuguese wine industry in order to cope with climate change. 3.2. Methods Five bioclimatic indices, specifically 1) Huglin Index (HI), 2) Cool Night Index (CI), 3) Hydrothermic Index (HyI), 4) Dryness Index (DI) and 5) a Composite Index (CompI) were calculated over Portugal, using the mathematical definitions found in Table 3.1. For assessing climate change impacts, two periods were considered in this study. The baseline period representing current-past condition (1961-2000) was calculated using data from an observational dataset (E-OBS, version-5, Haylock et al. 2008), while for the future conditions (2041-2070) data from a 16-member ensemble, produced by the ENSEMBLES project (Table 3.2), was considered. The future period (2041-2070) was chosen to better characterize midcentury climatic conditions under the A1B IPCC-SRES scenario, a moderate anthropogenic radiative forcing scenario, but with already high emission levels (Nakićenović et al. 2000). Additionally, differences in the number of days with extreme temperatures (above 40ºC) and in the growing-season precipitations between future and current climates are also discussed. 49 Chapter 3. Climate change impacts on the Portuguese viticulture from a multi-model ensemble Table 3.1 - List of all the bioclimatic indices used in this study, their definitions and references. Bioclimatic Index Definition References a: HI ≥ 900ºC; Composite index b: DI ≥ -100 mm; Adapted from (CompI) c: HyI ≤ 7500ºC.mm; (Malheiro et al. 2010) d: Tmin always > -17ºC Cool Night Index (CI) September average Tmin (ºC) (Tonietto 1999) Sept. (Wo P Tv Es ) April Dryness index (DI) Huglin index (HI) (Riou et al. 1994) Wo - Initial available soil water reserve (mm); (Tonietto and Carbonneau 2004) P – Precipitation (mm); Tv - the potential transpiration in the vineyard (mm); Es - Direct evaporation from the soil (mm) Sept . (T 10) (Tmax 10) d 2 April T - Mean air temperature (ºC); (Huglin 1978) Tmax - Maximum air temperature (ºC); d - Length of day coefficient, ranging from 1.02 to 1.06 Aug . Hydrothermic index (HyI) (T * P) April (Branas et al. 1946) T - Mean air temperature (ºC); P – Precipitation (mm) All data fields, from each model run, were bilinearly interpolated from their original resolution (Table 3.2) onto a 0.25º × 0.25º grid, the same grid as in the observational dataset. This interpolation allowed applying a statistical error correction to the model-derived indices (model output statistics; MOS). The application of MOS to the calculated bioclimatic indices resulted in some losses of data along coastal areas of Portugal, where E-OBS grid-cells have missing data (cf. Fig. 3.2). Due to the low representativeness of these littoral areas to the national wine production, this is not an important shortcoming and no other statistical approach was thereby applied to estimate the indices in the blank cells (e.g. nearest neighbour with values). For the model calibration, since the calculated bioclimatic indices are normally distributed (according to the Lilliefors test; not shown), adjustments (transfer-functions) using multiple linear regressions were carried out. The same linear transformations were then applied to all future indices. This type of statistical model error correction has been used in previous studies (Alexandrov and Hoogenboom 2000) and can be used to provide reliable climate change scenarios (Jakob Themeßl et al. 2011). The use of a 16-member ensemble in the assessment of viticultural zoning in the future is an innovative methodology, by taking 50 Chapter 3. Climate change impacts on the Portuguese viticulture from a multi-model ensemble into account model uncertainties. Due to the large amount of outcomes, the results shown in this study are only focused on the 16-member ensemble mean patterns. Table 3.2 - Summary of all GCM / RCM model chains, original grid resolutions, institutions and references used in this study. GCM RCM Original grid Institution References 0.22ºx0.22º ARPEGE HIRHAM Danish Meteorological Institute (Christensen et al. 1996) rotated ARPEGE0.22ºx0.22º Centre National de Recherches (Gibelin and Deque Aladin RM5.1 rotated Météorologiques 2003) 0.22ºx0.22º Sweden's Meteorological and (Kjellström et al. 2005) BCM RCA rotated Hydrological Institute (Samuelsson et al. 2011) 0.20ºx0.20º (Böhm et al. 2006) ECHAM5-r1 CLM-1 Max Planck Institute regular (Steppeler et al. 2003) 0.20ºx0.20º (Böhm et al. 2006) ECHAM5-r2 CLM-2 Max Planck Institute regular (Steppeler et al. 2003) 0.22ºx0.22º ECHAM5-r3 HIRHAM Danish Meteorological Institute (Christensen et al. 1996) rotated 0.22ºx0.22º Koninklijk Nederlands ECHAM5-r3 RACMO (Lenderink et al. 2003) rotated Meteorologisch Instituut 0.22ºx0.22º Sweden's Meteorological and (Kjellström et al. 2005) ECHAM5-r3 RCA rotated Hydrological Institute (Samuelsson et al. 2011) 0.22ºx0.22º International Centre for (Elguindi et al. 2007) ECHAM5-r3 RegCM3 rotated Theoretical Physics (Pal et al. 2007) 0.22ºx0.22º (Jacob and Podzun 1997) ECHAM5-r3 REMO Max Planck Institute rotated (Jacob 2001) 0.22ºx0.22º (Kjellström et al. 2005) HadCM3Q16 RCA3 C4I Center rotated (Samuelsson et al. 2011) 0.22ºx0.22º Sweden's Meteorological and (Kjellström et al. 2005) HadCM3Q3 RCA rotated Hydrological Institute (Samuelsson et al. 2011) 0.22ºx0.22º Eidgenössische Technische (Steppeler et al. 2003) HadRM3Q0 CLM rotated Hochschule Zürich (Jaeger et al. 2008) 0.22ºx0.22º HadRM3Q0 HadRM3Q0 Hadley Centre (Collins et al. 2011) rotated 0.22ºx0.22º HadRM3Q16 HadRM3Q16 Hadley Centre (Collins et al. 2011) rotated 0.22ºx0.22º HadRM3Q3 HadRM3Q3 Hadley Centre (Collins et al. 2011) rotated In order to better analyse the spatial distribution of the resulting bioclimatic indices, an overlay of the current vineyard land cover was applied, using the Corine Land Cover Map (CLC 2000; EEA 2002; Büttner et al. 2006). This dataset provides an inventory of the land cover over Europe and has proven to have high accuracy in representing the cartography over mainland Portugal (Caetano et al. 2006). 51 Chapter 3. Climate change impacts on the Portuguese viticulture from a multi-model ensemble 3.3. Results The 16-member ensemble mean pattern of the HI (Fig. 3.2a), for the baseline period, shows relatively high values in the central and southern regions of Portugal (1800-3000ºC), while in the northern regions it reveals much lower values (900-1800ºC), which highlights the strong north-south contrast in the climatic conditions over Portugal. Similar results were reported by Magalhães (2008) using weather station data. For the future period (Fig. 3.2b), an overall increase in these values is expected (Fig. 3.2c), especially in the innermost regions, reaching values above 3000ºC (highest HI class). This pattern is in clear agreement with Malheiro et al. (2010) in a study for Europe and using a single model (COSMO-CLM). In fact, these increases will indeed lead to shifts to higher classes in the HI throughout Portugal. Fig. 3.2 - Huglin Index (HI; in ºC) for a) the baseline period (1961-2000), b) the future time period (2041-2070), and c) overall differences in 2041-2070 minus 1961-2000 (under the A1B IPCC-SRES scenario). Black areas represent the current vineyard land cover. This warming is also apparent in the night temperatures, as is suggested by the CI patterns. The CI (Fig. 3.3a) mean ensemble pattern shows a clear difference between south and coastal regions (warmer nights) and north and innermost regions (cooler nights). The CI future pattern shows a clearer distinction between southern (>18ºC), central (16-18ºC) and northern (12-16ºC) Portugal (Fig. 3.3b); increases of 2-4ºC are expected to occur, particularly over inland Portugal (Fig. 3.3c). This overall warming is also attested by an increase in the 52 Chapter 3. Climate change impacts on the Portuguese viticulture from a multi-model ensemble frequency of occurrence of extreme temperatures in the future. In fact, for some regions of the Douro Valley and Alentejo, projected changes include a significant rise in the number of days with maximum daily temperature above or equal to 40ºC (up to 8 days, Fig. 3.4a). Fig. 3.3 - Cool night Index (CI; in ºC) for a) the baseline period (1961-2000), b) the future time period (20412070), and c) overall differences in 2041-2070 minus 1961-2000 (under the A1B IPCC-SRES scenario). Black areas represent the current vineyard land cover. Fig. 3.4 - a) Differences in the number of days with maximum daily temperature equal to or above 40ºC (2041-2070 minus 1961-2000). b) Differences in the precipitation totals (in mm) during the growing season (2041-2070 minus 19612000). In addition to the overall warming in both the mean temperature and its extremes, a decrease in precipitation is also projected to occur, particularly in the northern and coastal 53 Chapter 3. Climate change impacts on the Portuguese viticulture from a multi-model ensemble Fig. 3.5 - Dryness Index (DI; in mm) for a) the baseline period (1961-2000), b) the future time period (20412070), and c) overall differences in 2041-2070 minus 1961-2000 (under the A1B IPCC-SRES scenario). Black areas represent the current vineyard land cover. areas (Fig. 3.4b). This drying leads to changes in the DI mean pattern (Fig. 3.5), where southern Portugal already shows moderate dryness under current climatic conditions (-100 to -50 mm). Changes in this index thereby suggest an important threat or challenge to the viticultural sector owing to the severe dryness that is likely to occur in the future (Fig. 3.5b), particularly in the innermost southern regions (Fig. 3.5c). This excessive dryness is in effect considered unsuitable for winegrape growth without irrigation (Koundouras et al. 1999). Conversely, the mean pattern of the HyI (Fig. 3.6a) for the baseline period shows low to moderate risk of downy mildew disease in most of the Portuguese mainland, and high risk in a small region in north-western Portugal (Alto Minho). The combined effect of the projected future warming and drying will yield a decrease in the HyI (Fig. 3.6b, c), leading to lower risks of crop contamination, which may have beneficial impacts on the sector. 54 Chapter 3. Climate change impacts on the Portuguese viticulture from a multi-model ensemble Fig. 3.6 - Hydrothermal Index (HyI; in ºC.mm) for a) the baseline period (1961-2000), b) the future time period (2041-2070), and c) overall differences in 2041-2070 minus 1961-2000 (under the A1B IPCC-SRES scenario). Black areas represent the current vineyard land cover. The CompI has proven to be effective in detecting the most suitable European regions for winegrape growth and wine production (Santos et al. 2012). For the baseline period, its mean pattern over Portugal highlights that most of Portugal has a very high suitability for viticulture (values above 0.99; Fig. 3.7a). However, in the future period, a tendency for lower values is depicted (Fig. 3.7b). In fact, the excessive dryness underlies this decrease (Fig. 3.7c) and may represent a detrimental impact on viticulture. Further, some regions in south-eastern Portugal (e.g. Alentejo) will have CompI values that suggest unsuitability for viticulture if mitigation measures (e.g. irrigation) are not implemented. Fig. 3.7 - Composite Index (CompI) for a) the baseline period (1961-2000), b) the future time period (20412070), and c) overall differences in 2041-2070 minus 1961-2000 (under the A1B IPCC-SRES scenario). Black areas represent the current vineyard land cover. Not statistically significant differences at the 99% confidence level are grey shaded. 55 Chapter 3. Climate change impacts on the Portuguese viticulture from a multi-model ensemble From Fig. 3.8 it is clear that over most of Portugal (36-41 ºN) a decrease in the area of suitable climate (CompI above 0.5) is expected to occur in the future, while over northwestern Portugal (polewards of 41ºN) small increases in suitability are projected instead. Fig. 3.8 - Latitudinal differences between the periods 2041-2070 and 1961-2000 in the number of grid cells equal to or above 0.5 in the composite index (CompI). 3.4. Summary and Conclusions Aiming at analyzing the climatic viticultural zoning in Portugal, five bioclimatic indices (HI, CI, HyI, DI and CompI) were computed and mapped over mainland Portugal. Their recent-past spatial patterns allow the isolation of the most suitable regions for winegrape growth and high-quality wine production. On the other hand, their changes under humandriven climate change (A1B SRES scenario) suggest a reshaping of the suitability throughout the country in the next few decades (until 2070), shedding some light onto the measures that can be adopted to adapt to or mitigate climate change impacts on the Portuguese wine industry. The HI pattern reveals a significant increase in the values of this index, particularly over the inland and southern areas of the country (Fig. 3.2). Increases in the HI have already been reported in other European countries, such as for Germany ((Neumann and Matzarakis (2011); Stock et al. (2005)) and for France (Duchene and Schneider (2005). This is indeed a clear manifestation of the expected warming under the A1B emission scenario, which is more accentuated in continental rather than coastal areas (Meehl et al. 2007; Knutti et al. 2008). Since HI values are largely tied to grapevine thermal demands, including strong correlations 56 Chapter 3. Climate change impacts on the Portuguese viticulture from a multi-model ensemble with some phenological events (Jones et al. 2005a), this shift may have important impacts on the Portuguese viticultural sector. A careful selection of the winegrape varieties to be grown at a given location in the future, based on the new HI classes, is certainly one important adaptation measure. The CI pattern clearly shows a warming of the nights in September, which may have important implications on the wine quality that must be taken into full consideration, particularly in southern Portugal (e.g. Alentejo), where average minimum temperatures are currently already above 18ºC (Fig. 3.3). This projected night-time warming is also supported by previous findings of Malheiro and Santos (2011); for the Iberian Peninsula and expected to yield altered wine typicity and quality (Bock et al. 2011). Additionally, the frequencies of occurrence of extremely high maximum temperatures (over 40ºC) will also increase (Fig. 3.4), which enhance thermal stresses and may eventually result in severe damages or even sunburns in leaves and berries. Grape harvest dates may also be affected by these extreme heat events, resulting from the high sensitivity of this crop to summer temperatures (Menzel et al. 2011; Chuine et al. 2004). Despite the changes in the thermal conditions, changes in rainfall patterns can also strongly impact on the winegrapes. In particular, the overall decrease in the growing season precipitation, mainly in the northern half of Portugal (e.g. Douro Wine Region, Vinhos Verdes Wine Region), may substantially increase water stress symptoms (Fig. 3.4). In effect, the lack of rain may require compensation measures through irrigation. This idea is supported by the DI pattern that displays values below the minimum threshold of -100 mm over southeastern Portugal (Fig. 3.5). Such decrease in the DI values can thus lead to great reductions in grapevine productivity, caused by damaging water stress (Moutinho-Pereira et al. 2004). In contrast, the decrease in the HyI values, accompanied by the lowering of the rainfall during the growing season, mainly in north-western Portugal (Vinhos Verdes controlled appellation), suggests a weakening of the risks of some pests and diseases in the vineyards, such as the downy mildew disease, among others (Fig. 3.6). The CompI summarizes the main changes in the previous indices and shows high agreement with the current distribution of the wine types in Portugal (Fig. 3.7 and Fig. 3.8). In the current climate, most of mainland Portugal tends to present optimal conditions for winegrape growth, whereas in the future large areas of south-eastern (e.g. Alentejo) will become less suitable. This outcome is supported by previous findings (Jones et al. 2005b; Stock et al. 2005; Malheiro et al. 2010). This climatic unsuitability is largely tied to the 57 Chapter 3. Climate change impacts on the Portuguese viticulture from a multi-model ensemble excessively dryness that can, however, be widely overcome by implementing both short and long term measures. Amongst the short term measures, it is worth highlighting agronomic practices such as soil (e.g. tillage), fertilization and irrigation management and use of chemical sunscreens for leaf protection. With respect to the long term measures, just to mention, changes in the altitude or solar exposure of the vineyards (vineyard microclimatic conditions), adjustments in the training systems, rootstocks and winegrape varieties of each region and genetic breeding of new varieties, necessarily less sensitive to water and thermal stresses, can be carried out. Furthermore, it is also important to emphasize the genetic pool given by a vast range (over 200) of autochthonous/indigenous varieties (Magalhães 2008), which can provide a key tool for climate change adaptation. These measures need to be thoroughly evaluated for each specific region, since some can be more easily implemented in southern Portugal (plain areas more prone to irrigation) and others in northern Portugal (e.g. changes in altitude and solar exposure). Altogether, these measures can effectively mitigate the potentially adverse impacts of climate change on the wine production sector in Portugal and can decisively contribute to its sustainable development in the next decades. Hence, the present study provides some insight into future strategies for the maintenance of a highly competitive wine industry despite climate change threats. 58 Chapter 4. Very high resolution bioclimatic zoning of Portuguese wine regions: present and future scenarios Regional Environmental Change, 2014, 14: 295-306 Impact Factor (5-yr): 2.082 Helder Fraga, Aureliano C. Malheiro, José Moutinho-Pereira, Gregory V. Jones, Fernando Alves, Joaquim G. Pinto, João A. Santos 59 Chapter 4. Very high resolution bioclimatic zoning of Portuguese wine regions: present and future scenarios Abstract Wine production is strongly affected by weather and climate and thus highly vulnerable to climate change. In Portugal, viticulture and wine production are an important economic activity. In the present study, current bioclimatic zoning in Portugal (1950-2000) and its projected changes under future climate conditions (2041-2070) are assessed through the analysis of an aggregated, categorized bioclimatic index (CatI) at a very high spatial resolution (near 1 km). CatI incorporates the most relevant bioclimatic characteristics of a given region, thus allowing the direct comparison between different regions. Future viticultural zoning is achieved using data from 13 climate model transient experiments following the A1B emission scenario. This data is downscaled using a two-step method of spatial pattern downscaling. This downscaling approach allows characterizing mesoclimatic influences on viticulture throughout Portugal. Results for the recent past depict the current spatial variability of Portuguese viticultural regions. Under future climate conditions, the current viticultural zoning is projected to undergo significant changes, which may represent important challenges for the Portuguese winemaking sector. The changes are quite robust across the different climate models. A lower bioclimatic diversity is also projected, resulting from a more homogeneous warm and dry climate in most of the wine regions. This will lead to changes in varietal suitability and wine characteristics of each region. 60 Chapter 4. Very high resolution bioclimatic zoning of Portuguese wine regions: present and future scenarios 4.1. Introduction The study of possible climatic impacts on natural and agricultural ecosystems is of major importance for a wide range of socioeconomic activities (IPCC 2001). More specifically, the vitivinicultural sector is strongly controlled by weather and climate and is thereby highly vulnerable to climate change (Fraga et al. 2012a; Kenny and Harrison 1992). In Portugal, this sector plays an important role in economic stability and growth. Its importance is evident with vineyards being grown over most of the country (Fig. 4.1a), covering an area of nearly 240 000 ha (3% of global vineyard surface area) and producing approximately 6×106 hl of wine (2% of world production) each year (OIV 2012). Roughly half of this production is exported, corresponding to more than €670×10 6 (OIV 2012) and representing approximately 2% of national export revenues (IVV 2011). In 2011, Portugal was the 11th/10th largest wine producer/exporter in the world (OIV 2012). Mainland Portugal is divided into 12 winemaking regions (Fig. 4.1b), many of which are world renowned such as the Douro/Porto, Minho and Alentejo. The Douro/Porto region, renowned for its Port Wine, is the oldest Demarcated Region in the world to be regulated by law and was designated World Heritage by UNESCO since 2001, which has obvious implications in tourism and other economic-stimulating activities. Viticultural microclimatic and mesoclimatic characteristics are key factors in determining varietal suitability and wine types of a given region (Carbonneau 2003; Jones 2006). Portuguese viticultural regions are in fact very distinct regarding their climatic characteristics (Fraga et al. 2012b). As an example, Alentejo (in southern Portugal) is mostly flatland (Figs. 4.1a, c), with a relatively homogenous warm and dry climate, while Douro/Porto (in northern Portugal) is a very mountainous region, characterized by the steep slopes of the Douro Valley. Furthermore, the Douro Valley contains a significant part of the country’s vineyard land cover (Fig. 4.1a) and comprises a large diversity of mesoclimates (Santos et al. 2013). Another climatically distinct viticultural region is Minho (in northwestern Portugal), which is characterized by relatively high annual precipitation (above 1200 mm) and relatively mild summers. These climatic features provide distinct viticultural characteristics in each winemaking region (Magalhães 2008). The combination of favorable climatic conditions, geomorphological aspects, experienced management decisions and numerous native winegrape varieties (e.g. Castelão, Touriga Nacional, Touriga Franca, Tinta Barroca) result in unique wine types (Jones 2012). 61 Chapter 4. Very high resolution bioclimatic zoning of Portuguese wine regions: present and future scenarios Spatially interpolated climate datasets have been a valuable tool to study recent-past climate and to assess possible climate change impacts, particularly in terms of environmental and agricultural changes (IPCC 2001). The use of high spatial resolution datasets is of significant importance for viticulture, since it is necessary to properly capture environmental variability within and between regions. This is particularly relevant for regions with steep climate gradients, where low resolution datasets cannot truly represent vineyard mesoclimates. Furthermore, as vineyards are usually dependent on slope, aspect and other yield influencing elements of the landscape, the complexity of the Portuguese topography (cf. Fig. 4.1c) is also an important factor controlling both development and spatial distribution of this crop throughout the country. The diverse climatic conditions and large spatial variability in which vineyards are grown in Portugal must be taken into account in order to produce a valuable viticultural zoning. Several previous studies used weather station data (Blanco-Ward et al. 2007; Jones and Goodrich 2008), other studies used Global Climate Model (GCM) data with spatial resolutions of about 100-250 km, and, more recently, Regional Climate Model (RCM) data with higher spatial resolutions of about 25 km (Fraga et al. 2012b; Fraga et al. 2013; Malheiro et al. 2010). Although these studies represent a valuable tool for climatic characterization of viticultural regions, the data used may not entirely reflect the vineyard climate. Therefore, the use of high-resolution data (1 km or less) has proven to be an important advance for regional/local scale assessments (Hall and Jones 2010; Anderson et al. 2012; Jones et al. 2010). Bioclimatic indices are useful metrics for assessing climatic influences on viticulture and thus can be used in viticultural zoning applications. One of the most widely used climatic indicators in viticultural zoning is the concept of degree-day for a base temperature of 10°C (Amerine and Winkler 1944). Using a degree-day approach, the Huglin Index (Huglin 1978) allows assessing the thermal potential of a given region and links the thermal demands for the ripening of each variety, reflecting the potential grape sugar content. The Dryness index (DI; Riou et al. 1994) evaluates soil water availability for vine development, by estimating soil water reserves, precipitation and potential evapotranspiration (potential water balance). This is also a good indicator for winegrape growth, as water availability can influence yield and quality (Hardie and Martin 2000). The Cool Night Index (CI; Tonietto and Carbonneau 2004) accounts for minimum temperatures during September, providing a coarse estimate of the ripening stage, as large temperature ranges during ripening tend to be favorable to highquality wines. Furthermore, by combining these bioclimatic aspects of heat accumulation 62 Chapter 4. Very high resolution bioclimatic zoning of Portuguese wine regions: present and future scenarios (HI), dryness (DI), and ripening conditions (CI) it is possible to determine optimum viticulture suitability and direct comparison between regions. The present study aims to improve the current knowledge concerning the bioclimatic conditions of the Portuguese viticultural regions, also focusing on future implications of climate change in the sustainability of this agribusiness. As such, the objectives of this study Fig. 4.1 - (a) Vineyard land cover over mainland Portugal using Corine Land Cover (EEA 2002). (b) Winemaking regions in Portugal (IVV 2011). (c) Elevation in meters using the GTOPO30 digital elevation model (http://eros.usgs.gov/#/Find_Data/Products_and_Data_Available/gtopo30_info), the red line differentiates the Atlantic/Mediterranean climatic zones. are three-fold: 1) to provide a high-resolution viticultural climatic zoning and comparison of bioclimatic indices over mainland Portugal; 2) to develop an aggregated categorized index that provides combined information regarding the bioclimatic characteristics of each region and; 3) to discuss the impacts of climate change on this zoning. 4.2. Materials and methods 4.2.1. Bioclimatic zoning A very high resolution climate dataset, produced by the WorldClim project (Hijmans et al. 2005), is used over the baseline period (1950-2000) for index computations. This gridded 63 Chapter 4. Very high resolution bioclimatic zoning of Portuguese wine regions: present and future scenarios dataset provides monthly climatic normals of precipitation, minimum, maximum and mean temperatures over a 0.008º × 0.008º latitude-longitude grid (~1 km). The above-mentioned variables are extracted over the Portuguese mainland (649 × 865 grid cells). Using the WorldClim data, three bioclimatic indices CI, DI and HI are calculated over mainland Portugal (cf. Table S2.1 in supplementary material). For validation purposes, differences between the indices calculated using this dataset and the E-OBS observational data (Haylock et al. 2008) are computed, revealing an overall good agreement (Fig. S2.1). Differences can be mostly explained by the predominant location of the E-OBS stations at low-elevations, leading to slightly warmer and drier conditions in E-OBS than in WorldClim. Although other bioclimatic indices are also calculated (Table S2.1), their spatial patterns provide overlapping information (Pearson correlation coefficients ≥ 0.95 with the previous three indices, Table S2.2), and are thus not presented here. To assess the impacts of climate change on future viticultural suitability in Portugal, the period 2041-2070 under the Intergovernmental Panel on Climate Change Special Report on Emission Scenarios A1B scenario is considered (Nakićenović et al. 2000). An ensemble of 13 RCM simulations driven by 3 different GCMs (Table S2.3) produced by the ENSEMBLES project (van der Linden and Mitchell 2009) is used for this purpose. 4.2.2. Spatial pattern downscaling As raw RCM datasets are available on a 0.22º × 0.22º latitude-longitude grid (~25 km), which is a coarse grid when compared to the baseline pattern (BP; <1 km), a two-step downscaling of the RCM-derived indices is undertaken. First, the climate change signal (CCS) for each index, i.e. the difference pattern between its 13-member ensemble means for 2041-2070 and 1951-2000 (‘deltas’ or anomalies), is bi-cubically interpolated for the same grid as in BP. Second, the interpolated climate change signal of each index is then added to BP, resulting in an index future pattern (FP) at a very high resolution. Therefore, the entire procedure for each index is as follows: CCS → interpolated CCS + BP = FP. This approach, also called pattern downscaling or ‘delta method’, relies on the assumption that the CCS presents high spatial correlations within a 25 km cell (RCM resolution). This is indeed a very reasonable assumption and does not present a significant limitation in the methodology. In fact, similar methodologies have been used in previous studies (Ramirez-Villegas and Jarvis 2010; Hay et al. 2000), but using GCM data (significantly lower resolutions than herein). 64 Chapter 4. Very high resolution bioclimatic zoning of Portuguese wine regions: present and future scenarios Besides the very high resolution of the downscaled fields, this approach has the important advantage of maintaining the original baseline dataset independent variables (longitude, latitude and elevation). Another important advantage of this methodology is that it does not require model output statistics, since the CCS itself is independent of these corrections. 4.2.3. Categorized index Lastly, an aggregated categorization is developed that integrates the CI, DI, and HI into a Categorized Index (CatI hereafter). This single, multipurpose index provides combined information regarding regional bioclimatic characteristics. The combination of classes is indeed similar to the multi-criteria climatic classification (MCC) by Tonietto and Carbonneau (2004). However, CatI includes only 17 different categories (Table 4.1), instead of 96 classes in the MCC, which simplifies mapping over large areas and allows an easier regional comparison. Also, the CatI includes limiting thresholds for viticultural suitability, namely DI > -100 mm and HI > 900°C (Fraga et al. 2013). Area-average minima, maxima and means of the CatI composing indices (HI, DI and CI) are also summarized for the 12 wine regions in Table S2.4. Table 4.1. Categories of CatI, along with the corresponding classes of the combined indices: Huglin, Dryness and Cool Night indices. A short description of the each category is also provided. Cool Night Index Huglin Index Dryness Index (°C units) Category (C° units) (mm) (Tonietto and Description (Huglin 1978) (Riou et al. 1994) Carbonneau 2004) 0 <900 <-100 Unsuitably cold or excessively dry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 < 14 > 14 < 14 > 14 < 14 > 14 < 14 > 14 < 14 > 14 < 14 > 14 < 14 > 14 < 14 > 14 -100 – 50 900 – 1500 > 50 -100 – 50 1500 – 2100 > 50 -100 – 50 2100 – 2700 > 50 -100 – 50 > 2700 > 50 65 Cool, dry with cool nights Cool, dry with warm nights Cool, humid with cool nights Cool, humid with warm nights Temperate, dry with cool nights Temperate, dry with warm nights Temperate, humid with cool nights Temperate, humid with warm nights Warm, dry with cool nights Warm, dry with warm nights Warm, humid with cool nights Warm, humid with warm nights Very warm, dry with cool nights Very warm, dry with warm nights Very warm, humid with cool nights Very warm, humid with warm nights Chapter 4. Very high resolution bioclimatic zoning of Portuguese wine regions: present and future scenarios 4.3. Results 4.3.1. Regional climatic zoning For the different wine regions in mainland Portugal, the patterns of CatI (Table 4.1) are depicted in Figs. 4.2-5 (cf. CI, DI and HI patterns in Figs. S2.2-S2.4). High correspondences between CatI and MCC are found for Portugal (Table S2.5), with the classes showing a Spearman rank correlation coefficient of 0.81 (the MCC nominal scale is converted into an equivalent ordinal scale for computing this correlation). For Minho (Fig. 4.2a,b), CatI is projected to undergo changes that are mainly from category 3 to 6 and from 7 or 8 to 10. This is the most humid wine region in Portugal under current conditions and will likely retain the maximum DI local value in the future (Table S2.4). The Trás-os-Montes region (Fig. 4.2c,d) is characterized by a strong bioclimatic heterogeneity, with warm and dry climates coexisting with cool and humid climates due to elevation differences over the region (Fig. 4.1c). In the future, contrary to the results seen for other regions, where a homogenization of the climate categories is projected, this region is expected to retain strong regional contrasts. In fact, some areas that are currently too cold are projected to become suitable (changes from category 0 to 1 or 3). Other areas will likely see changes from category 3 to 6, 6 to 14 and 7 to 9 or 10. It is also important to note that Trás-os-Montes exhibits the lowest minimum CI and HI in both present and future periods for the whole of Portugal (Table S2.4). The Douro/Porto region (Fig. 4.2e,f) presents a high diversity of bioclimatic regions due to the complex topography of the Douro Valley. In low elevation areas, categories 9 or 10 are projected to change to category 14, whereas categories 6 (in eastern areas) and 7 (in central and western areas) are projected to change to category 10. The Beira-Atlântico region (Fig. 4.3a,b) shows a large diversity of distinct bioclimatic categories, which can be largely explained by orographic forcing inland from the coast. Most of the region is currently in category 8, while more northern areas are category 7 (cooler nights) and some higher elevation areas are in category 3 (due to generally cooler climate). In the future, however, large areas of the region are projected to change to category 10, with the exception of some higher elevation and northern areas that will likely change to category 6. Currently, this region is the only showing category 12, combining a warm and humid climate with warm nights. In the future, this area will likely change to category 14, as a result of warming and increased dryness. Regarding Terras-do-Dão (Fig. 4.3c,d), shifts mainly occur 66 Chapter 4. Very high resolution bioclimatic zoning of Portuguese wine regions: present and future scenarios from categories 3, 7 and 8 to 6, 10 and 14. The small region of Terras-de-Cister (Fig. 4.3e,f) undergoes changes from categories 3 and 7 to 6 and 10. In the future, according to HI and DI, this region will likely become the coolest and least dry region (less affected by the overall drying trend) in the whole country (Table S2.4). Fig. 4.2 - CatI for the winemaking regions of (a, b) Minho, (c, d) Trás-os-Montes and (e, f) Douro/Porto for (left panel) 1950-2000 and (right panel) 2041-2070 under the A1B scenario. The current vineyard land cover is shown in dark stipple. See Table 4.1 for class structure for the CatI. 67 Chapter 4. Very high resolution bioclimatic zoning of Portuguese wine regions: present and future scenarios Fig. 4.3 - CatI for the winemaking regions of (a, b) Beira-Atlântico, (c, d) Terras-do-Dão and (e, f) Terras-deCister for (left panel) 1950-2000 and (right panel) 2041-2070 under the A1B scenario. The current vineyard land cover is shown in dark stipple. See Table 4.1 for class structure for the CatI. 68 Chapter 4. Very high resolution bioclimatic zoning of Portuguese wine regions: present and future scenarios For Terras-da-Beira (Fig. 4.4a,b), CatI clearly shows two distinct areas that either reflect the dominance of Atlantic (north) or Mediterranean (south) influences on climate. Southeastern areas are in category 10, with a warmer climate, while northern areas present temperate (CatI 7) or even cool (CatI 3) climates. Intermediate areas, between these climatically distinct zones, present CatI 6 or 8 (dry or humid temperate climates with warmer nights; Table 4.1). In the future, the Mediterranean area is projected to change to categories 14 or 0, being the last transition always due to excessive dryness. In the more Atlantic areas, changes are projected to be mostly to categories 6 and 10. The Lisboa region (Fig. 4.4c,d) is mostly characterized by categories 6 and 8 (western coastal areas), which is projected to change to category 10, though category 6 will likely remain in some coastal areas and category 14 is likely to be residual at some easternmost locations. For Tejo (Fig. 4.4e,f), changes are projected to be mostly from categories 6 and 10 to 10 and 14, respectively. For the Península-de-Setúbal (Fig. 4.5a,b) region, projected changes are expected to be similar to the Tejo region with categories 6 and 10 changing to 10 and 14, respectively. However it is worth noting that Península-de-Setúbal is expected to have the highest CI in the future (Table S2.4). For the Alentejo (Fig. 4.5c,d), CatI shows that most of the region is projected to transition from category 10 to 14, resulting in a future with a dry and very warm climate with warm nights, or even category 0, where viticulture is constrained, in this case due to excessive dryness (Fig. S2.3). In the future, Alentejo will not only likely experience the highest area-mean HI in Portugal, but will also keep the absolute nationwide maximum. Furthermore, the lowest DI value over Portugal is also projected to be found in this region (127 mm), where its area-mean value is near the suitability limit of -100 mm (Table S2.4). Although some areas will likely change from category 6 to 10, they currently show relatively low vineyard coverage. Lastly, in the Algarve wine region (Fig. 4.5e,f), similar future changes are visible, which is not totally surprising due to the proximity between these two regions. In fact, most of category 10 areas are projected to change to either category 14 or 0 (mainly in the Guadiana river basin, eastern Algarve), while category 6 areas (high elevation or northwestern Algarve) will likely change to category 10. Algarve, in both present conditions and future projections, shows the highest CI values in Portugal (Table S2.4). 69 Chapter 4. Very high resolution bioclimatic zoning of Portuguese wine regions: present and future scenarios Fig. 4.4 - CatI for the winemaking regions of (a, b) Terras-da-Beira, (c, d) Lisboa and (e, f) Tejo for (left panel) 1950-2000 and (right panel) 2041-2070 under the A1B scenario. The current vineyard land cover is shown in dark stipple. See Table 4.1 for class structure for the CatI. 70 Chapter 4. Very high resolution bioclimatic zoning of Portuguese wine regions: present and future scenarios Fig. 4.5 - CatI for the winemaking regions of (a, b) Península-de-Setúbal, (c, d) Alentejo and (e, f) Algarve for (left panel) 1950-2000 and (right panel) 2041-2070 under the A1B scenario. The current vineyard land cover is shown in dark stipple. See Table 4.1 for class structure for the CatI. 71 Chapter 4. Very high resolution bioclimatic zoning of Portuguese wine regions: present and future scenarios 4.3.2. Projected category transitions Overall, CatI transitions will mainly occur to higher categories (Fig. S2.5), while opposite changes will cumulatively account for less than 1% (transitions to category 0 not included). It is visible that most changes are from category 10 to 14 (33% of all grid boxes). This outcome is largely influenced by changes depicted in vast areas of southern Portugal. Changes from category 5, 6, 7 or 8 to 10 are also frequent (2%, 7%, 17% and 12%, respectively), especially in the coastal regions. Most of the changes to category 0 (with viticultural unsuitability) come from current category 10 (8%), mainly occurring in the innermost regions of southern Portugal, as other transitions to category 0 are residual. Category 3 areas will mostly change to categories 6 and 10 (5% and 3%) in the north. Moreover, remarkable climate anomalies, depicting changes from temperate to very warm climates, are visible from category 6 to 14 (3%). Conversely, positive CatI changes from category 0 to others, account for 1% of the total transitions over the Portuguese mainland, due to changes from very cool to cool or temperate climates in some high-elevation areas in the northern half of Portugal. 4.3.3. Ensemble variability The ensemble variability in the projected CatI transitions reveals low model spread for each region. Although other categories occur within each region, only the future dominant categories (with the highest spatial coverage) are considered here for the sake of succinctness (Table 4.2). The highest agreement between models is found for Lisboa, with all models projecting the same future category as in the ensemble mean (CatI 10; Fig 4.4d), whereas for Terras-de-Cister, only 7 from 13 models (53.8%) show the same dominant category as in the ensemble mean (CatI 10; Fig. 4.3f). 72 Chapter 4. Very high resolution bioclimatic zoning of Portuguese wine regions: present and future scenarios Table 4.2 - Wine regions with the corresponding dominant category (with the highest spatial coverage) in the ensemble-mean CatI for the future period (2041-2070; cf. Figs. 4.2-5). The respective single-model dominant categories are also listed, along with the percentages of occurrence across the 13-model ensemble (ensemblemean dominant category in bold). Ensemble-mean dominant category Single-model dominant categories Region (CatI) (CatI – %) Alentejo 14 14 – 76.9%; 0 – 23.1% Algarve 14 14 – 69.2%; 0 – 15.4%; 10 – 15.4% Beira-Atlântico 10 10 – 84.6%; 14 – 15.4% Douro/Porto 10 10 – 84.6%; 14 – 15.4% Lisboa 10 10 – 100% Minho 10 10 – 92.3%; 14 – 7.7% Península-de-Setúbal 14 14 – 92.3%; 0 – 7.7% Tejo 14 14 – 76.9%; 10 – 23.1% Terras-da-Beira 10 10 – 69.2%; 14 – 23.1%; 0 – 7.7% Terras-de-Cister 10 10 – 53.8%; 6 – 38.5%; 14 – 7.7% Terras-do-Dão 10 10 – 84.6%; 14 – 15.4% Trás-os-Montes 10 10 – 84.6%; 14 – 7.7%; 0 – 7.7% 4.4. Discussion and conclusions In the present study, a high-resolution current viticultural bioclimatic zoning over mainland Portugal is produced using an aggregated index that combines aspects of heat accumulation, dryness, and ripening conditions (CatI) allowing for direct comparison between regions. The new CatI also provides an easier and more practical integration with high resolution climatic dataset than other indices, such as the MCC, by using fewer classes more adapted to the Portuguese climate (17 instead of 96). The index computation is based on a very high resolution dataset (WorldClim Project) for a baseline period (1950-2000). Climate change projections are carried out by using a 13-member GCM/RCM ensemble of transient model experiments over 2041-2070 and following the A1B scenario. This data is downscaled to a near 1 km grid spacing, through a two-step method of spatial pattern downscaling, allowing the characterization of mesoclimatic influences in viticulture. Although a moderate emission scenario is selected, the variability in the climate change projections associated to 73 Chapter 4. Very high resolution bioclimatic zoning of Portuguese wine regions: present and future scenarios more/less severe scenarios are commonly of the same magnitude as the variability across a model ensemble (IPCC 2001). The viticultural zoning clearly depicts an Atlantic/Mediterranean climatic contrast over Portugal (roughly delineated by the red line Fig. 1c), highlighting not only significant spatial variability throughout the country, but also in each winemaking region. The current vineyard land cover in mainland Portugal takes advantage of the current bioclimatic diversity, explaining the relatively high number of wine regions and sub-regions with Denomination of Origin (Jones and Alves 2012b), specialized in the production of very specific wines made from a large number of grapevine varieties grown all across the country (Lopes et al. 2008). Current grapevine growing areas are not only determined by climate, but also by historic factors, soil and varietal characteristics, land-use limitations, and food politics (e.g. other crop priorities). Despite the importance of all these factors, the very high resolution maps of the bioclimatic indices provided herein represent an important tool for better understanding viticultural zoning in Portugal, especially if interpreted in conjunction with these factors. Under human-induced climate change, however, the current viticultural zoning is projected to undergo remarkable changes that can represent important challenges to the Portuguese winemaking sector. Future shifts in the CatI to higher classes are expected, which are in close agreement with recent and current warming reported in several European countries (Orlandini et al. 2009; Neumann and Matzarakis 2011; Duchene and Schneider 2005; Lorenzo et al. 2012). This may also lead to earlier phenological events throughout the country (Jones et al. 2005b; Webb et al. 2012; Jones and Davis 2000a). Future CatI spatial characteristics may imply a lower diversity of bioclimatic categories, resulting in a more homogeneous warm and dry climate throughout the country (mostly CatI 10 and 14), with the exception of Trás-os-Montes, where this warming will result in more heterogeneity. A more homogeneous climate undoubtedly will lead to changes in varietal selection and wine typicity of each region. In particular, in the more Mediterranean-like climatic regions (Alentejo, Algarve, Península-de-Setúbal, Tejo, southern half of Terras-da-Beira and innermost part of Douro/Porto), where CatI categories such as 10 characterize the current conditions, the projected increase in HI (warmer climates) may threaten less adapted cultivars. Moreover, in these regions, projected severe dryness (DI; Fig. S2.3) may lead to poor yields or even to unsuitable conditions for viticulture wherever irrigation is not feasible. In fact, the combined effect of increased warming and drying may cause damaging stresses to grapevines and 74 Chapter 4. Very high resolution bioclimatic zoning of Portuguese wine regions: present and future scenarios grapes at biochemical and physiological levels (Moutinho-Pereira et al. 2004; Chaves et al. 2010). Nocturnal warming and decreased diurnal thermal amplitude (CI; Fig. S2.2) in the last stages of development may also reduce wine quality (Tonietto and Carbonneau 2004). Although not covered by CatI, extreme high summertime temperatures may substantially increase water demands, which, when unmitigated, may cause physiological issues (e.g., leaves and berries sun burn, dehydration). Other weather extremes (e.g. hailstorms and late frosts) can also play a key role, but are not taken into account herein. All these changes can ultimately result in negative impacts on wine typicity and/or quality. On the other hand, northern and coastal regions of Portugal (Minho, Trás-os-Montes, Lisboa, Terras-do-Dão, Terras-de-Cister and Beira-Atlântico), which are more exposed to Atlantic air masses, where lower CatI categories are found, the results herein suggest some beneficial impacts of climate change. While projections for HI (Fig. S2.4) in these regions also give evidence for strong warming trends in the growing season, changes in these regions will likely enable changes to late maturing varieties, different wine styles, and potentially better quality. However, increased dryness in these regions, although neither excessive nor limiting for grapevine development, may have impacts on yield and quality regularity. To a certain extent, the overall drying in future climates can be advantageous, since lower humidity levels in vineyard microclimates, should reduce the risk of diseases (e.g. downy mildew) potentially requiring less chemical control measures in vineyards. The results from this study indicate that climate change adaptation measures should be focused on regions that are currently in category 10 (CatI; Fig. S2.5), as they are projected to change to category 14 or 0 (unsuitable conditions). The excessive dryness in these regions will likely require improved grapevine water use efficiency (Flexas et al. 2010). Furthermore, changes in some traditionally developed training systems may need to be applied (Pieri and Gaudillere 2003) to deal with lower water availability. In addition, proper selection of scion and rootstock will become more essential under increasingly warm and dry climate regimes (Harbertson and Keller 2012). Exploring irrigation management techniques, such as deficit irrigation strategies (Ferreira et al. 2012; Alves et al. 2012), should also be considered to provide an optimum yield/cost ratio. Solar radiation interception is also a major factor that can influence grapevine physiology. In the future, the optimization of canopy geometry and row orientation, or even employing shading materials, will likely be key measures that should be considered. Changes in soil management practices (e.g. cover cropping) should also be included in a broader adaptation strategy (Lopes et al. 2011). Furthermore, the beneficial 75 Chapter 4. Very high resolution bioclimatic zoning of Portuguese wine regions: present and future scenarios effect of the enhanced atmospheric CO2 concentrations in grapevine physiology may also play an important role in future viticulture (Moutinho-Pereira et al. 2009). In the future, warmer and dryer climates will challenge Portuguese viticulture, but may also represent an opportunity for a more sustainable sector. Although the current study focuses on a single future scenario (A1B), it provides robust and comprehensive projections using a multi-model ensemble. Previous studies have established the importance of viticultural zoning (e.g. Jones et al. 2010; Santos et al. 2012; Webb et al. 2011; Jones and Alves 2012a) and also emphasized the relevance of various bioclimatic factors in the geographical distribution of winegrapes, phenological developmental stages and varietal suitability. However, until the present, due to the lack of a high resolution climatic zoning in Portugal, direct comparisons between the winemaking regions have been based on large-scale assessments or weather station data, which may overestimate the climatic structure in a given wine region (Hall and Jones 2010; Jones and Alves 2012a). Although the country has a large number of legally authorized varieties (341; Veloso et al. 2010), most of them autochthonous, only a few studies have been devoted to their optimum suitability in some specific Portuguese wine regions (e.g. Lopes et al. 2008). By depicting bioclimatic conditions at a very high resolution for both current and future climates in Portugal, along with advances in understanding the optimal requirements of the traditionally grown or autochthonous winegrape varieties, a better understanding of adaptation strategies to climate change impacts can be fostered within this key sector of the Portuguese economy. 76 Chapter 5. Climate factors driving wine production in the Portuguese Minho region Agricultural and Forest Meteorology, 2014, 185: 26-36 Impact Factor (5-yr): 4.118 Helder Fraga, Aureliano C. Malheiro, José Moutinho-Pereira, João A. Santos 77 Chapter 5. Climate factors driving wine production in the Portuguese Minho region Abstract Establishing the role of climate on wine production is one major goal of the winemaking sector. Portuguese viticulture plays a key role in national exports of agro-food products. The Minho Wine Region, in particular, produces a unique wine type ‘Vinho Verde’ that has been taking its stand as a prominent international brand. The present study aims at improving the understanding of climate-yield relationships in this region. A long wine production series (1945-2010) is used and some transformations are undertaken for robust statistical relationships. A stepwise methodology is applied to select regressors for logistic modeling of production classes (low, normal and high). New weather regimes are developed to assess large-scale atmospheric forcing and cycles in production are isolated by a spectral analysis. Ten regressors are selected: Dryness and Hydrothermal indices, 3-yr lagged production, mean temperatures in March and June, precipitation in June and frequencies of occurrence of two regimes in May, and of one in February and September. Overall, moderate water stress during the growing season, high production 3-yrs before, cool weather in February-March, settledwarm weather in May, warm moist weather in June and relatively cool conditions preceding harvest are generally favorable to high wine production. Some of these relationships demonstrate the singularity of Minho Wine Region and justify the present study. The model shows high skill (72% after cross-validation), stressing not only the important role played by atmospheric conditions, but also its value for prediction and management. 78 Chapter 5. Climate factors driving wine production in the Portuguese Minho region 5.1. Introduction Wine production from Vitis vinifera L. is largely controlled by atmospheric forcings, since wine type, yield and quality are strongly dependent on weather conditions, mainly during the growing season (Malheiro et al. 2010; Jones and Davis 2000a). In fact, a 10ºC base temperature is needed for the growing season onset (Winkler 1974), which should be preferably preceded by a period of moderate cool weather, enabling an adequate bud dormancy (Webb et al. 2007). Vine phenophases (e.g. budburst, flowering and veraison) are triggered by adequate air temperatures, precipitation and solar exposures (Fraga et al. 2012a; Jones et al. 2005a). Furthermore, due to the very specific climatic demands of grapevines, suitable viticultural zones are restricted to relatively narrow areas where mean near-surface air temperatures range from 12 to 22°C during the growing season (Jones 2006), which makes this crop particularly suited for warm temperate climates. However, heat stress, with temperatures above 35ºC, can severely damage grapevine leaves and grapes (Crespy 1987; Chuine et al. 2004). Precipitation is also a central atmospheric element, as it widely governs soil moisture and grapevine water potential, particularly in vineyards without irrigation. Low water availability can lead to a wide range of effects, though largely dependent on the stage of plant development (Austin and Bondari 1988). For instance, severe water stress during the early stages may considerably delay growth and grapevine development (Hardie and Considine 1976). In contrast, high soil moisture throughout the growing season may cause excessive vigor, increased risks of pests and diseases and other problems related to wine quality and to the balance between its chemical compounds (During 1986; Magalhães 2008). Overall, highquality and well-balanced wines are commonly associated with mild water stress during the maturation period (from veraison onwards) and to relatively cool and moderately wet weather during the early stages (Koundouras et al. 1999; van Leeuwen et al. 2004; Storchi et al. 2005). Although these basic requirements are dependent on vine variety, they are fulfilled by Mediterranean-like rainfall regimes. A favorable and regular climate is also crucial to stabilize yields, while high interannual variability in atmospheric conditions can strongly influence winegrape quality (Jones and Goodrich 2008). Furthermore, extreme weather events, such as hail and frost during the growing season, can have harmful impacts on vineyards, thus contributing to noteworthy fluctuations in yield and wine quality. Taking into account all the aforementioned climatic 79 Chapter 5. Climate factors driving wine production in the Portuguese Minho region influences in viticultural production, further knowledge into these relationships is of foremost relevance for modern wine industry, which aims at stabilizing wine production (WP) and quality for brand marketing. Apart from the direct influence of near-surface temperatures and precipitation, other factors, such as weather regimes may also play a key role in yield prediction, as they combine several elements in their typical atmospheric conditions (e.g. temperature, precipitation, wind, humidity and solar radiation). Furthermore, viticultural bioclimatic indices are often used when assessing climatic suitability of a given region to winegrape zoning (Jones et al. 2010). Well-known bioclimatic indices accounting for water availability, such as the Hydrothermal Index (HyI; Branas et al. 1946) and the Dryness Index (DI; Riou et al. 1994), should also be tested as potential predictors of grapevine yield. Other factors, such as variability in wine production resulting from commercial cycles, economic fluctuations or even physiological constraints, may result in regular oscillations in the annual yield series. This information may also be relevant for WP modeling. More specifically, Portugal (the 11th wine producer in the world; OIV 2012) is divided into several wine growing areas, with the Minho Wine Region (MWR henceforth) located in the very northwestern part of the country (Fig. 5.1). MWR covers a broad area of about 8.8×103 km2 that geographically coincides with the ‘Vinhos Verdes’ Appellation of Origin. Regarding its topography, MWR has a rather irregular terrain, with elevations ranging from 0 to approximately 1500 m asl, characterized by a compact valley system with a predominance of deep granitic soils (Magalhães 2008). Despite having a Mediterranean climate, the Atlantic influences are noteworthy. The relatively high exposure to maritime winds, high annual precipitations (1200-2400 mm) and mild summers (summer mean temperatures ranging from 18 to 22ºC), explains the distinctiveness of the MWR, setting it apart from other Portuguese wine regions (e.g. Alentejo, Douro), where heat and water stresses can be important limiting factors of WP. In fact, Fraga et al. (2012b) showed that the MWR depicts the highest HyI and DI values of all Portuguese winemaking regions. Another distinctive characteristic of the MWR is its relatively low yearly solar insolation when compared to other Portuguese regions (< 2500 hours). In response to these climatic features, traditional training systems, such as pergola, stakes and tall growing vines (‘enforcado’), are still used today, which are rare in the rest of the country. Most wines are usually made from earlier maturity varieties (e.g. Alvarinho, Alvarelhão, Trajadura), with a predominance of white varieties, as a response to moderate humidity levels in summer, which enhance the risk of trunk and leaf diseases (IVV 80 Chapter 5. Climate factors driving wine production in the Portuguese Minho region 2011). All these characteristics explain the typicity of its famous ‘Vinho Verde’ wine, which is largely consumed, representing the country’s second most exported wine, right after Port Wine (IVV 2011). Fig. 5.1 - Right panel: map showing the geographical location of the Minho Wine Region (MWR) in northwestern Portugal, along with the other Portuguese wine regions. Left panel: magnified map of the MWR. Grid boxes represent the E-OBS original resolution of 0.25º latitude × 0.25º longitude. Highlighted grid boxes show where data was extracted for area-mean computations. The main current vineyard land cover is also shown in black stipple. Previous studies have revealed that monthly temperatures and precipitations are significantly correlated with grapevine yield in the nearby Douro Wine Region, Portugal (Gouveia et al. 2011; Santos et al. 2013). However, due to the aforementioned climatic specificities of MWR, this climate-yield linkage can significantly differ from other Portuguese regions and needs to be studied separately. The present study aims to examine the influence of climate on WP of MWR. So this study addresses the effects of climate on the WP of this distinct region. The objectives are three-fold: 1) Quantify the cycles in wine production (WP) in the region over the last 60 years. 2) Differentiate the climate of the region into a set of distinct weather regimes, and determine their relationship to WP. 3) Develop a multinomial logistic model using weather data, previous WP and weather regimes to predict classes of WP values. 81 Chapter 5. Climate factors driving wine production in the Portuguese Minho region 5.2. Materials and methods 5.2.1. Wine production data A relatively long time series of WP (in hl) in MWR, from 1945 to 2010 (66 years), is provided by the Portuguese ‘Instituto da Vinha e do Vinho’. This data is annually reported by producers directly to this governmental institution. The current vineyard land coverage is of about 22 × 103 ha (Fig. 5.1), estimated using the Corine Land Cover (EEA 2002), which has previously shown a good accuracy over Portugal (Caetano et al. 2006). This vector dataset is originally available on a 1:100 000 scale, but for display proposes it is aggregated within a 1 km range. Although significant changes in the total vineyard area might have occurred during the study period, no reliable records are available with this information. 5.2.2. Climatic data Four climatic elements used in this study are extracted from the E-OBS observational interpolated/gridded dataset, version 7 (Haylock et al. 2008). Despite some already documented limitations (Hofstra et al. 2009), this dataset provides uninterrupted and homogeneous gridded fields of daily minimum (Tmin), maximum (Tmax) and mean (Tavr) temperatures and daily precipitation (P) over Europe from 1950 onwards, on a ~ 27 km spatial resolution. Since no climatic data is available before 1950, all computations are carried out in the overlapping period between climate and WP data (1950-2010). Climatic data are extracted for 9 grid-boxes within the MWR, which jointly enclose most of the vineyard area (Fig. 5.1). The monthly means of Tmin, Tmax and Tavr and the monthly totals of P are computed and spatially averaged over the 9 grid-boxes. As the WP series reflect the entire region, these areameans better represent the atmospheric conditions of the region as a whole. 5.2.3. Production cycles In order to verify the existence of well-defined cycles in the WP series, a spectral analysis is implemented using the Burg algorithm (Burg 1975). It has been successfully applied in many different areas (e.g. Santos and Leite 2009). This parametric spectral estimation method (Stoica and Moses 1997) is applied to the 2nd order stationary and normally 82 Chapter 5. Climate factors driving wine production in the Portuguese Minho region distributed time series of percentile-transformed wine productions, as will be further explained below. The Haykin criteria (Haykin 1991) is applied to determine the optimal order of the autoregressive filter embedded in Burg’s estimation, which is of 19 for the current WP series. The outcomes of this method are then validated by the Fourier transform method, revealing very similar results (not shown). 5.2.4. Weather regimes Aiming at establishing statistical relationships between large-scale atmospheric circulation and WP, the daily atmospheric circulation patterns over a North Atlantic-Iberian sector are classified into a set of weather regimes. For this purpose, daily mean sea level pressure fields in 1950-2010 from the NCEP-NCAR reanalysis are used (Kistler et al. 2001). These fields have previously proven their effectiveness in isolating the large-scale weather regimes over Portugal (Trigo and DaCamara 2000; Santos et al. 2005a) and, for this purpose, are preferred over other atmospheric fields, such as geopotential height. In fact, the sea level pressure fields better reflect surface features e.g. driven by strong land-sea thermal contracts that are indeed important when identifying weather regimes in the Iberian Peninsula. Daily pressure anomalies at each grid point are computed with respect to the corresponding longterm mean pressures for each calendar day. A K-means clustering is applied to the leading 20 principal components of the daily mean sea level pressure anomalies within a Euro-Atlantic sector (30ºW-10ºE; 25-65ºN). It not only covers Western Europe, but also the Eastern North Atlantic, which is known to play a key role in the weather conditions in Europe and more specifically in Portugal (Santos and Corte-Real 2006; Santos et al. 2007). As stated before, this is particularly patent in the MWR, as it is the Portuguese region most exposed to the Atlantic influences, owing to its location in the northwestern corner of Portugal. The 20 retained components cumulatively explain 98% of the total temporal variance, but substantially reduce the number of input variables for clustering, thus increasing the signal-to-noise ratio and the robustness of the results (Wilks 2011). The choice of the number of weather regimes (clusters) is always subjective. However, here it is (1) based on previous studies (Santos et al. 2005a), (2) targeted to evenly distribute days across different regimes and (3) seeks parsimony between the number of regimes and the discrimination of different atmospheric flows. Six weather regimes are chosen to represent the 83 Chapter 5. Climate factors driving wine production in the Portuguese Minho region main features of large-scale atmospheric circulation in Eastern North Atlantic. The clustering is based on Euclidean distances from each day to the six centroids. A preliminary clustering, using a random 10% subsample collected from the 20 input components, is performed to define initial centroids (seeding). This process is repeated 10 times so as to identify the best solution that minimizes total within-cluster variance. Lastly, the resulting calendar (day vs. regime) is used to compute frequencies of occurrence of weather regimes on a monthly basis and over the full period (1950-2010). 5.2.5. Bioclimatic indices The HyI considers precipitation and temperature as follows (Branas et al. 1946): 31 Aug (T avr P) , 1 Apr where Tavr is daily mean temperature (ºC) and P daily precipitation (mm). This index estimates the risk of downy mildew disease, which is a common limiting factor for grapevine yield (Carbonneau 2003). This risk is usually considered low when HyI values are below 2500ºC mm, high for values higher than 5100ºC mm and very high for values higher than 7500ºC mm (Santos et al. 2012). The DI allows the description of available soil water content for the vine in the growing season, taking into account precipitation and reference evapotranspiration (Riou et al. 1994), as described below: 30Sep (W 0 P Tv Es ) , 1 Apr where W0 is soil water reserve, which should be equal to 200 mm in April and takes the previous month DI value in the following months. P is precipitation; Tv is potential transpiration and Es is soil evaporation (all variables in mm). Daily variables are used for this calculation and four climate classes can be distinguished (defined in Tonietto and Carbonneau 2004): Very dry, where viticulture is limited by severe dryness (DI ≤ −100 mm), Moderately dry (−100 < DI ≤ 50 mm), Sub-humid (50 < DI ≤ 150 mm) and Humid (DI >150 mm). For 84 Chapter 5. Climate factors driving wine production in the Portuguese Minho region the inclusion of HyI and DI in statistical modeling, their yearly values are averaged over the 9 grid-boxes (Fig. 5.1) and using the E-OBS dataset. 5.2.6. Logistic model A multinomial logistic modeling of WP classes is developed in the present study (Wilks 2011). One important advantage of logistic over linear regression is that it delivers probabilities of occurrence of predefined classes instead of productions. Logistic modeling has been used in many fields, including agricultural research (Gobin et al. 2001; Meyer and Mart ne -Casasnovas 1999), and is frequently preferred over linear or polynomial models (Overman et al. 2003). All years are then classified in three probability classes: Class 1 - low production years (below the 25th percentile); Class 2 - normal years (between the 25th and 75th percentiles); and Class 3 - high production years (above 75th percentile). Their probabilities of occurrence are thereby modeled by logistic regression using a number of regressors. Each year is keyed to the highest probability class and the model skill ratio corresponds to the percentage of correct matches between modeled and observed WP classes. Owing to the high number of potential regressors, a stepwise multivariate regression is used for a preliminary selection (Wilks 2011). The stepwise criteria for forward selection and backward elimination correspond to p-levels of 0.05 and 0.10, respectively. The initial pool of predictors comprises monthly Tmin, Tmax, Tavr and P, yearly HyI and DI, monthly frequencies of occurrence of the six weather regimes and time lagged productions. As shown below, a percentile-transformed WP series is used in the stepwise process instead of the original time series, as this transformed series better fulfills two basic assumptions: weak stationarity and normally distributed. 5.3. Results and discussion 5.3.1. Wine production series Although WP shows a linear decrease of –26.4×103 hl yr-1, an 11-year moving average suggests that a 2nd order polynomial can be more skillfully adjusted to the time series than the linear trend (Fig. 5.2a). Therefore, prior to further statistical analysis, the best-fit least squares 85 Chapter 5. Climate factors driving wine production in the Portuguese Minho region polynomial is subtracted from the raw series, thus removing most of the long-term trends and oscillations. Contrary to the original WP series, the transformed time series (Fig. 5.2b) is now nearly stationary in mean (no linear trend at a 5% significance level) and normally distributed (according to the Lilliefors test at a 5% significance level). However, this transformed time series still presents some non-stationarity in variance (e.g. higher variance in 1960-1980 than in 1980-2010). Therefore, in order to overcome this remaining heterogeneity, the percentiletransformed series is used (i.e. productions replaced by their corresponding percentiles), which is also 2nd order stationary and normally distributed (Fig. 5.2c). The underlying reasons for these heterogeneities are rather difficult to address, as many factors can have interfered (e.g. changes in vineyard area, agricultural practices and human-driven policies). Nevertheless, they are mostly long-term changes that are out of the scope of the present study, as their attribution is not possible with the available data. Moreover, they do not account for the bulk of year-to-year variability to be modeled herein, which is still well-captured by the percentile-transformed series. 86 Chapter 5. Climate factors driving wine production in the Portuguese Minho region Fig. 5.2 - (a) On the left: chronogram of the wine production in the MWR (in 10 3 hl) in 1950-2010 (grey line), along with the corresponding best-fit second order polynomial and the 11-yr moving average. On the right: corresponding box-plot, where the median is the horizontal line within the box, the 75th and 25th percentiles are the upper and lower box limits, and the whiskers limits are the maximum and minimum. (b, c) The same as in (a), but for the transformed wine production series (b) and for the percentiles of the transformed series (c). The horizontal dashed lines correspond to the 25th and 75th percentiles, the thresholds used in the class definitions for logistic regression. 87 Chapter 5. Climate factors driving wine production in the Portuguese Minho region 5.3.2. Candidate regressors As previously stated, monthly Tmin, Tmax, Tavr and P are strong candidate regressors for WP modeling. In fact, some of them have been used in wine production/yield modeling for other nearby wine regions, such as the Douro Valley, Portugal (Santos et al. 2013; Santos et al. 2011; Gouveia et al. 2011) and Rías Baixas (in northwestern Spain; Lorenzo et al. 2012). However, as will be shown below, these regressors alone are not sufficient to reach a reasonable skill in modeling WP. The spectrum of the percentile-transformed WP series hints at the presence of 2.5 and 3yr cycles (Fig. 5.3). A 5-yr cycle is also apparent and no further cycles are visible series after 10-yrs (not-shown). The existence of these cycles is a remarkable result that can be due to several causes. One possible explanation is that yields respond to grapevine vegetative and reproductive cycles, the later extending over 2 years. Furthermore, production limitations and commercial policies, enforced by wine industry decision-makers, cannot be disregarded as driving Fig. 5.3 - Normalized power spectral density (dimensionless) of the percentile-transformed wine production series in the MWR for 1950-2010. factors of these cycles. Nevertheless, this kind of assessment cannot be properly addressed using the available data. Thus, lagged production series (1-5 yrs. lags) are tested as potential predictors of WP classes. For these lagged time series, productions in 1945-1949 are also used. The HyI shows some noteworthy yearly fluctuations, spanning from about 2000 up to 8000ºC mm, but reveals no statistically significant trend (Fig. 5.4a). In fact, when analyzing the spectral density of this index a 5-yr cycle is depicted (not-shown). This suggests that the presence of this cycle in the WP series can be attributed to climatic factors. Although this periodicity has been detected in many time series of climatic parameters and in teleconnection patterns, such as in the El-Niño Southern Oscillation (Latif et al. 1998), an attribution analysis is out of the scope of the present study. Although temperature increases have been recorded in the MWR (linear trend of 0.16 ºC yr-1 over 1950-2010), they tend to be offset by decreasing 88 Chapter 5. Climate factors driving wine production in the Portuguese Minho region Fig. 5.4 - (a) Chronogram of the HyI from 1950 to 2010, using a 9 grid-box area-mean over the MWR and the respective linear trend. (b) The same as in (a), but for the DI. precipitation (linear trend of -6.16 mm yr-1 over the same period). Further, the strong interannual variability in HyI highlights the irregularity of the climatic conditions in the MWR, with potentially detrimental impacts on yields. Although a drying trend is clearly shown in DI (Fig. 5.4b), with a downward trend of -0.59 mm yr-1, severe dryness is still not a major concern in the MWR (DI generally above -100 mm). As such, despite the increased dryness over the last six decades, the downy mildew risk has remained largely unchanged in the MWR. 89 Chapter 5. Climate factors driving wine production in the Portuguese Minho region 5.3.3. Weather regimes Previous studies suggest that the inter-annual variability in WP is considerably controlled by large-scale atmospheric circulation (Jones and Davis 2000b; Lorenzo et al. 2012). Different regimes generate different weather conditions that are not entirely captured by temperature or precipitation anomalies, as their associated conditions are also manifested by other parameters, such as relative humidity, solar radiation and wind. In fact, some of these weather conditions can influence vineyard microclimates, e.g. different wind speeds can trigger different stomatal responses (Brewer 1992). Thus, each regime combines several atmospheric parameters that can jointly have a much stronger impact on vineyards than when analyzed separately, thus justifying their inclusion as candidate regressors. Hence, the monthly frequencies of occurrence of six weather regimes (AA, E, NW, C, R and A) are also considered as potential regressors for production modeling. The composites of mean sea level pressure for each regime indicate prevailing weather conditions (Fig. 5.5). Wind bearing and speed can be roughly estimated by the geostrophic wind balance, in which wind is approximately parallel to the isobars, leaving the highest pressures to the right (Holton 2004). However, it is worth noting that at near-surface levels the wind vector can be significantly diverted due to friction. The designation of some regimes is related to the predominant wind direction over Portugal. The AA (dual anticyclonic) regime is characterized by two high pressure systems, one corresponding to the Azores high and another located over Central Europe. A strong low pressure system is located at higher latitudes over the North Atlantic (Fig. 5.5a). The E (easterly wind) regime features a strong high pressure system over the British Isles, driving easterly winds over Iberia (Fig. 5.5b). The NW (northwesterly wind) regime is characterized by a near-average pressure field over the North Atlantic, which may lead to important westerly flows over Portugal (Fig. 5.5c). The C (cyclonic) regime presents a strong low pressure system northwestwards of Iberia, inducing westerly-southwesterly winds over Portugal (Fig. 5.5d). The R (ridge) regime exhibits a strong ridge over the Eastern North Atlantic, producing strong northerly-northwesterly winds over Portugal (Fig. 5.5e). Finally, the A (anticyclonic) regime reflects a positive North Atlantic Oscillation phase, with a strong mid-latitude pressure system extending from the North Atlantic towards Iberia (Fig. 5.5f). These regimes are in clear agreement with other regimes developed for the same region, e.g. by Plaut and Simonnet (2001) (cf. their Figure 4). The most noteworthy difference is the split of their AR regime into the present A and R 90 Chapter 5. Climate factors driving wine production in the Portuguese Minho region regimes. As will be shown below, this distinction is advantageous for the purposes of this study, since the latter two regimes induce different mesoscale conditions in the MWR, with different impacts on production. Fig. 5.5 - (a-f) Composites of the daily mean sea level pressure fields (in hPa) over the Euro-Atlantic sector (25-65ºN; 60ºW-20ºE) for the six outlined weather regimes (AA, E, NW, C, R and A) and for all days in 1950-2010. The interplay between these regimes and the associated weather conditions in Portugal is widely dependent on month/season (strong seasonality). When analyzing the monthly frequencies of occurrence of each weather regime in 1950-2000 (Fig. 5.6), it is clear that A and E dominate wintertime months, while the frequency of R tends to gradually increase in the warmer half of the year, surpassing 50% in summer months (June-August). In these 91 Chapter 5. Climate factors driving wine production in the Portuguese Minho region months, A, C and NW have very low frequencies of occurrence. In the last months of the year, as the frequency of R decreases, A, C and NW have higher occurrences. The AA regime presents a very uniform distribution throughout the year, but with frequencies peaking in September-October. For the sake of succinctness, the study will be focused only on the relevant regimes for WP in the MWR. Fig. 5.6 - Relative monthly frequencies of the occurrence (in %) of each weather regime for the period 1950-2010 and across the year. 5.3.4. Significant regressors Ten predictors are selected by the stepwise process: HyI, DI, 3-yr lagged WP (Prod-3), Tavr in March (Tavr_Mar) and June (Tavr_Jun), P in June (PJun), frequencies of occurrence of the A regime in February (AFeb), May (AMay) and September (ASep), and frequencies of occurrence of R in May (RMay). The leading regressor, HyI, presents a negative regression coefficient (b = −0.005), which means that it negatively influences WP. In fact, anomalously high HyI values may represent high risks of downy mildew attack in vineyards. Significant correlations between WP and this index were also found by Lorenzo et al. (2012). The second leading regressor is Tavr_Jun, now with a positive regression coefficient (b = +10.164). As June roughly 92 Chapter 5. Climate factors driving wine production in the Portuguese Minho region corresponds to the first phase of berry growth in MWR, moderately high temperatures are considered favorable to production (Nemani et al. 2001; Jones and Davis 2000a). The third regressor is Prod-3, showing a positive regression coefficient (b = +0.012), i.e. production in a given year is positively correlated with production 3 years before. The fourth selected regressor is DI. Since this index is usually a measure of water stress in vineyards, its negative correlation coefficient (b = −0.289) means that elevated soil moisture during the growing season may be harmful for WP. This is in agreement with HyI, since high precipitation levels tend to bring higher risks of diseases and physiological problems. On the other hand, moderate dryness found in some years (Fig. 5.4b), which tends to be beneficial for quality, may reinforce this relationship. As referred above, excessive precipitation levels during the growing season tend to be unfavorable to WP in the MWR. However, PJun (fifth selected regressor) presents a positive regression coefficient (b = +0.063), which means that anomalously high precipitation in this month tends to be beneficial for WP. This is in agreement with Lorenzo et al. (2012), where rainfall days are usually correlated with higher production during bloom-veraison. The seeming contradiction between the signals of the regression coefficients of HyI or DI and P Jun can be explained by the different time spans of the whole growing season (used in HyI and DI calculations) and of June. Other studies also found a negative impact in grapevine yield when an increase in evapotranspiration was combined with a decrease in accumulated precipitation from bloom to veraison (Ramos et al. 2008; Camps and Ramos 2012). Anomalously high temperatures and precipitations in June may in fact promote berry growth through cell division and cell enlargement, thus increasing size and weight of the fruits at harvest (Ojeda et al. 2001). Tavr_Mar presents a negative regression coefficient (b = −10.248), which shows that anomalously low temperatures in March tend to be favorable to WP. This is in agreement with results for other winemaking regions in the country (Santos et al. 2013). Insufficient winter chill may bring harmful effects, such as later and irregular budburst (Webb et al. 2007), which may in fact result in lower yields. The frequencies of occurrence of two regimes in three months (AFeb, AMay, ASep and RMay) are also selected. The differences in the growing season occurrences of A and R between low and high production years (Fig. 5.7) are in accordance with their regression coefficient signals (b = +0.498 for AFeb; b = +1.676 for AMay; b = −2.419 for ASep; b = −0.540 for RMay). Despite the relatively high anomalies in AMar and RJun, they are not selected by the stepwise method, as they are negatively correlated with Tavr_Mar and PJun, respectively 93 Chapter 5. Climate factors driving wine production in the Portuguese Minho region (collinearity). Anomalously high (low) frequencies of occurrence of A Feb and AMay (ASep) are beneficial for high production. The A regime tends to be associated with settled weather conditions and relatively dry air masses, as the anticyclone over Iberia suppresses local instability and advections (Fig. 5.5f). It is generally favorable to settled-cool weather in February (in the winter half of the year through radiative cooling) and settled-warm weather in May and September (in the summer half of the year through radiative warming). Therefore, for February (May), with positive regression coefficient, settled-cool (settled-warm) weather is beneficial, which also supports the relationship found with Tavr for the following March (June). As ASep presents a negative regression coefficient, settled-warm weather conditions are not favorable in this period. Further, since R tends to be favorable to WP in September (Fig. 5.7), which is cooler and moister than A owing to the prevailing northerly maritime winds (Fig. 5.5e), cool weather is then preferred over warm weather. Cool night temperatures in the period preceding harvest might be indeed beneficial for high quality wines and yields (Kliewer and Torres 1972; Tonietto and Carbonneau 2004). Lastly, the frequency of occurrence of RMay negatively influences WP, as settled-cool weather is not beneficial during this period of the growing season, which also corroborates the AMay-WP relationship above. In general, R is not favorable in the period May-July, but becomes favorable in AugustSeptember (Fig. 5.7), which stresses the importance of settled-warm weather in the floweringveraison phenophase and settled-cool weather during late maturation. Additionally, the settled weather prevailing in both regimes is commonly linked to low risks of thunderstorms and hailstorms, which are also prominent limiting factors for WP. This can partially explain the low relevance of other potentially favorable regimes, but being relatively unsettled their positive contributions can be offset by hazardous events (e.g. hailstorms). 94 Chapter 5. Climate factors driving wine production in the Portuguese Minho region Fig. 5.7 - Relative monthly frequencies of the occurrence (in %) of each weather regime for the period 1950-2010 and across the year. 5.3.5. Logistic model The logistic model is then based on ten regressors for modeling the WP classes (low, normal and high production years), as follows: X t 1 DI HyI Prod 3 TGMar TGJun PJun exp Bi X t pi ,t 2 1 expB X i AFeb AMay ASep RMay , i 1, 2 t i 1 p3,t 1 p1 p2 where Xt is the regressor matrix as a function of time t (1950-2010), Bi the regression coefficients and pi,t the probabilities of occurrence of the i-th WP class as a function of time. 95 Chapter 5. Climate factors driving wine production in the Portuguese Minho region The modeled class (Fig. 5.8a) successfully corresponds to the observed class in 92% of the years (61 years), failing only in five years. When analyzing the model matches/errors (Fig. 5.8b), class 1 shows one error, while 2 and 3 show two errors each. It is clear that class 3 shows a higher percentage of errors (13%), while 1 and 2 show error percentages of about 7%. A leave-one-out cross-validation scheme is also applied to take into account model overfitting (Wilks 2011), now revealing a model skill ratio of 72%. These outcomes highlight the skillfulness of the logistic model, allowing future applications. This is particularly noteworthy in view of the complexity of factors governing the variability of WP, some of them being non-climatic factors. Furthermore, a sensitivity analysis leaving the frequencies of occurrence of the regimes out of the model revealed a much lower skill (55% after crossvalidation), which hints at their importance in modeling WP. Fig. 5.8 - (a) Chornogram of the observed (modeled) classes represented by black (white) circles. (b) The number of matches (light grey bars), the number of errors (dark grey bars) and the percentage of errors in each class are pointed out. 96 Chapter 5. Climate factors driving wine production in the Portuguese Minho region 5.4. Discussion and conclusions Viticulture in general is strongly influenced by weather and climate, a relationship that is clearly reinforced by the present study for the MWR in northwestern Portugal. For modeling its WP series, ten regressors are selected by a stepwise methodology: the Dryness and Hydrothermal indices, the 3-yr lagged production, monthly mean temperatures in March and June, total precipitation in June and the frequencies of occurrence of the A and R regimes in May, and of the A regime in February and September. Although some of these regressors are in agreement with WP-climate relationships obtained for the nearby Douro Wine Region (Tavr_Mar and Tavr_Jun), others clearly demonstrate the singularity of the MWR. A logistic modeling of the WP classes (low, normal and high), using these ten selected regressors, reveals a high model skill ratio of 72 % after cross-validation. The model can prove useful for viticulturists, as predicting WP classes provide a more reliable assessment than estimating annual WP with a large uncertainty (large confidence interval in the estimate). Therefore, the model not only isolates the main climatic driving factors (favorable and unfavorable) of interannual variability in WP, but may also be used as a valuable tool for WP prediction. The relatively long WP series (1945-2010, 66 years) allows a statistically robust isolation of the main climatic forcings on its inter-annual variability, particularly its short-term variability, as the long-term oscillations are removed in the percentile-transformed series. However, the selected climate-based indices show only weak trends that cannot explain the strong longterm trends and oscillations observed in WP. Other non-climatic factors, such as changes in vineyard area and location, agricultural and enological practices, commercial decisions and economic strategies, market policies, among others, tend to be manifested as low-frequency oscillations owing to their gradual implementation. Nevertheless, thorough assessments of these factors are difficult to undertake using the available information. Overall, despite other potentially important non-climatic factors, the atmospheric factors satisfactorily explain the year-to-year changes in WP. Results presented herein suggest that moderate growing season dryness, high production 3-yrs before, moderate cool conditions in February and March, followed by settled-warm weather in May, warm and moist weather in June, and settled-cool weather in September are favorable to WP. The results suggest that the precipitation and humidity levels may be deeply tied to WP in this region. High precipitation in the growing season seems to be harmful, while moderate water stress is generally beneficial. This is not verified in other Iberian regions, such 97 Chapter 5. Climate factors driving wine production in the Portuguese Minho region as in northeastern Spain, where lowest yields are related to the driest and hottest years (Camps and Ramos 2012). This different behavior hints at the specific characteristics of the MWR, when compared to other usually much drier Iberian regions. Relatively high annual precipitation and humidity are reinforced by deep soils, with high water holding capacity, and by compact valley systems where most grapevines grow, with enhanced run-offs. Another critical factor is the duality between wine quality and yield. It is well-known that the optimum quality-yield balance is usually established by growers, since higher water availability tends to result in higher production but lower quality (Ramos et al. 2008). Since high yields and high quality seem extremely difficult to achieve simultaneously (and often not economically viable), compromises are usually made through yield reduction (van Leeuwen et al. 2009). This may have also happened in the MWR, particularly in the more recent years, but the identification of producer decisions is not possible. In fact, the stabilization of WP towards the end of the series suggests that the wine quality-yield balance may have been pursued in recent years and, to some extent, successfully accomplished. It should be noted that the assessment of wine quality and its annual variation in relation to climatic factors is also a critical issue for winemakers, however, this assessment is out of the scope of the current research. The projected future warming and drying in the MWR (Fraga et al. 2012b) may bring further challenges to local winemakers and may modify the above-mentioned quality-yield balance. In effect, a recent study suggests that the winemaking industry should focus on minimizing the impact of yearly fluctuations and on optimizing production of high-quality grapes (Keller 2010a). This issue is especially relevant under scenarios of future warmer climates (Fraga et al. 2013; Webb et al. 2008a; Duchene et al. 2010). Changes in traditional training systems and in grown varieties are already underway in the MWR (Magalhães 2008). Growers are currently changing traditional varieties to later maturity varieties (e.g. Touriga Nacional), aiming to adapt to climate change, but maintaining the production of high-quality and competitive wines. The model here developed may be used as a valuable tool for WP stabilization, since it provides an inherent probability assessment that may improve future planning within the winemaking sector (Meyer and Mart ne -Casasnovas 1999). However, as it is tuned for the transformed production series, it does not take into account long-term trends and lowfrequency oscillations. Therefore, this model can only be used in near-term predictions, as extrapolation to the long-term future projections is not possible. Furthermore, the current 98 Chapter 5. Climate factors driving wine production in the Portuguese Minho region study provides evidence for the connection between several atmospheric factors and the WP in the MWR, which greatly contributes to the understanding of how climate may influence viticultural parameters. 99 Chapter 5. Climate factors driving wine production in the Portuguese Minho region 100 Chapter 6. Winegrape phenology and temperature relationships in the Lisbon Wine Region, Portugal Journal International des Sciences de la Vigne et du Vin, 2013, 47(4): 287-299 Impact Factor (5-yr): 1.265 Aureliano C. Malheiro, Rita Campos, Helder Fraga*, José Eiras-Dias, José Silvestre, João A. Santos *- Author provided assistance in the statistical methodologies, climatological assessments, figure and table preparation, and in the writing of the article 101 Chapter 6. Winegrape phenology and temperature relationships in the Lisbon Wine Region, Portugal Abstract Aim: To investigate the characteristics, relationships and trends in the phenology of four winegrape varieties and associated temperature relationships in the Lisbon wine region (LWR), between 1990 and 2011. Methods and results: Budburst, flowering and véraison dates of red (Castelão and Aragonez, syn. Tempranillo) and white (Chasselas and Fernão Pires) varieties were taken from an experimental vineyard in the LWR. Harvest dates were determined based on a similar maturity level for all varieties. From these data, varietal characteristics, temporal trends as well as relationships between phenology and temperature were assessed through stepwise multivariate linear regressions. Flowering was the most sensitive to temperature in the preceding months (March-April). Differences/similarities between the phenological timing of the different varieties are presented. With few exceptions, no trends were found in phenophases over the 1990-2011 period, whereas several significant negative slopes were displayed for phenological intervals, suggesting a role for accumulated thermal effects in phenological timing. Strong correlations were observed between phenophases, especially between flowering and véraison. Conclusion: The study highlights the key role played by temperature on phenology, particularly during springtime. Furthermore, an increase in temperature during that period will cause an advance in the timing of the following phenological events. Given the significant trends found, phenological shifts may occur in the long term, emphasizing the need to assess varietal characteristics and responses to regional climate. Significance and impact of the study: The present work is the first attempt to systematically examine temporal trends in phenology and corresponding relationships with temperature in a Portuguese viticultural area, providing valuable information on the development and suitability of grapevine varieties, which determine viticultural practices and winegrower’s income. 102 Chapter 6. Winegrape phenology and temperature relationships in the Lisbon Wine Region, Portugal 6.1. Introduction It is widely acknowledged that vitiviniculture is mainly governed by weather and climate conditions (e.g. Due et al. 1993; Jackson and Lombard 1993; Keller 2010a; Fraga et al. 2014a). In fact, Vitis vinifera L. is a perennial and deciduous crop, whose vegetative cycle, morphologically described by phenological stages, is mostly driven by air temperature. Precipitation and radiation are also important, though to a lesser extent (Jones 2003; Bock et al. 2011; Tomasi et al. 2011; Urhausen et al. 2011). After a period of chilling temperatures in winter for breaking dormancy, the growing season begins with budburst in early spring. A cumulative effect of temperatures above a threshold of 10°C (base temperature) is the classical thermal requirement for this event to take place (Winkler 1974). The following main events are flowering, véraison (berry softening and colour change) and harvest (with fruit ripening in summer), during which warm temperatures (dry and stable atmospheric conditions) are required for balanced crop yield and wine quality (Jones and Davis 2000b; Santos et al. 2013; Santos et al. 2011). Later, the vines lose their leaves (leaf fall). Mild temperatures and considerable precipitation totals during the dormant period are critical for the next growing season, particularly in Mediterranean winegrowing regions (Magalhães 2008). All these climatic requirements are reflected in the global geographic distribution of this species, preferably located where growing season mean temperatures range from 12 to 22°C (Jones 2006). Within these boundaries are the typical Mediterranean climates, which includes the Portuguese winegrowing areas, despite important regional differences and strong inter-annual variability that may significantly influence winegrape phenology (Fraga et al. 2012b). In addition, the timing of each phenophase differs according to each grapevine variety, and is generally tied to local thermal conditions (e.g. flowering and véraison; Parker et al. 2011). Indeed, phenological models, based on temperature accumulations above a base temperature between events, such as Growing Degree Days (GDD), have been extensively reported in the literature (e.g. Chuine et al. 2003; Jones 2003). More complex models using different threshold values and chilling requirements have also been developed (de CortazarAtauri et al. 2009; Caffarra and Eccel 2011). Taking into account these heat requirements, different derived (bioclimatic) indices have been used for viticultural climatic zoning and inter-regional and global comparisons (Malheiro et al. 2010; Fraga et al. 2014a; Fraga et al. 2013). Other studies linking phenology and climate variability (especially temperature) used temperature averages rather than summations (Grifoni et al. 2006; Garcia-Mozo et al. 2010; 103 Chapter 6. Winegrape phenology and temperature relationships in the Lisbon Wine Region, Portugal Dalla Marta et al. 2010; Bock et al. 2011; Tomasi et al. 2011) in order to avoid correlations between accumulated values and elapsed time (Due et al. 1993; Jones and Davis 2000a; Urhausen et al. 2011). Most studies that addressed these links have reported earlier phenophase dates, shorter phenological intervals and warmer grape maturation periods, suggesting that these changes are mostly likely due to the temperature rise observed in recent years (Jones and Davis 2000a; Duchene and Schneider 2005; Jones et al. 2005a; Bock et al. 2011; Tomasi et al. 2011; Urhausen et al. 2011). Furthermore, phenological changes are projected to be particularly striking in the upcoming decades due to warming trends (Webb et al. 2007; Duchene et al. 2010; Caffarra and Eccel 2011). Therefore, shifts in grapevine varieties, wine quality and typicity, and even in the location of the main winegrowing regions may occur in the long term (Malheiro et al. 2010; Santos et al. 2012; Fraga et al. 2013). Most importantly, these authors projected detrimental impacts in most wine regions of southern Europe, which would require short term measures (e.g. irrigation; Ferreira et al. 2012) to cope with environmental stresses. The phenological characteristics of individual grapevine varieties, as well as their responses to climate, are of utmost importance for activity planning and decision making in viticulture. Indeed, they may have direct impacts on vine yield and wine production, and, ultimately, on winegrower’s income (Magalhães 2008). Therefore, the present study evaluated the influence of temperature variability on the phenology of four winegrape varieties in the Lisbon Wine Region (LWR henceforth), of south-western Portugal. In addition, the relationships between the onset and duration of the main phenological stages and their trends were analysed. To our knowledge, this is the first attempt to systematically examine temporal trends in phenology and corresponding climate links in a Portuguese viticultural area. As the selected region is the second most important with respect to production volume (after the Douro Valley), it also plays a central role in the Portuguese economy (IVV 2011). 104 Chapter 6. Winegrape phenology and temperature relationships in the Lisbon Wine Region, Portugal 6.2. Material and methods 6.2.1. Location and climate data The study was carried out in the LWR (former Estremadura region), consisting of a coastal north-south strip about 30 km wide, stretching from just north of Lisbon to central Portugal (Fig. 6.1). The winegrowing area is characterized by a Mediterranean climate (warm temperate climate with a dry and warm summer; Kottek et al. 2006), but with a noteworthy Atlantic influence. The ‘WorldClim’ database was used (Hijmans et al. 2005), which provides gridded data (with spatial resolution of about 1 km at the Equator) of monthly precipitation and maximum, minimum and mean air temperatures over 1950-2000, was used for the description of regional climate conditions. These variables were then extracted over the target area (8183 grid cells), overcoming the considerable lack of data provided by the few weather stations located within the region. Fig. 6.1 – a) Average annual (left panel) and April-September (right panel) precipitation (in mm) for the 1950-2000 period, Lisbon Wine Region, Portugal. b): the same as a) but for the air temperature (in ºC). 105 Chapter 6. Winegrape phenology and temperature relationships in the Lisbon Wine Region, Portugal At the vineyard level, two sets of daily maximum, minimum and mean air temperature records were used, for the calculation of monthly values. The first set (1999-2011) was recorded from a standard meteorological weather station close to the experimental site (Dois Portos: 39.0ºN, 9.2ºW, 110 m asl). As this set did not cover the entire 22 years of the study period (1990-2011), daily data from the Lisbon meteorological station (Lisboa-Geofísico: 38.7°N, 9.1°W, 80 m asl) were selected over the study period. By applying linear regressions between these two datasets in their common period (1999-2011), high coefficients of determination were obtained for monthly maximum, minimum and mean temperatures (0.96 ≤ R2 ≤ 0.99). This result highlights the similarity of the inter-annual climate variability within the LWR, which is fundamentally forced by large-scale atmospheric patterns (Santos et al. 2011). Using the regression coefficients from these linear regression equations, a third and complete set of monthly (maximum, minimum and mean) temperature data was estimated for Dois Portos in the extended period of 1990-2011 (reconstructed series); the correlation coefficients between the raw and the reconstructed time series were ≥ 0.98 for the three variables in their common period (1999-2011). The yearly evolution of maximum, mean and minimum temperatures (reconstructed series) for the four trimesters is shown in Figure 6.2. Even though these reconstructed series are not free of bias, the very high correlation coefficients referred to above suggest that their inter-annual variability was reliably reproduced, which is by far the most important property when modelling the inter-annual variability of phenological variables. In effect, this methodology allows overcoming an important limitation: the absence of simultaneous meteorological and phenological data in the same location and over a relatively long time period. However, these data constraints prevented the use of daily temperatures. Therefore, temperature thresholds (Jones and Davis 2000a), varietal thermal requirements (Duchene et al. 2010; Parker et al. 2011) and derived indices (Fraga et al. 2014a; Jones 2003) were not computed in the present work. The monthly series are then only related to grapevine phenology. Although precipitation was also considered in a first exploratory study, it presented no statistically significant correlations with phenological stages (95% confidence level) and was thus discarded from the analysis (not shown). 106 Chapter 6. Winegrape phenology and temperature relationships in the Lisbon Wine Region, Portugal Fig. 6.2 - Yearly evolution (1990-2011) of maximum (TX), mean (TG) and minimum (TN) temperatures (reconstructed series) for the (a) winter (Jan-Feb-Mar), (b) spring (Apr-May-Jun), (c) summer (Jul-Aug-Sep) and (d) autumn (Oct-Nov-Dec) in Dois Portos, Portugal. Straight lines represent linear regressions for significant trends, LT their trends, and R2 their coefficients of determination. p-values < 0.05 are shown. 6.2.2. Plant material and phenology The dates of key phenophases of two red (Aragonez and Castelão) and two white (Chasselas and Fernão Pires) Vitis vinifera L. varieties were obtained from the Portuguese national ampelographic collection at the state laboratory ‘Instituto Nacional de Investigação Agrária e Veterinária’, in Dois Portos (which includes the former ‘Estação Vitivinícola Nacional’). These varieties cover the longest period of phenological records within the collection. The vineyard, which covers about 2 ha and 724 accessions or botanical clones, aims to preserve the native Portuguese varieties, allowing characterization, identification and synonymy/homonymy study. It has also the goal of being used as a reference collection for varieties and rootstocks grown in Portugal, including those widely planted worldwide. The experimental vines, are trained on vertical trellis at a plant density of about 3200 vines/ha, and grafted onto SO4 rootstock. The soil is classified as a calcic fluvisol (eutric) (FAO 2006). Castelão and Fernão Pires are two native winegrape varieties, typically planted in the study area. Aragonez (syn. Tempranillo) is largely grown in the Iberian Peninsula and Chasselas in Switzerland (known as Fendant) and France (Magalhães 2008). Regarding phenological precocity, budding ranges from early (Castelão, Chasselas and Fernão Pires) to 107 Chapter 6. Winegrape phenology and temperature relationships in the Lisbon Wine Region, Portugal average (Aragonez), and all are early ripening (Lopes et al. 2008; Magalhães 2008), though temperature sum requirements along the growing season may vary with variety and climate (Parker et al. 2011). The observed date of budburst, flowering and véraison of the four varieties were monitored from 1990 to 2011 (22 years), except for Aragonez, for which no data was available for the first four years. These phenological stages were considered to occur when 50% of the plants had reached a particular stage. Budburst, flowering and véraison correspond to stages 07, 65 and 81, according to the BBCH scale, respectively (Lorenz et al. 1995). In addition, harvest dates (stage 89 on the BBCH scale) were available for the 19902011 period (except for the 1994 vintage). The white varieties were harvested when probable alcohol content of Fernão Pires fruits were about 11.5% (vol.), which is considered as a reference, though Chasselas berries generally had lower probable alcohol content at harvest (not shown). A similar approach was carried out for the red varieties using Castelão as a reference (11.5% vol.). For this purpose, berry samples from these varieties were collected from the end of August to early September (2-4 samplings) of every vintage. As the harvest date is generally considered as a ‘false’ phenophase influenced by weather conditions, oenological decisions and winery logistics (Petrie and Sadras 2008), a comparison of harvest dates with similar level of ripeness was developed. Therefore, an estimated harvest date for each variety reaching a probable alcohol content of 11.5% (vol.; hereafter designated as harvest) was defined by taking the linear rate of change in probable alcohol content during maturation. The observed phenological events were also used to derive the intervals between each event (budburst to flowering, budburst to véraison, flowering to véraison and véraison to harvest). The growing season length corresponds to the number of days between budburst and harvest. 6.2.3. Statistical analysis Statistical analysis of phenology from the four varieties was undertaken using the tStudent’s test in conjunction with the least significant differences of means. The results are then presented as mean estimates ± standard deviation, for inter-annual variability assessment. In addition, medians, interquartile ranges and outliers were determined. In order to examine linear dependencies between these variables, Pearson correlation coefficients were also computed. The differences between means and linear regression trends were assumed to be statistically significant at the 5% level. A multivariate linear regression model (stepwise 108 Chapter 6. Winegrape phenology and temperature relationships in the Lisbon Wine Region, Portugal methodology) was applied (using a least-squares approach) to grapevine phenology using the full set of selected potential predictors (monthly maximum, minimum and mean temperatures separately and combination) as independent variables. 6.3. Results and discussion 6.3.1. Climatic conditions Average of annual total precipitation and mean temperature patterns within the LWR were 780 mm and 15.4ºC, respectively, over the 1950-2000 period (Fig. 6.1). Typically, the northern hilly areas and the western coast were much wetter and cooler than the southeasternmost area (where Lisbon is located). Indeed, the average climatic conditions in Lisbon were warmer and drier than those typically found in the LWR. Furthermore, high elevation areas, such as in the Montejunto mountain range and the coastal strip (sea breeze influence), presented the coolest conditions. For the growing season (April-September), the area-mean temperatures were clearly higher than their annual means (nearly +3ºC), and the area-mean precipitation were significantly lower (nearly -600 mm), which is typical of the LWR Mediterranean climate. The remarkable seasonality in precipitation can be demonstrated at the experimental site (Dois Portos), where the long-term average precipitation from April to September was about 160 mm, which represents roughly 22% of the total annual amount. At the same site, annual and growing season mean temperatures were 15.7 and 18.4ºC, respectively. During the study period (1990-2011), relevant seasonal differences between years were found (Fig. 6.2). In fact, there were summer periods (July, August and September) presenting anomalously high mean temperatures (e.g. 1990, 1991, 2003 and 2010, anomalies of +0.71.6ºC) and winter periods (January, February and March) presenting anomalously low mean temperatures (e.g. 1991, 1992, 2005 and 2006, anomalies ranging from -1.0 to -1.4ºC). Furthermore, significant temperature trends were displayed for autumn (maximum temperature) and springtime (Fig. 6.2). Regarding the annual values (Fig. 6.3), some years (e.g. 1995, 1997 and 2011) were anomalously warm (anomalies in the annual mean temperature of +0.6-0.8ºC), whereas others (e.g. 1992, 1993 and 1999) were anomalously cold (anomalies in the annual mean temperature from -1.1 to -0.4ºC). Significant positive trends were also found for the annual means of daily maximum and mean temperatures 109 Chapter 6. Winegrape phenology and temperature relationships in the Lisbon Wine Region, Portugal (+0.04-0.03ºC/yr, p ≤ 0.016). This inter-annual variability is expected to have forced the whole phenology of the winegrapes in the LWR and to be reflected in the winegrape phenology of the study site in particular. 6.3.2. Characteristics and relationships in phenology Overall, the time series of the main phenological events for the four selected winegrape Fig. 6.3 - Annual values of maximum (TX), mean (TG) and minimum (TN) temperatures (reconstructed series) for the 1990-2011 period in Dois Portos, Portugal. Enclosed symbols of mean temperatures correspond to values that fell below first (squares) and above third (circles) quartiles and straight lines represent linear regressions for significant trends, LT their trends, and R 2 their coefficients of determination. p-values < 0.05 are shown. varieties revealed a very similar behaviour, with clear inter-annual variability during 19902011 (Fig. 6.4). Furthermore, median dates were quite similar to mean dates for each phenophase of each variety, but significant differences in means were found between varieties (not shown). However, the mean budburst date (March 14 ± 8 days) of Castelão, Chasselas and Fernão Pires did not differ significantly, but was distributed over nearly one month, from February 27 (2001) to March 30 (2005) (Fig. 6.4). On the other hand, this phenological event took place from March 8 (1998 and 2001) to April 6 (2005) for Aragonez, which was, on average, near a week later (March 20 ± 9 days) compared with the other three varieties, revealing its (mean) budburst precocity (Lopes et al. 2008). 110 Chapter 6. Winegrape phenology and temperature relationships in the Lisbon Wine Region, Portugal Fig. 6.4 - On the left: Box plots of budburst, flowering, véraison and harvest dates of Aragonez, Castelão, Chasselas and Fernão Pires varieties from 1990 to 2011 in Dois Portos, Portugal. Medians correspond to horizontal black lines within boxes, lower (upper) box limits to the first (third) quartiles and whiskers to the non-outlier maxima and minima. First (second) order outliers are indicated by circles (asterisks) and represent events above/below the box upper/lower limit by at least 1.5 (3.0) times the respective box height (interquartile range). On the right: Time series of the same phenophases and varieties for the outlined period. The dashed lines represent the linear regression for the budburst of Aragonez and harvest of Fernão Pires, LT represents its trend, and R 2 represents the coefficient of determination. 111 Chapter 6. Winegrape phenology and temperature relationships in the Lisbon Wine Region, Portugal There were no significant differences between varieties in the mean flowering dates (May 20 ± 9 days), with a maximum about two/three weeks later (June 2-7). Outlier minima were also reported for all varieties in 1997 (April 24) that corresponded to an anomalously warm March-April period (anomaly of +2.8ºC in the two-month mean). Regarding the véraison dates, Chasselas presented higher range (42 days) compared with Castelão (34 days), but significant earlier onsets (July 27 ± 11 days and August 4 ± 8 days, respectively; Fig. 6.4 and Table 6.1). Despite no significant differences, Aragonez and Fernão Pires tended to onset maturation between the former two varieties. As previously, 1997 was an outlier for all varieties, except Fernão Pires, and 2011 (anomaly of +3.0ºC in the April-May mean) was an outlier for Aragonez and Chasselas. Table 6.1 - Means, standard deviations and linear trend parameters (R2, p-value and annual trend) of main phenophase dates and corresponding intervals for the four varieties in Dois Portos, Portugal (1990-2011). Variable Variety n Mean Std. Dev. R2 p-value Trend (days/yr) Aragonez 18 Mar 20 9 0.45 0.002 +1.1 Castelão 22 Mar 12 8 ns Budburst Chasselas 22 Mar 14 8 ns Fernão Pires 22 Mar 15 7 ns Aragonez 18 May 23 9 ns Castelão 22 May 18 8 ns Flowering Chasselas 22 May 20 9 ns Fernão Pires 22 May 21 8 ns Aragonez 18 Jul 29 9 ns Castelão 22 Ago 4 8 ns Véraison Chasselas 22 Jul 27 11 ns Fernão Pires 22 Ago 2 10 ns Aragonez 17 Sep 15 6 ns Castelão 21 Sep 20 8 ns Harvest Chasselas 21 Sep 25 10 ns Fernão Pires 21 Sep 16 10 0.49 <0.001 -1.0 Aragonez 18 64 9 ns Budburst Castelão 22 67 8 ns to Chasselas 22 68 8 ns Flowering Fernão Pires 22 67 8 ns Aragonez 18 131 12 0.45 0.003 -1.5 Budburst Castelão 22 145 10 0.21 0.035 -0.7 to Chasselas 22 135 10 ns Véraison Fernão Pires 22 140 10 ns Aragonez 17 179 11 0.51 0.001 -1.5 Budburst Castelão 21 192 10 ns to Chasselas 21 195 12 0.24 0.023 -0.9 Harvest Fernão Pires 21 185 13 0.49 <0.001 -1.4 Aragonez 18 68 6 0.67 <0.001 -0.9 Flowering Castelão 22 78 5 0.40 0.002 -0.5 to Chasselas 22 68 8 ns Véraison Fernão Pires 22 73 7 ns Aragonez 17 48 8 ns Véraison Castelão 21 47 7 ns to Chasselas 21 60 13 ns Harvest Fernão Pires 21 45 10 0.25 0.021 -0.7 112 Chapter 6. Winegrape phenology and temperature relationships in the Lisbon Wine Region, Portugal Average harvesting of Chasselas occurred around 10 days later compared with Aragonez and Fernão Pires (September 15 ± 8 days), which should be explained by its generally lower probable alcohol content (compared with the other three varieties) at harvest. Furthermore, Chasselas showed a high range (36 days), similar to Fernão Pires (38 days) but lower than Aragonez (21 days) and Castelão (27 days). Importantly, despite environmental effects, days to berry maturity is also varietal dependent (Petrie and Sadras 2008; Webb et al. 2007, 2011) and genetically determined (Magalhães 2008). An evidence of the latter characteristic might be found for Shiraz, syn. Syrah, in Australia, where over four consecutive seasons, ripening parameters (°Brix and berry weight) were more closely correlated with the number of days after flowering than with GDD (McCarthy 1999). In the present study, an assessment of the intervals between the main phenological events showed high inter-annual variability (Table 6.1). Commonly, the flowering-véraison interval was the longest, but with relatively low inter-annual variability, followed by the budburst-flowering interval. The véraison-harvest interval presented the highest inter-annual variability. The budburst-flowering interval ranged from about 50 days, for the shortest period in 1997 to about 79 days for the longest period in 2001, averaging 67 (±8) days during the study period. Conversely, significant differences between varieties were found for the growth period between flowering and veraison, with the exception (no difference) of Aragonez and Chasselas, where the averaged interval was 68 (±7) days, being also the shortest interval. On the other hand, Castelão revealed the longest average interval (78 ±5 days), but with the lowest range (18 days). The length of the véraison to harvest interval (47 ± 8 days, Aragonez, Castelão and Fernão Pires) did not differ significantly between varieties, except for Chasselas, which was the last to reach maturity (60 ± 13 days), and ranged from 30 (year 2005) to 77 (year 2011) days. The length from budburst to harvest is particularly important for defining the growing season period, which for Chasselas lasted from 160 (in 2005) to 209 (in 1990) days, with an average of 195 (± 12) days. This last value was significantly extended compared to Fernão Pires (185 ± 13 days) and Aragonez (179 ± 11 days), but not to Castelão. Significant differences were also detected between Aragonez and Castelão for this same interval, which tended to have shorter and longer growing seasons, respectively. Relationships between phenophases and varieties were clear (Table 6.2). Overall, significant correlations were observed between successive phenological stages; these were 113 Chapter 6. Winegrape phenology and temperature relationships in the Lisbon Wine Region, Portugal particularly strong between flowering and véraison (r ≥ 0.73 for all varieties). Additionally, the latter phase was also correlated with harvest for Aragonez, Castelão and Fernão Pires (r ≥ 0.50), whereas no significant correlations were found between budburst and harvest. Interestingly, the spring (a critical period as discussed in the next section) of 2011 was clearly the warmest of the study period (Fig. 6.2), promoting the shortest flowering-véraison followed by the longest véraison-harvest interphases (indicating less suitable conditions) for Chasselas. However, intervals between phenophases were not significantly correlated (not shown), revealing that, for example, a shorter (longer) budburst-flowering period does not necessarily imply a shorter (extended) flowering-véraison period. These results are in agreement with previous studies (Tomasi et al. 2011; Bock et al. 2011; Jones and Davis 2000a). Table 6.2 - Pearson correlations coefficients between main phenophases of the four varieties in Dois Portos, Portugal (1990-2011). Values in bold are significant (p < 0.05). Budburst Flowering Véraison Aragonez Castelão Chasselas Fernão Pires Budburst Flowering Véraison Harvest 0.45 0.09 -0.02 0.77 0.34 0.50 Budburst Flowering Véraison Harvest 0.53 0.30 0.32 0.80 0.47 0.61 Budburst Flowering Véraison Harvest 0.57 0.44 0.11 0.73 0.58 0.29 Budburst Flowering Véraison Harvest 0.51 0.37 -0.07 0.77 0.45 0.55 6.3.3. Temperature and phenology: effects and trends Close relationships between main phenophases and temperature of the preceding months were reported using a multivariate regression approach. Moreover, all variables were negatively correlated (r < 0, not shown), meaning that increased temperatures during the selected months anticipate a specific event. The stepwise method was used to select the most significant predictors (individual and combined maximum, minimum and mean temperatures) at monthly (and multi-monthly) timescales in 1990-2011 (Table 6.3). 114 Chapter 6. Winegrape phenology and temperature relationships in the Lisbon Wine Region, Portugal Table 6.3 - Summary of most significant and not inter-correlated (99% confidence level) monthly (or multimonthly) temperature variables (TX, TN and TG combined and analysed separately) and corresponding coefficients of determination after cross-validation (Rcv2) for key phenophases of the four varieties in Dois Portos, Portugal (1990-2011). TX, TN and TG: maximum, minimum and mean temperatures, respectively. nd: not determined. Note that all temperatures are negatively correlated with phenophase dates. Budburst Flowering Véraison Harvest 2 2 2 Regressor Regressor Rcv Regressor Rcv Regressor Rcv2 Rcv TX, TN and TG as variables Aragonez TN Jan-Feb-Mar 0.46 TX Mar-Apr 0.75 TG Mar-Apr 0.57 nd TG Mar-Apr TX Apr Castelão TN Jan-Feb-Mar 0.69 TX Mar-Apr 0.78 TN Jun 0.87 TN Jun 0.63 TX Jul TX Sep TG Mar-Apr TN Mar-Apr TG Mar-Apr Chasselas TN Feb-Mar 0.60 0.84 0.57 0.61 TN Apr TN Jul TN Aug TX Mar Fernão TN Mar-Apr TX Apr TN Feb-Mar 0.63 TX Mar-Apr 0.79 0.68 0.77 Pires TN Jul TX Jun TX Aug TX as variable Aragonez TX Jan-Feb-Mar 0.29 TX Mar-Apr 0.75 TX Mar-Apr 0.56 nd TX Mar-Apr TX Mar Castelão TX Jan-Feb-Mar 0.41 TX Mar-Apr 0.78 TX Jun 0.84 TX Apr 0.51 TX Jul TX Jun Chasselas TX Feb-Mar 0.33 TX Mar-Apr 0.77 TX Mar-Apr 0.35 TX Apr 0.32 TX Mar Fernão TX Apr TX Jan-Feb-Mar 0.45 TX Mar-Apr 0.79 TX Mar-Apr 0.50 0.77 Pires TX Jun TX Aug TN as variable Aragonez TN Jan-Feb-Mar 0.46 TN Mar-Apr 0.52 TN Mar-Apr 0.48 nd TN Mar-Apr Castelão TN Jan-Feb-Mar 0.69 TN Mar-Apr 0.54 TN Mar-Apr 0.63 0.34 TN Jun TN Mar-Apr TN Mar-Apr Chasselas TN Feb-Mar 0.60 TN Mar-Apr 0.67 0.57 0.56 TN Jul TN Aug Fernão TN Mar-Apr TN Apr TN Feb-Mar 0.63 TN Mar-Apr 0.60 0.68 0.53 Pires TN Jul TN Jun TG as variable Aragonez TG Jan-Feb-Mar 0.43 TG Mar-Apr 0.74 TG Mar-Apr 0.57 nd TG Mar-Apr TG Mar-Apr Castelão TG Jan-Feb-Mar 0.62 TG Mar-Apr 0.75 TG Jun 0.86 0.52 TG Jun TG Jul TG Mar-Apr TG Mar-Apr Chasselas TG Feb-Mar 0.47 0.82 TG Mar-Apr 0.44 0.58 TG Apr TG Aug TG Apr Fernão TG Jan-Feb-Mar 0.61 TG Mar-Apr 0.78 TG Mar-Apr 0.59 TG Jun 0.70 Pires TG Aug Combining the three temperatures increased the coefficient of determination in several cases, though the selected months were generally the same. In fact, the discrepancy in the selection of the combined climatic variables can be primarily explained by the stepwise methodology. As maximum and minimum temperatures (besides mean temperature) for a given month were highly correlated (0.76 ≤ r ≤ 0.89 for all months, except September with r = 115 Chapter 6. Winegrape phenology and temperature relationships in the Lisbon Wine Region, Portugal 0.56), this method automatically selects the variable with better statistical performance in order to avoid multi-collinearity and overfitting. Hence, after the selection of one of the three variables for a given month, the others for the same month are automatically excluded, regardless of the physiological responses. Therefore, the isolation of the predictor months is greatly important, though multiple variables should be tested (avoiding previous selection of variables) in a statistical analysis (Due et al. 1993). Average minimum temperatures of the two-month February-March period (Chasselas and Fernão Pires) or the three-month January-February-March period (Aragonez and Castelão) were the major variables controlling grapevine budburst for the studied varieties (0.46 ≤ Rcv2 ≤ 0.69), with Aragone presenting the lower response. Only the latter variety displayed a significant trend of +1.1 days/yr (p = 0.002, Table 6.1), indicating that this event has been delayed along the period of study (Fig. 6.4). Nonetheless, the results need to be cautiously interpreted as the data series is shorter for Aragonez (18 years) and may be substantially influenced by anomalously high (or low) values (e.g. April 6, 2005). In fact, Petrie and Sadras (2008) highlighted similar times series constraint using a 1993-2006 period. Additionally, budburst is projected to occur later in a warmer area of Australia (Margaret River) in response to future warming (Webb et al. 2007). The authors pointed out the insufficient winter chilling for breaking dormancy in the region for these projections. However, no significant temperature trends were found over the winter periods in our study (Fig. 6.2), though short term (e.g. daily) responses may not be totally expressed in monthly averages (Due et al. 1993). Practices such as pruning should not have a determinant impact on that response, as it was performed in similar time periods of each year (early January). On the other hand, other factors, such as soil temperature, soil water availability (precipitation effect) and starch content in the roots should influence this stage (Jackson and Lombard 1993; Lavee and May 1997; Keller 2010a). These results highlight the importance of developing phenological models (Chuine et al. 2003; Parker et al. 2011) quantifying the precise base temperatures and thermal requirements (defining varietal suitability), which may vary with variety and location (de Cortazar-Atauri et al. 2009; Oliveira 1998). As an illustration, calculated (budburst) threshold temperature may vary from 5.1 to 6.9ºC for Müller-Thurgau in cool winter climates (Nendel 2010). Conversely (and similarly to our results in three varieties), no significant budburst trends were described for a long-term (1964-2009) dataset of eighteen varieties in the Veneto region, Italy (Tomasi et al. 2011) and in Lower Franconia, 116 Chapter 6. Winegrape phenology and temperature relationships in the Lisbon Wine Region, Portugal Germany (Bock et al. 2011), whereas earlier budburst were reported in Alsace (Duchene and Schneider 2005) and Burgundy (Jones et al. 2005a), France, as a result of temperature rise. Flowering timing also showed to be mostly influenced by March temperature, but now combined with April (preceding two-month period). Regarding the analysis of the combined three climatic variables, maximum March-April temperature was chosen for all the varieties except Chasselas, for which mean March-April and minimum April temperatures were selected. The model explained 75-84% of the total variance of this event, which was generally the highest significance. The effect of springtime thermal conditions was particularly clear in the advanced flowering of 2007 (outlier for all varieties) as described above. Other studies (Jones and Davis 2000a; Bock et al. 2011; Tomasi et al. 2011) have reported that both flowering and budburst are mainly controlled by (maximum) March and April temperatures. The former result may be ascribed to the warmer climate conditions within the LWR (and then earlier budburst) compared with the above studies (e.g. Germany). Moreover, strong linear relationships between flowering dates and spring temperatures were also found for other plant species (e.g. Menzel et al. 2005). Véraison was still mainly driven by March and April temperatures (0.57 ≤ Rcv2 ≤ 0.87 for the combined analysis), which is consistent with the strong correlation between flowering and véraison onsets detailed above. Another determinant variable was late monthly temperature (e.g. July). Once again, April temperatures were crucial for harvest, though preceding months/variables were also selected (0.61 ≤ Rcv2 ≤ 0.77 for the combined analysis), suggesting the increased influence of cultural practices over the season. However, no significant predictors could be isolated for the harvest of Aragonez, due to its smaller sample size (17 years). Therefore, the strong link between winegrape phenology and springtime thermal conditions, as well as the increased influence of viticultural practices (non-climatic drivers) during maturity, are determinant in the LWR, as in other regions (Jones and Davis 2000a; Duchene and Schneider 2005; Webb et al. 2007; Petrie and Sadras 2008; Bock et al. 2011; Tomasi et al. 2011). Significant positive trends in temperatures were found for the spring period (Fig. 6.2). However, no significant trends were observed for flowering, véraison and harvest onsets, with the exception of Fernão Pires where anticipation in harvest (negative trend) of roughly -1.0 days/yr was depicted (p ≤ 0.001, Table 6.1). This result may be ascribed to a higher climatic sensitivity over maturation, which was also exposed in interphase trends (as described below). Indeed, based on acid and sugar levels over a 45-year period, Jones and Davis (2000a) 117 Chapter 6. Winegrape phenology and temperature relationships in the Lisbon Wine Region, Portugal concluded that Merlot is a more phenologically and climatically sensitive winegrape variety compared with Cabernet Sauvignon in Bordeaux, France. Additionally, negative trends (shortest interval) in the time series were mostly found for interphases: flowering to véraison (Aragonez and Castelão), véraison to harvest (Fernão Pires) and budburst to harvest (Aragonez, Chasselas and Fernão Pires) (Table 6.1), which are generally linked to better climatic conditions for growth and development (Jones and Davis 2000a). However, in warmer Portuguese regions, harvest dates may be brought forward into a warmer period of the year, with detrimental effects on yield and quality (Jones 2006; Duchene et al. 2010; Santos et al. 2011). The significance of the apparent discrepancy in slopes between dates of events and intervals in phenology may be attributed to the accumulated thermal effect along the season (Tomasi et al. 2011), which cannot be clearly disclosed by the date of each phenophase. Though generally not statistically significant, all seasonal (and yearly) mean temperatures displayed positive slopes (not shown). Furthermore, a persistent effect of spring (near budburst) temperatures was also reported to influence the vegetative growth of Cabernet Sauvignon grapevines in Washington, USA (Keller and Tarara 2010). Other studies (Jones and Davis 2000a; Duchene and Schneider 2005; Bock et al. 2011; Tomasi et al. 2011; Neethling et al. 2012; Webb et al. 2012) showed significant phenological shifts, particularly for the last events (véraison and harvest) and intervals (e.g. budburst to harvest). However, these authors generally reported lower rates (e.g. roughly -0.3 days/yr, for the budburstharvest interval in Germany and northern Italy), which may be attributed to the different temporal windows and lengths used for the trend assessments in the different studies. In fact, Webb et al. (2012) computed significant higher trends in maturity (e.g. -1.25 days/yr for Shiraz in Central Victoria, Australia), when a shorter and common period in their study (1985-2009) was defined. In the same country, Sadras and Petrie (2011) concluded that the early maturity of Chardonnay, Shiraz and Cabernet Sauvignon varieties was mostly related to early onset (86% of the variation in maturity) rather than faster ripening in response to warming. Moreover, increases in potential alcohol level at harvest were reported during the last decades in Alsace (Duchene and Schneider 2005). It is worthwhile noting that, contrary to usual wine production practices (e.g. harvest dates are varietal and oenological dependent), the trends in maturity were assessed by a biophysical parameter (potential alcohol content) similar to all varieties, which diminishes the effect of non-climatic drivers (Webb et al. 2011; Webb et al. 2012). Nevertheless, the period 118 Chapter 6. Winegrape phenology and temperature relationships in the Lisbon Wine Region, Portugal between budburst and véraison was also determined in order to disregard the harvest date. Non-significant trends were displayed in the former interval for the white varieties (Chasselas and Fernão Pires), indicating the major influence of the later phase (i.e., harvest) (Table 6.1). Conversely, for the red varieties, which are commonly more heat demanding, significant trends were found between budburst and véraison (e.g. -1.5 days/yr for Aragonez), revealing that effects may occur at earlier periods (flowering to véraison) of the cycle (Table 6.1). 6.4. Conclusions This study presents the first analysis of the temporal trends in winegrape phenology and corresponding connections to temperature in a major Portuguese wine region. Our results highlight the key role played by springtime thermal conditions in phenology, particularly in flowering, which in turn influence the following phases, though non-climate drivers (e.g. viticultural practices) are also important during ripening. While only a few trends were found for phenophases, several significant negative slopes were displayed for interphases, which was a reflection of the accumulated thermal effects along the seasons. This issue is especially relevant under a scenario of projected future warmer climates (where these relationships may be changed), although this assessment is out of the scope of the present study. As a result, phenological shifts may occur in the long term, emphasizing the need to assess varietal characteristics and responses to climate at regional level. Despite the considerable lack of relatively long time series of phenological data for the Portuguese wine regions, further analysis on the impacts of climate change in winegrape phenology is currently underway, which may provide crucial information on the development and suitability of grapevine varieties, determining viticultural practices and winegrower’s income. 119 120 Chapter 7. Integrated analysis of climate, soil, topography and vegetative growth in Iberian viticultural areas Submitted Impact Factor (5-yr): 2.837 Helder Fraga, Aureliano C. Malheiro, José Moutinho-Pereira, Rita M. Cardoso, Pedro M. Soares, Javier J. Cancela, Joaquim G. Pinto and João A. Santos 121 Chapter 7. Integrated analysis of climate, soil, topography and vegetative growth in Iberian viticultural areas Abstract The Iberian viticultural regions are convened according to the Denomination of Origin (DO) and present different climates, soils, topography and management practices. All these elements influence the vegetative growth of different varieties throughout the Iberian Peninsula, and are tied to quality and wine type. In the current study, an integrated analysis of climate, soil, topography and vegetative growth is performed for the Iberian DO regions, using state-of-the-art datasets. For climatic assessment, a categorized index, accounting for phenological/thermal development, water availability and ripening conditions is computed. Soil textural classes are established to distinguish soil types. Elevation and aspect (orientation) are also taken into account, as the leading topographic elements. Using a spectral vegetation index, all the previous elements are then integrated as a function of the vegetative growth of the Iberian vineyards. Results show that most Iberian vineyards are grown in temperate dry climates with loamy soils, presenting low vegetative growth. Other vineyards, in temperate humid conditions tend to show higher vegetative growth. In these climates, soil type and topography play a minor role for grapevine growth. Conversely, in cooler and warmer climates, soil type and water availability acquire more important roles. Soils with higher clay content limit vegetative growth, whereas higher sand content provide better growing conditions. These assessments allow direct comparison between DO regions and may be of great value to viticulturists, as they play a key role when including vineyards into a given DO. This is particularly important taking into account the projected climate for future decades. 122 Chapter 7. Integrated analysis of climate, soil, topography and vegetative growth in Iberian viticultural areas 7.1. Introduction The most renowned viticultural regions in the Iberian Peninsula (Portugal and Spain) have a long standing tradition in winemaking and are considered world-class grapevine (Vitis vinifera L.) producing regions. Spain, which currently has the largest vineyard area in the world (over 1×106 ha), is the 3rd wine producer worldwide, while Portugal ranks in the 11th place, with internationally acclaimed wines, such as the Port wine (OIV 2012). In these countries, viticultural regions are convened according to Denominations of Origin (DO), or Qualified Denomination of Origin (DOCa), which are imposed by governmental institutions and controlled by strict regulations. In fact, the Portuguese Douro/Porto DO was the first viticultural region worldwide to implement these regulations (in 1756; Magalhães 2008). Such regulations aim at obtaining a superior wine quality, while establishing the specific wine characteristics of each region (IVV 2011). The concept of terroir, which includes specific soil, topography, climate, landscape characteristics and biodiversity features of each winemaking region (OIV 2010), is entrenched within the classification of a DO. Each DO is expected to be a recognized trademark, whilst other vineyards/products not included in the DO are not allowed to bear this denomination. Being part of the terroir, soil is one of the most important factors for a sustainable viticulture (Magalhães 2008). It supports the root system, which accumulates carbohydrates, water and other nutrients, being crucial for grapevine growth, physiology and yield attributes (Winkler 1974; Morlat and Jacquet 2003). Soil structure and chemistry can influence grapevine composition and consequently wine quality (Mackenzie and Christy 2005). Compact and shallow soils can obstruct root access to oxygen, water and nutrients, limiting root growth and development and inhibiting proper grapevine behaviour (Jackson and Lombard 1993). In grapevines, nutrient and water uptake occur mostly within 0.5–1.0 m soil profile (Magalhães 2008; Keller 2010b). Therefore, deep soils with good drainage (either natural or manmade) are usually preferred for vineyard implementation (Morlat and Jacquet 1993). Additionally, soil heat and water retention properties are also important, as they can affect the phenological development of grapevines (Yau et al. 2013; Field et al. 2009). A high soil water storage capacity is indeed important in regions where grapevines are subjected to high levels of heat and water stress, as is the case of the Mediterranean regions (Flexas et al. 2010). 123 Chapter 7. Integrated analysis of climate, soil, topography and vegetative growth in Iberian viticultural areas Climate, also an important component of the terroir, is widely acknowledged as one of the most important factors for grapevine development and growth (Fraga et al. 2014b, 2013; Jackson and Lombard 1993; Keller 2010a; Malheiro et al. 2010). During the growing season (April–October in the Northern Hemisphere) of this perennial and deciduous crop, climatic conditions are responsible for numerous morphologically and physiological changes. One of the most well-known climatic limitations of grapevine is the 10°C base temperature, needed for the onset of its yearly vegetative cycle (Winkler 1974). Throughout its different stages of development, sunlight, heat and water demands vary, so as to guarantee optimum and balanced grape ripening. In fact, it has been shown that the timings and duration of the grapevine phenological stages are deeply tied to the prevailing atmospheric conditions (e.g. Jones and Davis 2000a; Malheiro et al. 2013), which also contribute to variability in grapevine yield (e.g. Santos et al. 2011; Bindi et al. 1996), wine production (e.g. Fraga et al. 2014b; Santos et al. 2013; Camps and Ramos 2012) and quality (e.g. Jones et al. 2004b; Jones and Goodrich 2008). All these climatic factors limit the geographic distribution of this crop (Jones et al. 2005b; Jones 2006; Fraga et al. 2013), being also key factors in determining the varietal suitability and wine type of a given region (Carbonneau 2003; Jones 2006; Fraga et al. 2014a). The topographic elements represent yet another key factor that influences viticultural and oenological characteristics of a given region. In Europe, vineyards are traditionally grown in a large array of orographic conditions, from extended flatland areas to steep mountainous regions (Nascimbene et al. 2013). Amongst the most important topographic elements for viticulture are elevation, slope and aspect/exposition (Yau et al. 2013; Jones et al. 2004a). These elements further enhance the singularity of viticultural regions, since the cultural and management practices are deeply related to these elements (Magalhães 2008). As an example, growers tend to select row orientation according to the geographical aspect (orientation) of the vineyard, e.g. horizontal direction to which a terrain slope faces. Usually, in the Iberian viticulture, growers generally tend to prefer southern facing areas, increasing the solar radiation interception by foliar organs (Magalhães 2008). While elevation is generally a determining element for site selection, due to the decreased (increased) thermal accumulation at higher (lower) altitudes (e.g. Fraga et al. 2012b), steep-slopes can be overcome through the implementation of agricultural practices, such as walled terraces or vertical planting. The Iberian Peninsula presents a wide range of all these grapevine-influencing elements. From a climatic perspective it delivers a relatively large set of mesoclimates, 124 Chapter 7. Integrated analysis of climate, soil, topography and vegetative growth in Iberian viticultural areas spanning from dryer regions, in the inner south, to more humid regions, in the north and northwest (AEMET/IMP 2011; Santos et al. 2012). Topography and soils are also quite distinct throughout the peninsula, influencing crop selection and settlements in each region. All these elements are reflected in the different varieties grown throughout the peninsula (Böhm 2010). While in southern Iberia, red varieties usually prevail, in the north ,white varieties are more common (Malheiro and Santos 2011). Thus, each Iberian viticultural region presents a large number of autochthonous grapevine varieties, best adapted to the different climates, soils and topographic elements (Fraga et al. 2012b). Given the heterogeneous conditions in which grapevines are grown in Iberia, understanding the complex relationships between all these factors represents a serious concern for grapevine growers and winemakers. The present study aims to evaluate the conditions of the viticultural regions in Iberia, regarding the main features of the terroir, also focusing on the sustainability and development of this valuable agribusiness. This is first integrated analysis of this kind over the entire Iberian Peninsula. Therefore, the objectives of this work are three-fold: 1) to assess the prevailing conditions in terms of climate, soil and topography in the Iberian viticultural regions; 2) to develop an integrated analysis of the previous three elements and their impact on vegetative growth; and 3) to establish a zoning of homogeneous climate-soil-topographyvegetative growth areas. 7.2. Material and methods 7.2.1. Viticultural regions and vineyard area As previously referred, the Iberian DO regions are subjected to strict regulatory requirements. Although the spatial distribution and limits of each DO are subjected to different rules in Portugal and Spain, the significance of the DO is nearly the same. In particular, the law enforcements are similar to both Portugal and Spain, but they still differ in certain aspects of national legislation. To assess the spatial characteristics of each DO region, boundaries of each DO or DOCa (Rioja and Priorat) are defined using data available in the Portuguese ‘Instituto do Vinho e da Vinha’ (IVV; http://www.ivv.min-agricultura.pt) and Spanish ‘Ministerio de Agricultura, Alimentación y Medio Ambiente’ (MAGRAMA; http://sig.magrama.es). The viticultural regions in the islands of Madeira, Azores (Portugal) and Canarias (Spain) are not assessed due to limitations in the soil and climatic datasets. 125 Chapter 7. Integrated analysis of climate, soil, topography and vegetative growth in Iberian viticultural areas Other viticultural regulated regions, such as, quality wine with specific geographical indication, estate wine, qualified estate wine and contry wines, Indicação de Proveniencia Regulamentada (in Spain) and Vinho Regional (in Portugal), are out of the scope of the current study, since DO regions are usually considered of higher importance. Note that the DO regions of Málaga and Sierras de Málaga (in Spain) are two different DO regions that coincide geographically and therefore are treated jointly (henceforth DO Málaga & Sierras de Málaga). As a result, the spatial boundaries of 81 DO regions (82 effectively), 25 in Portugal and 56 (57) in Spain, are identified within Iberian Peninsula (Fig. 7.1a). In order to analyse the spatial distribution of the vineyards in Iberia, the Corine Land Cover Map, version 13–2012, is used (EEA 2002; Büttner et al. 2006). This dataset, last updated in 2012, provides an inventory of the land cover over Europe and has previously proven to have high accuracy in representing the land cover over mainland Portugal and Spain (Caetano et al. 2006; Genovese et al. 2001). The vineyard fraction of the land cover over Iberia extracted from this dataset for subsequent analysis (Fig. 7.1b). 7.2.2. Topography For the topographic analysis, elevation and geographical aspect are selected as the leading topographic elements to be studied, due to their importance in Iberian viticulture. Although slope is also considered a key landscape element, it is not included in the current analysis, as in European vineyards certain practices, such as walled terraces, often adjust to natural slopes. As detailed information such practices like walled terraces is not available, this cannot be taken into account. For the elevation assessment, the GTOPO30 digital elevation model (DEM) is used at the 30 arc-seconds spatial resolution (https://lta.cr.usgs.gov/GTOPO30). Elevation ranges are isolated inside each region (according to Fig. 7.1a, b). Aspect is calculated using the same GTOPO30 DEM and geographical information systems. 126 Chapter 7. Integrated analysis of climate, soil, topography and vegetative growth in Iberian viticultural areas Fig. 7.1 - - a) Location of the viticultural regions in Iberia, along with their denomination. b) Spatial distribution of the vineyard land cover over Iberia (dark-red), assessed using the Corine Land Cover, version 13- 2012, along with the viticultural regions (light-grey). 127 Chapter 7. Integrated analysis of climate, soil, topography and vegetative growth in Iberian viticultural areas 7.2.3. Climate In order to analyse the climatic conditions of each region, a categorized bioclimatic index for viticultural zoning is used (CatI; Fraga et al. 2014a). CatI establishes climatic categories by combining three bioclimatic indices (Table 7.1): Huglin Index (Huglin 1978), Dryness Index (Riou et al. 1994) and Cool Night Index (Tonietto and Carbonneau 2004). The Huglin Index expresses the thermal potential of a given region and relates the heat accumulation to the physiological development of grapevines. The Dryness Index assesses water availability for grapevines, by estimating potential water balance over the growing season. The Cool Night Index accounts for minimum temperatures at the end of the vegetative cycle (September in the Northern Hemisphere), as lower nocturnal temperatures during this stage tend to be favourable for wine quality. Thus, CatI allows determining the optimum climatic suitability in terms of phenological development, heat and water stress, as well as ripening conditions (Supplementary Table S3.1). Although these conditions may vary according to each variety, it is taken into account that grapevines in Iberia are well adapted to their local climatic conditions. Table 7.1 - Categorized Index (CatI), along with the corresponding classes of the combined indices: Huglin, Dryness and Cool Night indices, according to Fraga et al. (2014a) Cool Night Index Huglin Index Dryness Index (°C) CatI (°C) (mm) Description (Tonietto and (Huglin 1978) (Riou et al. 1994) Carbonneau 2004) <900 <-100 Unsuitably cold or excessively dry 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 -100 – 50 900 – 1500 > 50 -100 – 50 1500 – 2100 > 50 -100 – 50 2100 – 2700 > 50 -100 – 50 > 2700 > 50 < 14 > 14 < 14 > 14 Cool, dry with cool nights Cool, dry with warm nights Cool, humid with cool nights Cool, humid with warm nights < 14 > 14 < 14 > 14 Temperate, dry with cool nights Temperate, dry with warm nights Temperate, humid with cool nights Temperate, humid with warm nights < 14 > 14 < 14 > 14 Warm, dry with cool nights Warm, dry with warm nights Warm, humid with cool nights Warm, humid with warm nights < 14 > 14 < 14 > 14 Very warm, dry with cool nights Very warm, dry with warm nights Very warm, humid with cool nights Very warm, humid with warm nights 128 Chapter 7. Integrated analysis of climate, soil, topography and vegetative growth in Iberian viticultural areas For CatI calculation (and combining indices), data from a regional climate model (RCM) Weather Research and Forecast model (WRF; Skamarock et al. 2008) version 3.1.1 is used. The WRF model is a non-hydrostatic model and has been widely used for dynamical downscaling regional climate. For the present climate, a simulation with a horizontal grid resolution of 9-km (nested in a 27-km grid) was used, with both grids centred in the Iberian Peninsula. The RCM simulation started at 00 00 UTC 1 January 1989 and ended at 18 00 UTC 31 January 2013, with initial, lateral and lower boundary conditions derived from ERAInterim. From the model output, precipitation and temperature over Iberia are considered for this study. A more detailed description of the model set-up can be found in Soares et al. (2012) and Cardoso et al. (2013), where the simulation results were validated for inland maximum and minimum temperatures and precipitation, showing a good agreement with observations. Patterns of the simulated CatI are them compared to previously established patterns using observational data (Fraga et al. 2014a), showing a good agreement. Additionally, solar radiation (surface net downward shortwave flux), from the Modern Era Retrospective-analysis for Research and Applications (MERRA; http://gmao.gsfc.nasa.gov/merra/) at a 0.6º × 0.6º longitude/latitude spatial resolution, is also obtained. Mean values over the growing season are calculated for 1989-2012. 7.2.4. Soils For soil analysis, the predominant soil texture (SoilT) is assessed according to the United States Department of Agriculture soil textural classification (USDA 2006). The texture of a soil refers to its relative content of clay, sand and silt particles (Table 7.2). Each soil texture class presents its own properties in terms agricultural applicability. Clay soils have fine particles and retain large amounts of water, but are poorly drained and usually difficult to manage (Juma 1999; McKenzie et al. 2008). Conversely, sandy soils are coarse and usually excessively drained, with low water retention capacity (Juma 1999; McKenzie et al. 2008). With relatively even proportions between particles, the loamy soils are typically well drained and provide sufficient nutrient retention and are thus usually preferable for agricultural use (Juma 1999; McKenzie et al. 2008). Soil texture is a fundamental soil property used as a qualitative classification tool to determine other soils properties (USDA 2006). Therefore, soil texture establishes relationships between textural classes and soil plasticity, drainage, available water content, amongst others (McKenzie et al. 2008). Soil texture classes (SoilT) 129 Chapter 7. Integrated analysis of climate, soil, topography and vegetative growth in Iberian viticultural areas are obtained from the Harmonized World Soil Database (HWSD; FAO/IIASA/ISRIC/ISSCAS/JRC 2012). Table 7.2 - Soil texture categories, along with the respective percentages of Clay, Silt and Sand, according to USDA soil textural classification (USDA 2006). Soil Category 1 2 3 4 5 6 7 8 9 10 11 12 13 Clay (%) Silt (%) Sand (%) 60-100 40-60 40-60 27-40 27-40 0-12 0-27 35-55 7-27 20-35 0-20 0-15 0-10 0-40 40-60 40-60 40-73 15-52 88-100 74-88 0-20 28-50 0-28 0-50 0-30 0-14 0-45 0-20 0-45 0-20 20-45 0-20 20-50 45-65 23-52 45-80 50-70 70-86 86-100 Texture heavy clay silty clay clay silty clay loam clay loam silt silty loam sandy clay loam sandy clay loam sandy loam loamy sand sand 7.2.5. Vegetative growth The Enhanced Vegetation Index (EVI) is used for the analysis of the vegetative growth. Spectral vegetation indices are based on visible and near-infrared radiation fluxes, captured by sensors on-board of polar orbiting satellites, and are a measure of the concentration of green leaf vegetation (Huete et al. 2002). The EVI algorithm accounts for canopy background (e.g. soil and bare earth) and atmospheric effects (e.g. clouds), while also being barely affected by manmade structures (Huete et al. 2002; Pennec et al. 2011). Applied to viticulture, spectral vegetation indices have shown relationships with production and wine attributes (Usha and Singh 2013; Johnson et al. 2001), also assisting management activities (Johnson et al. 2003). In this study, the EVI from the Moderate Resolution Imaging Spectroradiometer (MODIS - MOD13A2 Collection 5) is extracted from the National Aeronautic and Space Administration (NASA) Land Processes Distributed Active Archive Center (LP DAAC; https://lpdaac.usgs.gov/). Four MODIS tiles (h17v04, h17v05, h18v04 and h18v05), covering all of the Iberian mainland are obtained at a 1-km spatial resolution for the 2012 growing season. This single year was selected to match the latest available land cover update (described in section 2.1), which ensures land cover changes do not interfere with the analysis 130 Chapter 7. Integrated analysis of climate, soil, topography and vegetative growth in Iberian viticultural areas of grapevine vegetative growth. The mean growing season EVI is then calculated (AprilOctober mean) and the spatial average for the vineyard areas over Iberia is 0.23. For categorization of each DO according to the vegetative growth, two EVI classes (EVIc-1 and 2) are defined: EVIc-1 (EVI ≤ 0.23), for low vegetative growth areas, and EVIc-2 (EVI > 0.23), for high vegetative growth. 7.3. Results 7.3.1. Mesoscale patterns Overall, topography over Iberia displays large differences in both elevation and aspect (Fig 7.2a, b). Effective solar radiation depicts a strong north/south contrast, with higher solar radiation values in the south (Fig. 7.2c). Noticeable is the relatively low solar radiation in centre western Iberia (near La Mancha DO, #33), when compared to the surrounding areas. Grapevines in these areas are generally less affected by excessive solar radiation. CatI (Fig. 3a) depicts a clear distinction between the northern regions, generally cooler, and the southern regions, with higher thermal accumulation. In fact, most of northern Iberia presents CatI-3 (cool, humid, with cool nights), making this the predominant category regarding all of Iberia. Also in the north, CatI-0 climates are present, showing the lack of viticultural suitability owing to insufficient thermal accumulation. Apart from the CatI-0 areas, the lower category, CatI-1 (Cool, dry, with cool nights), is observed in an isolated area near the centre of the peninsula. Conversely, the warmer climatic region CatI-14 is located in the centre-south. CatI-10/9 (warm, dry, with warm/cool nights), represent the 2nd/3rd dominant climatic categories, scattered across the south and northeast. CatI-5 occurs over a large area in centre/northern Iberia, being the 4th dominant category (taking into account all of the mesoscale pattern in Iberia). Other secondary categories appear in transitional areas, between cool and warm climates. For example, CatI-11/12, with warm and humid climates, appear isolated in opposite sides of the peninsula (east/west), depicting the singularity of these regions. Soil textural classes present a very homogeneous pattern (Fig. 7.3b and Table 7.2). Loamy soils (SoilT-9), which are commonly considered highly suitable for agriculture, are prevalent in most of the Iberian Peninsula. However, in central Iberia, sandy-loam soils (SoilT-11) are more frequent, whereas in southwestern Iberia loamy-sand soils (SoilT-12) are 131 Chapter 7. Integrated analysis of climate, soil, topography and vegetative growth in Iberian viticultural areas also common. Clay loam soils (SoilT-5) are isolated in a small region in southwestern Iberia. Clay soils (SoilT-3) dominate most of southern and eastern Iberia. In northwest Iberia, some intrusions of silt-loam soil (SoilT-7) are also depicted. Fig. 7.2 - a) Elevation (m) in the Iberian Peninsula, calculated using the GTOPO30 dataset. b) As in (a) but for the aspect. c) Solar radiation over Iberia, mean growing season values in 1989-2012 calculated using MERRA data at a 0.6º spatial resolution. 132 Chapter 7. Integrated analysis of climate, soil, topography and vegetative growth in Iberian viticultural areas Fig. 7.3 - a) CatI over Iberia calculated according to Table 1, for the period of 1989-2012 using WRF simulations. b) SoilT according to Table 2 using HWSD data. c) Mean EVI and EVIc for the grapevine growth period (April-October) in 2012, using MODIS data. The spatial-average of the vineyard areas corresponds to 0.23. Below this value the EVIc equals 1 (transparent overlay), above this value EVIc equals 2 (hatched overlay). 133 Chapter 7. Integrated analysis of climate, soil, topography and vegetative growth in Iberian viticultural areas Regarding the vegetative development, represented by EVI and EVIc (Fig. 7.3c), the north/south and west/east contrasts are clear. Northern and western areas generally present a much higher EVI than southern and eastern areas. This is particularly clear in the EVIc, where the coastal areas in northern and western Iberia are classified as EVIc-2, while the rest of Iberia is generally keyed to EVIc-1. This distinction in the vegetative growth could in fact be tied to the strength of the Atlantic influence in the Iberian Mediterranean-like climates, with much higher precipitation amounts over northern and western Iberia than over central and eastern Iberia (Fraga et al. 2014b). Spearman ranked correlation coefficients between the previous mesoscale patterns (Supplementary Table S3.2) reveal weak to moderate correlations. The highest positive correlation is found between CatI and surface net solar radiation flux (0.61), as both indirectly reflect the latitudinal effect over temperature and incoming solar radiation. The strongest negative correlation (-0.66) is found is between CatI and elevation, undoubtedly reflecting the temperature lapse rate already embedded into this climatic index. Regarding the grapevine growth, a negative moderate correlation (-0.35) is found between EVI and CatI, while a weak positive correlation is found between EVI and SoilT (0.12). Fig. 7.4a depicts the grapevine vegetative growth in Iberia (EVIc) as a function of their climates (CatI) and soil characteristics (SoilT). It is clear that most vineyard areas are keyed to SoilT-9 (loam). This soil type reveals a strong influence of the climatic conditions on grapevine growth, in which water availability is fundamental. Grapevine growing in regions with SoilT-9 seems much more dependent on water availability than other soil types, such as SoilT-3 and SoilT-11/12, with predominantly EVIc-1 and EVIc-2 (respectively). Vineyards with SoilT-3 tend to show much lower vegetative growth, when compared to other soils with less clay content (SoilT-11/12), and regardless of the prevailing climatic conditions (with the exception of CatI-6 and CatI-8 areas). Concerning the climatic influences on vegetative growth, the CatI-6 and 8 regions predominantly EVIc-2. These two climate categories seem to induce higher grapevine vegetative growths, independently of the SoilT category, which is also true for CatI-7 vineyards (except in SoilT-3 areas). This suggests that in temperate climates (CatI-6, 7 and 8), water availability, thermal amplitudes and soil characteristics tend to be less relevant for higher vegetative growth than in cooler or warmer climates. While both climate and soil characteristics seem to induce different responses on grapevine vegetative growth, in terms of importance, the prevailing soil textural class seems to overpass climatic conditions in regions 134 Chapter 7. Integrated analysis of climate, soil, topography and vegetative growth in Iberian viticultural areas with SoilT-11 and 12, where all vineyards show higher grapevine vegetative growths, regardless of the prevailing climatic category. Conversely, in temperate climates soil characteristics loses some of its influence on grapevine growth. Fig. 7.4 - a) Circular accumulated EVIc (1 – red, 2 – green) as a function of the SoilT and CatI for all vineyards in Iberia. b) Circular accumulated EVIc as a function of the elevation and aspect for all vineyards in Iberia. The size of each circular chart depicts the accumulated vineyard area belonging to that EVI class, and the inner (outer) circular class depicts the largest (smallest) EVI class. The same integrated analysis is also performed taking into account grapevine vegetative growth as a function of elevation and aspect (Fig. 7.4b). The largest vineyard concentrations are located at 600 - 800m elevation ranges. It is clear that low elevation vineyards (most of them along coastal strips) present much higher vegetative growths, which can be partly explained by the more humid coastal local climates. Regarding geographical aspect, it is visible that growers tend to prefer southern orientation for their vineyards. No 135 Chapter 7. Integrated analysis of climate, soil, topography and vegetative growth in Iberian viticultural areas other strong relationship can be established between elevation /aspect and vegetative growth, suggesting that these factors play a minor role, when compared to climate and soil. 7.3.2. DO analysis With respect to vineyard locations (Fig. 7.1b), they are mostly confined by the DO boundaries, with the largest visible concentrations of vineyards in La Mancha and Rioja (Spain, #33 and #60) and Douro (Portugal, #28). Overall, 81 DO are depicted in Fig. 7.1a. These regions show large differences in terms of shape and area (Fig. 7.1a; Table S3.3). The largest DO is La Mancha (Spain) and the smallest is Carcavelos (Portugal, #17). The elevation means and ranges of the vineyard areas for each DO (Fig. 7.5) highlight the large spatial variability in which vineyards are grown, ranging from near sea-level (e.g. DO Tavira, #65) to elevations over 850 m (DO Arlanza, #6). Near coastal regions tend to have vineyards at much lower mean elevations, while in the innermost DO regions vineyards show higher mean elevations and ranges (Fig. 7.2a). The largest vineyard areas are indeed located at mean elevations ranging from 600 to 800m, with the most heterogeneous being DO Valencia, which is also one of largest in terms of surface (Table S3.3). As expected, small DO regions, such as Tavira (#65), Pla i Llevant (#51), Lafões (#34), Monterrei (#43), Mondéjar (#42) and Arlanza (#6), tend to show much lower spatial variability. It should be noticed that La Mancha (with the largest vineyard area; Table S3.3) shows relatively low variability in elevation (relatively flat area) when compared to other smaller regions in mountainous areas (e.g. Douro, Table S3.3). When analysing the geographical aspect of the vineyards within each DO (Fig. 7.6), it is clear that terrains with an S-SW aspect are preferred for viticultural activities. This outcome mainly reflects the largest vineyard area located at 600-800 m elevations (Fig. 7.6a). Despite this fact, vineyards in 800-1000 m elevation tend to have S-SE aspects (Fig. 7.6b), while vineyards at lower elevations (0-200 m) have N-NW aspects (Fig. 7.6b). For elevations in the 200-600 m range, no clear distinction on aspect preference can be made. Regarding the solar radiation, while most DO regions in Portugal present a high solar radiation, DO regions in Spain are usually located in areas with lower radiations (with the exception of some regions in the south). 136 Chapter 7. Integrated analysis of climate, soil, topography and vegetative growth in Iberian viticultural areas Fig. 7.5 - Elevation (m) of the vineyards in each DO/DOCa in Iberia. The inner circle represents the mean elevation and the horizontal bars represent the minimum and maximum, of the locations of the vineyards inside the DO. Fig. 7.6 - a) Geographical aspect (orientation) vineyards in Iberia according to the mean elevation of each DO. b) as in (a) but normalized using the vineyard area. In Table 7.3, the 81 DO are defined according to their predominant categories of CatI, SoilT and EVIc. The cool, dry, with cool nights DO regions (CatI-1) of Tierra del Vino de Zamora (#70) and Toro (#71) exhibit similar clay soils (SoilT-3) with low vegetative growth (EVIc-1), while the also CatI-1 DO Rueda (#61) exhibits sandy-loam soils (SoilT-11) and high vegetative growth (EVIc-2). This may indicate that regions with similar climatic 137 Chapter 7. Integrated analysis of climate, soil, topography and vegetative growth in Iberian viticultural areas conditions, but with different soil types, can indeed present different vigour and yield attributes. This effect is less visible in the equally cool, but more humid, CatI-3 regions of Ribera del Duero (#57), Arlanza (#6), Cigales (#22), Chacolí de Álava (#19), Chacolí de Guetaria (#20), Tierra de León (#69) and Valdeorras (#76). While in the first two regions, SoilT-3 is predominant, each one shows a different EVIc (1 and 2 respectively). The other CatI-3 regions depict SoilT-9 (loam) and EVIc-2, with the exception of Cigales that exhibits EVIc-1. As previously mentioned, CatI-5 (Temperate, dry, with cool nights) was the 4th leading category (regarding all Iberia). Nonetheless, it is the dominant category with respect to just the vineyard area, being present in 19 DO regions (Bullas (#14), Conca de Barberà (#24), Manchuela (#39), Ribera del Júcar (#59), Mondéjar (#42), Penedès (#49), Alicante (#4), Almansa (#5), Cariñena (#18), Douro (#28), Priorat (#53), Trás-os-Montes (#73), Uclés (#74), Utiel-Requena (#75), Yecla (#81), Calatayud (#15), Távora-Varosa (#66), Arribes (#7) and Dão (#27)). These regions have SoilT-3, 9 and 11. For CatI-5, only SoilT-11 seems to present higher vegetative growth (in more than 50% of these DO regions). However, for CatI6 regions (same as CatI-5 but with warm nights), all SoilT-9, 11 and 12 regions are keyed to EVIc-2 (Alella (#1), Alenquer (#2), Arruda (#8), Colares (#23), Montsant (#45), Bucelas (#13), Obidos (#47), Torres Vedras (#72) and Lourinha (#37)). This suggests higher nocturnal temperatures may be beneficial for vegetative growth. The CatI-7 regions show the same relationship with soil as previously seen for CatI-3. Since these regions present humid conditions, SoilT-9 is clearly beneficial for vegetative growth when compared to SoilT-3. CatI-8 regions (Encostas d'Aire (#30) and Bairrada (#9)) are the only regions where EVIc-2 is present, regardless of SoilT. This suggests that temperate climates, with humid conditions and warm nights, are the ideal conditions for a higher grapevine growth. CatI-9 and CatI-10 regions display similar characteristics to those already reported, higher vegetative growths for SoilT-9, 11 and 12 than for SoilT-3. The only DO regions in Iberia that currently present very warm, dry, with warm nights climate (CatI14) is the DO Jerez (#31), that also exhibits clay soils (SoilT-3) and consequently low vegetative growth (EVIc-1). 138 Table 7.3 - CatI, SoilT and EVI class for each viticultural region in Iberia. Only the predominant categories are shown. # 70 71 61 57 6 22 19 20 69 76 14 24 39 59 42 49 4 5 18 28 53 73 74 75 81 15 66 7 27 17 36 Region Tierra del Vino de Zamora Toro Rueda Ribera del Duero Arlanza Cigales Chacolí de Álava Chacolí de Guetaria Tierra de León Valdeorras Bullas Conca de Barberà Manchuela Ribera del Júcar Mondéjar Penedès Alicante Almansa Cariñena Douro Priorat Trás-os-Montes Uclés Utiel-Requena Yecla Calatayud Távora-Varosa Arribes Dão Carcavelos Lagos CatI 1 1 1 3 3 3 3 3 3 3 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 6 6 SoilT 3 3 11 3 3 9 9 9 9 9 3 3 3 3 3 3 9 9 9 9 9 9 9 9 9 9 9 11 11 3 3 EVI 1 1 2 1 2 1 2 2 2 2 1 1 1 1 2 2 1 1 1 1 1 1 1 1 1 2 2 1 2 1 1 # 1 2 8 23 45 13 47 72 37 60 63 16 11 21 34 43 54 55 56 79 30 9 33 32 46 58 77 80 26 50 41 Region Alella Alenquer Arruda Colares Montsant Bucelas Obidos Torres Vedras Lourinhã Rioja Somontano Campo de Borja Bierzo Chacolí de Vizcaya Lafões Monterrei Rías Baixas Ribeira Sacra Ribeiro Vinho Verde Encostas d'Aire Bairrada La Mancha Jumilla Navarra Ribera del Guadiana Valdepeñas Vinos de Madrid Costers del Segre Pla de Bages Méntrida CatI 6 6 6 6 6 6 6 6 6 7 7 7 7 7 7 7 7 7 7 7 8 8 9 9 9 9 9 9 9 9 9 139 SoilT 9 9 9 9 9 11 11 11 12 3 3 9 9 9 9 9 9 9 9 9 11 12 3 9 9 9 9 9 9 9 11 EVI 2 2 2 2 2 2 2 2 2 1 1 1 2 2 2 2 2 2 2 2 2 2 1 1 1 1 1 1 2 2 1 # 10 67 48 35 44 68 38 65 3 12 52 78 29 51 64 62 40 25 31 Region Beira Interior Tejo Palmela Lagoa Montilla-Moriles Terra Alta Málaga & Sierras de Málaga Tavira Alentejo Binissalem Portimão Valencia Empordà Pla i Llevant Tarragona Setúbal Manzanilla Condado de Huelva Jerez CatI 9 9 9 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 14 SoilT 11 12 12 3 3 3 3 3 5 9 9 9 9 9 9 12 13 13 3 EVI 2 1 2 1 1 1 2 2 2 1 1 1 2 2 2 2 1 2 1 Chapter 7. Integrated analysis of climate, soil, topography and vegetative growth in Iberian viticultural areas 7.4. Discussion and conclusions In the current study, an integrated analysis of the climate, soil, topography and vegetative growth is undertaken for the Iberian viticultural areas, using state-of-the-art datasets. Until present, studies on viticultural zoning were mainly focused on climatic conditions (e.g. Jones and Davis 2000b; Fraga et al. 2013; Duchene and Schneider 2005; Webb et al. 2008b), while the combination of the terroir composing elements is still underexplored. While in the last decades great advances have been made regarding the quality and availability of these spatial datasets (Jones et al. 2010), few studies have been devoted to integrating these factors into viticultural zoning (Yau et al. 2013; van Leeuwen et al. 2004; Taylor 2004) and none for Iberia. Thus, understanding the spatial variability of these factors provides the basis for a viable characterization of each viticultural region. To our knowledge this is the first study in which climate, soil, topography and vegetative growth, are studied jointly to analyse the viticultural region in the Iberian Peninsula. Overall, temperate, dry climate, with cool nights (CatI-5) is the dominant climatic category for vineyards in Iberia. The DO regions with this climate tend to present lower vigour. Nevertheless, these conditions of moderate water stress and high diurnal temperature ranges during ripening are often beneficial for the production of high quality wines (Fraga et al. 2014b; Jones et al. 2004b; van Leeuwen et al. 2009). This may in fact explain some of the variability behind vineyard settlements, as grapevine quality attributes is usually favoured over yield and vegetative growth. Other temperate climate types (CatI-6, 7 and 8) tend to present higher levels of vegetative growth, regardless of soil type, topography or water availability. In warmer and cooler climates water availability is more important for vegetative growth than in temperate climates. This is in fact corroborated by other studies (Chaves et al. 2007; Chaves et al. 2010; Lopes et al. 2011; Koundouras et al. 1999; Zsofi et al. 2011). Results suggest soil plays a key role in grapevine growth and development. Soils with higher clay content seem to induce lower growths. Although clay soils retain large amounts of water, grapevine water uptake is more difficult, resulting in increased water stress (Tramontini et al. 2013). In opposition, soils with higher sand and lower clay contents seem to provide a better growing structure for roots, resulting in higher absorption capacity for water and mineral nutrients. Although these soils retain less water than clay soils, they are better drained, which may in part explain this outcome. Loamy soils are the most widespread in Iberia. Grapevines growing in these soils seem to depend much more on climatic factors, 140 Chapter 7. Integrated analysis of climate, soil, topography and vegetative growth in Iberian viticultural areas mainly on water availability, than in other soil types. As such, water availability becomes the leading element to govern vegetative growth in loamy soils. This enhances the importance of the type of soils in determinant for the wine characteristics and quality of each region. Topography seems to play only a minor role regarding the vegetative growth in Iberia. Although a clear distinction is apparent between lower and higher elevations, this can might explained by the proximity to coastal areas with enhanced Atlantic influence (Fraga et al. 2014b). For geographical aspect, no conclusions can be drawn. Results show the relative influence of climate, soil and topography on grapevine growth. Many previous studies indicate the predominant role played by climate on grapevine development (e.g. van Leeuwen et al. 2004; Fraga et al. 2014b). Other studies suggest that soil texture can influence grapevine physiological attributes (Carey et al. 2008) and may determine the spatial heterogeneity in grapevine biomass (Winkel et al. 1995). In the present approach, it is also clear that vineyards in Iberia are grown in a large range of terroir conditions. When considering the mesoscale patterns over Iberia, climate seems to play a more important role than soil (Table S2). However, this influence is balanced when analysing only the current vineyard areas. When comparing the characteristics of the DO regions, some similarities between them are clear. As an example the renowned Douro DO shows predominantly the same climatic, soil and growth categories as seen in Alicante, Almansa, Cariñena, Priorat, Trás-os-Montes, Uclés, Utiel-Requena and Yecla. Conversly, some regions succeed in maintaining their singularity regarding these factors, such as Jerez and Rueda. It should be noted that this assessment does not directly include other important terroir elements, such as the oenological and viticultural practices. However , some of these factors are inherent to the current assessment, as over the centuries winemakers have gradually adapted to best suite their local environmental conditions (Jones 2012). The assessments provided herein may be of great value to viticulturists and may also play a key role when including vineyards into a given DO. Usual methods for DO delineation rely on onsite analysis of climate, soil and topographic attributes. Although these traditional approaches are extremely useful, they are often based on erratic, insufficient or unreliable data (e.g. outdated land cover, soils surveys and topographic maps, assessments made on nearby weather station records), making comparison between regions rather difficult. Furthermore, taking into account the climate change projections (Fraga et al. 2014a; Fraga et al. 2013; Malheiro et al. 2010; Jones et al. 2005a), these delineations may require and more continuous update. Our fully integrated approach provides a feasible method for DO comparison on a 141 Chapter 7. Integrated analysis of climate, soil, topography and vegetative growth in Iberian viticultural areas mesoscale basis. This may allow growers to identify new management practices and grapevine varieties that can be easily adapted to other regions that share the same terroir elements. Additionally, the methodologies implemented can be extended to other viticultural regions of the world. 142 Chapter 8. Synthesis and final conclusions 143 Chapter 8. Final conclusions and future work perspectives 144 Chapter 8. Final conclusions and future work perspectives 8.1. Introduction The current research intended to expand the knowledge on how climate influences grapevine development, wine production and quality attributes, such, while accessing potential climate change impacts. In this final chapter, a synthesis of the performed research is provided, along with the final conclusions. To do so, it becomes essential provide a comprehensive and integrated overview of the work accomplished. As shown, the climatic forcing is one of the main factors that influence winegrape geographical distribution, physiology and development, and yield and quality attributes. Bearing this in mind, climate change can significantly affect the already delicate balance in the soil-plant-atmosphere continuum and thus, research on this field has shown to be of the utmost importance. Chapter 2 through 7, describe a planned research that started in a regionalscale assessment of viticultural zoning in Europe, moving on to a local-scale assessment of the main viticultural regions in Portugal using a very-high resolution dataset developed for this purpose, and case-studies of the wine production and phenological development in very specific winemaking regions in Portugal. Additionally, a viticultural zoning of the DO regions in Iberia is assessed taking into account the vineyard climate, soils, topography and vegetative growth. All the elements that contributed to the construction of the current doctoral thesis are thus represented in the Figure 8.1. Observational data Soil, topography and remote sensing Yield and phenological development Regional Climate models Viticultural zoning Spatial pattern downscaling Bioclimatic indices Weather regimes Fig. 8.1 – Schematic representation of the main topics that contributed to the thesis. 145 Chapter 8. Final conclusions and future work perspectives The abovementioned elements, allow a deeper understanding of winegrape growing conditions in Europe, and especially in Portugal, while also allowing a direct comparison between regions with different viticultural suitability and assessing the impacts of climate change on this crop (variability in yield and in quality). 8.2. Summary of the research findings The main conclusions can be summarized as follows: 1. The use of bioclimatic indices enables the characterization of different winemaking regions, at both regional and local scales. The analysis of the selected bioclimatic indices ensures the selection of the most suitable zones for winegrape growth in Europe, and enables to infer on the regional potential for wine quality, risk of contamination by pests and diseases, water demands, amongst other key issues. 2. According to historical/observational data applied to viticultural bioclimatic indices, it is clear that many countries in southern Europe already present a high thermal accumulation, with potential implications in grapevine maturity and wine quality. Some of these regions, with a Mediterranean-like climate, already present moderate to high levels of dryness and warming, which resulted in the introduction of some adaptation practices, such as irrigation (e.g. southern Iberia, southern Italy). Future projections, based on a multi-model ensemble, showed higher thermal accumulation throughout Europe. In southern Europe, projections point to future drying and warming, resulting in additional threats to the viticultural sector. Conversely, in northern and central Europe, warm and moist climates may result in higher risks for pests and diseases. In the latter regions, an enhanced thermal accumulation can actually increase the suitability of new winemaking regions, with implications on wine quality. Inter-annual variability in these bioclimatic indices, are also expected to increase in the future throughout Europe, resulting in a higher fluctuation in both yield and quality variables. All these projections present a certain level of uncertainty, inherent to the climate modelling approaches. These uncertainties remain higher in precipitation-based indices than in temperature-based indices, especially in the Mediterranean-like climatic regions. Despite the abovementioned uncertainties, these projections are quite robust due to the use of multi-model ensemble. Changes in viticultural suitability in Europe are thus expected under future climates with a high confidence. 146 Chapter 8. Final conclusions and future work perspectives 3. Focusing on the Portuguese winemaking regions, significant increases in the thermal accumulation and dryness in the growing season are expected to occur in the next decades, particularly over the south and innermost regions. Additionally an intensification of extreme weather events could result in an increased inter-annual variability in yield and quality attributes. Results indicate that a reshaping of the main Portuguese winemaking regions is likely to occur in the upcoming decades, emphasizing the need for the development of appropriate climate change adaptation measures, in order to preserve wine styles. 4. The use of high spatial resolution datasets has demonstrated to be of great importance in viticultural zoning research, since it is necessary to capture the effect of environmental variability on this crop. The very-high resolution categorized index (CatI) developed in this study acts as a reference index, which can be used for regional characterization of the viticultural regions. In future climate conditions, the viticultural zoning is projected to undergo significant changes. Results depict the Atlantic/Mediterranean climatic contrast over Portugal and allow capturing the current vineyard spatial distribution. Nonetheless, under human-induced climate change, increased thermal accumulation and dryness may result in loss of viticultural suitability, lower bioclimatic diversity and earlier phenological events. These changes may lead to changes in varietal selection and wine characteristics of each region. These climate change projections are robust across the different climate models, and highlight future changes in grapevine development and wine quality parameters of each region. These outcomes are particularly relevant for decision-makers in the winemaking sector, as this very-high resolution illustrates the small scale changes in the spatial variability of each region. 5. Several climatic elements have shown have great impact on yield and wine production. Results indicate that a moderate water stress during the growing season, high production 3-yrs before, cool weather in February-March, settled-warm weather in May, warm moist weather in June and relatively cool conditions preceding harvest are generally favourable to a high wine production. Our study highlights the key role played by precipitation and humidity levels in wine production. The linkage between large scale atmospheric circulation/patterns and wine production was also demonstrated. Furthermore, the developed logistic model provides a high skill in predicting wine production, which could prove to be of great value to the winemaking sector. 147 Chapter 8. Final conclusions and future work perspectives 6. This study also confirms the existence of significant links between climate and grapevine development stages. Temperatures preceding certain grapevine phenological stages have shown to have great influence on the onset and duration of these stages. Flowering as shown to be the most climatic sensitive phenophase, since its timings have shown to be deeply tied to air temperature in the preceding months (March-April). The developed statistical models show that this climatic forcing can explain most of the variability in phenological development. The increase in temperature during the study period, has also shown to be keyed to the advances in phenological timings. Given these significant trends, shifts may occur in the long term, highlighting the need to assess varietal characteristics and responses to climate change. 7. Climate has also shown to play a leading role for grapevine vegetative growth, when compared to other grapevine-influencing elements such as soil type and topography. Results show that most Iberian vineyards are grown in temperate dry climates with loamy soils, presenting low vegetative growth. Regarding grapevine growth, water availability has shown to be of be of utmost importance, especially in warm or cool climatic conditions. As such, taking into account the climate change projections provided in this thesis, the development of appropriate mitigation and adaptation measures if of utmost importance to cope with future impacts on this crop and improve sustainability of the winemaking sector. All these links can prove to be of major importance to the regional viticultural sector, by enabling the implementation of suitable strategies to cope with the inter-annual variability in both grapevine development and wine production and are particularly pertinent when planning the viticultural activities throughout the year. Furthermore, these assessments can support timely projections that can enable a suitable array of adaptation measures to cope with the inter-annual variability associated to viticultural practices. The assessments provided herein have demonstrated to be of crucial importance to the winemaking sector, as this knowledge may allow decision-makers to develop, and potentially adopt, suitable adaptation and mitigation measures that ensure the future sustainability of this sector. Also, the awareness to the climate change impacts in viticulture may also allow reducing the sector vulnerability. These results are particularly important for the winemaking sector in Europe and especially in Portugal. 148 Chapter 8. Final conclusions and future work perspectives 8.3. Adaptation measures: considerations at variety level Several adaptation/mitigation measures have been suggested throughout this study. Some of them can be implemented on a short-term basis such as, irrigation management, soil/tillage treatments and canopy microclimatic control, while other would require a much greater effort and time, namely the varietal and rootstock selection. Although these adaptation measures were already discussed throughout the thesis, the varietal suitability and adaptation to future climates requires special attention. In Portugal the large number of native varieties of Vitis vinifera L. (341; Portaria nº 428/2000 de 17 de Julho, changed by Portaria nº 380/2012 de 22 de Novembro), are considered a valuable asset due to their adaptive capacity different regions and climates. Therefore, based on the methodologies and results provided in this thesis, the adaptive capacity of the Portuguese grapevine varieties are now assessed and discussed, taking into account climate change impacts. Over the time, genetic mutations, coupled with environmental forcing and geographical spread, lead to differentiation of Vitis vinifera L. in several varieties, increasing the genetic heritage of this species. However, the geographic expansion, based on a global commercial scale, focused mainly on varieties of French origin (e.g. Cabernet-Sauvignon, Pinot-Noir), which are currently among the most renowned and most used in winemaking. However, these varieties have very different characteristics from the native Portuguese varieties, which have unique characteristics that greatly contribute to the distinctiveness of the national wines. Although each variety has its own unique characteristics, the same variety can produce wines with different properties depending on the region where it is grown, in particular taking into account the existing climatic conditions. As was previously shown, grapevines have remarkable adaptation strategies to various climatic conditions. Therefore, the concept of regional plasticity, i.e. the ability of each variety to adapt to different regions/climatic conditions, it is indeed important. Despite the effort to identify the climatic conditions in which some varieties are grown in specific regions (Lopes et al. 2008; Malheiro et al. 2013), the lack of a multi-regional assessment is still a major shortcoming in identifying the most suitable grapevine varieties to a future climate. The very-high resolution climatic information provided in this thesis, combined with the location of the most important grapevine varieties in Portugal (Böhm 2010), allows a multi-regional assessment of the thermal conditions in which these varieties are grown (Growing Degree Day – GDD; Fig. 8.2). 149 Chapter 8. Final conclusions and future work perspectives As expected, white varieties exhibit the lowest GDD mean values. However, they also show the largest ranges of GDD, with particular emphasis on the Alvarinho and Bical varieties, which attests to their high regional plasticity. Moreover, white varieties also present the highest GDD mean values, through Síria and Antao-Vaz varieties. These results are somewhat surprising, since red varieties are traditionally considered better suited to warmer climates. Nonetheless, these last two white varieties present a low regional plasticity, which may indicate that they are perfectly adapted to their current climatic conditions. Amongst the red varieties, Touriga-Franca and Tinto-Cão show a very high regional plasticity, while Malvasia-Preta shows the lowest GDD mean values. However, this latter variety presents a high GDD range, which suggests a high adaptability to different climatic conditions. The red variety with the lowest regional plasticity is Jaen (originally from the region of Bierzo, Spain), reflecting its relatively recent introduction into the national territory and little genetic variability. Also noteworthy are the Touriga-Nacional, Aragonez, Trincadeira and Castelão, presenting GDD values that fall in the optimal thermal range for obtaining high quality wines (Fraga et al. 2014a). While it is clear that some varieties have a higher adaptability to changes in climate then others, most Portuguese varieties indeed show a high range of GDD values (Fig. 8.2), which proves the high adaptability of Portuguese grapevine varieties to a wide range of climates and regions (1150-2300ºC). The projected future warming can increase GDD levels by about 600°C until 2070 (Chapters 2-4). It is likely that varieties with low GDD values and low regional plasticity face further challenges adapting to the future climates. In this category are varieties such as Azal, Avesso, Loureiro and Trajadura (white), and also Borraçal, Vinhão, Espadeiro and, to some extent, Jaen (red). It is expected that these varieties are less used/grown in the future, as they can achieve much faster or unbalanced ripening, leading to the production of low quality wines. Nonetheless, the selection of future varieties will be very dependent on the interest by oenologists/winemakers in their maintenance, for the production of specific wines. Given this context, the existing cultural practices may be used to adapt the negative impact of climate change. 150 Chapter 8. Final conclusions and future work perspectives Fig. 8.2 - Growing degree day (ºC) in which some grapevine varieties are grown in Portugal, calculated using the WorldClim dataset using the varietal spatial distribution. Medians correspond to horizontal black lines within boxes, lower (upper) box limits to the minimum and maximum of the GDD value. Other varieties, such as Síria, Antão-Vaz, Moscatel-Graúdo (white) and Alfrocheiro (red), despite having relatively low regional plasticity, are expected to have increased demand in the future, due to their good adaptation to warmer climates. Red varieties such as Touriga- 151 Chapter 8. Final conclusions and future work perspectives Franca, Touriga-Nacional and Aragonez (Tinta Roriz/Tempranillo) should remain as the elite varieties for winemaking in Portugal, due to their oenological characteristics and good adaptation to warm climates. It will also be interesting to verify the future suitability of white grapes with high regional plasticity, such as Alvarinho and Bical. Despite the existence of a large number of varieties worldwide, it is known that the market demand falls primarily on a very narrow group of varieties (e.g. Cabernet Sauvignon, Chardonnay, Merlot, Pinot-Noir, Aragonez/Tempranillo, Riesling). For the reasons described above, the Portuguese varieties offer an advantage, since they are already grown in conditions of high heat accumulation, when compared with other winemaking regions in other world (Fraga et al. 2013). This aspect, combined with the high regional plasticity/climatic adaptability of most of the Portuguese varieties and their high quality, can result in the expansion of Portuguese grapes to other regions in the world. However, the very strict regulations, specific to each wine region, regarding the varieties used in winemaking, and the oenological trends, can pose a strong resistance to this expansion. 8.4. Final remarks and future work The conclusion of this thesis fully complies with the planned objectives. This can be attested by the scientific productivity: 7 articles accepted or already published in international scientific journals, 14 communications in international meetings/conferences, including several oral presentations and 3 technical articles published in national journals of the winemaking sector. Lectures/seminars on some thesis topics were also given in the Oenology and in Agronomic Engineering graduation courses of the UTAD. The vast know-how acquired, during this doctoral course, in agricultural/viticultural sciences, computational sciences and programing (e.g. Matlab®, Python®), geographical information systems (e.g. ArcGis®), statistical analysis (e.g. SPSS®, R®) and methodologies, field trials, data mining and in human/social skills, were key points for the conclusion of this thesis. Furthermore, on a personal level, since this is my region of origin, I hope the results provided could prove useful for the winemaking sector and decision-makers in the Douro/Porto DO region, which is expected to face further threats due to climate change. Future work, as a Post-Doc, is now being prepared. New methodologies for a higher spatial resolution viticultural zoning could be achieved in the future. By integrating soil characteristics and other topographic elements at a very high spatial resolution we could 152 Chapter 8. Final conclusions and future work perspectives provide a more robust viticultural zoning on a microscale level. This could lead to new statistical modelling of grapevine yield attributes and development parameters. Furthermore, in the next years, the development of new future scenarios and advances in climate model simulations can improve our results. Additionally, based on our previous experience in grapevine zoning and its climate change projections for Europe and Portugal, as well as on grapevine yield and phenological modelling, we aim at applying these innovative methodologies to other Mediterranean crops. Our purpose is to apply similar methodologies to the most important crops in Portugal and to assess the impacts of climate change on them. Even though the entire methodology was already successfully tested for grapevine, adjustments to other crops are required, including the selection of proper bioclimatic indices for each crop. Furthermore, new climate model simulations, emission scenarios and crop models need to be tested and validated. These issues are vital for a strategic planning of the Portuguese economy, particularly in a highly and increasingly competing world and can be an opportunity to develop eco-innovative procedures towards a more sustainable agriculture. 153 154 References 155 156 AEMET/IMP (2011) Iberian Climate Atlas. Closas-Orcoyen S. L., Madrid, Spain Alexandrov VA, Hoogenboom G (2000) The impact of climate variability and change on crop yield in Bulgaria. Agric For Meteorol 104 (4):315-327. doi:10.1016/S01681923(00)00166-0 Alig RJ, Adams DM, McCarl BA (2002) Projecting impacts of global climate change on the US forest and agriculture sectors and carbon budgets. For Ecol Manag 169 (1-2):3-14. doi:10.1016/S0378-1127(02)00290-6 Alley R, Hewitson B, Hoskins BJ, Joos F, Jouzel J, Kattsov V, Lohmann U, Manning M, Matsuno T, Molina M, Nicholls N, Berntsen T, Overpeck J, Qin D, Raga G, Ramaswamy V, Ren J, Rusticucci M, Solomon S, Somerville R, Stocker TF, Stott PA, Bindoff NL, Stouffer RJ, Whetton P, Wood RA, Wratt D, Chen Z, Chidthaisong A, Friedlingstein P, Gregory JM, Hegerl GC, Heimann M (2007) Summary for policymakers. In Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change [Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M. Tignor and H.L. Miller (eds.)]. In. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA., Alonso AD, O'Neill MA (2011) Climate change from the perspective of Spanish wine growers: a three-region study. Br Food J 113 (2-3):205-221. doi:10.1108/00070701111105303 Alves F, Costa J, Costa P, Correia CM, Gonçalves BM, Soares R, Moutinho-Pereira J (2012) Influence of climate and deficit irrigation on grapevine physiology, yield and quality attributes, of the cv. Touriga Nacional at Douro Region. Proceeding of the IX International Terroirs Congress,25-29 June, Dijon-Reims, France Amerine MA, Winkler AJ (1944) Composition and quality of musts and wines of California grapes, vol 15. Hilgardia, California, USA Anderson JD, Jones GV, Tait A, Hall A, Trought MCT (2012) Analysis of Viticulture Region Climate Structure and Suitability in New Zealand. J Int Sci Vigne Vin 46 (3):149-165 Andrade C, Leite SM, Santos JA (2012) Temperature extremes in Europe: overview of their driving atmospheric patterns. Nat Hazard Earth Sys 12 (5):1671-1691. doi:10.5194/nhess-12-1671-2012 Archer E, Strauss HC (1990) The effect of vine spacing on some physiological aspects of Vitis vinifera L. (cv. Pinot noir). S Afr J Enol Vitic (10):49–58 Ashenfelter O, Storchmann K (2010) Measuring the Economic Effect of Global Warming on Viticulture Using Auction, Retail, and Wholesale Prices. Rev Ind Organ 37 (1):51-64. doi:10.1007/s11151-010-9256-6 Austin ME, Bondari K (1988) A Study of Cultural and Environmental-Factors on the Yield of Vitis-Rotundifolia. Sci hort 34 (3-4):219-227. doi:10.1016/0304-4238(88)90095-7 Bahar E, Yasasin AS (2010) The yield and berry quality under different soil tillage and clusters thinning treatments in grape (Vitis vinifera L.) cv. Cabernet-Sauvignon. Afr J Agr Res 5 (21):2986-2993 Basile B, Girona J, Behboudian MH, Mata M, Rosello J, Ferre M, Marsal J (2012) Responses of "Chardonnay" to deficit irrigation applied at different phenological stages: vine growth, must composition, and wine quality. Irrig Sci 30 (5):397-406. doi:10.1007/s00271-012-0353-1 Bates TR, Dunst RM, Joy P (2002) Seasonal dry matter, starch, and nutrient distribution in 'Concord' grapevine roots. HortScience 37 (2):313-316 157 Battaglini A, Barbeau G, Bindi M, Badeck FW (2009) European winegrowers' perceptions of climate change impact and options for adaptation. Reg Environ Change 9 (2):61-73. doi:10.1007/s10113-008-0053-9 Berry J, Bjorkman O (1980) Photosynthetic Response and Adaptation to Temperature in Higher-Plants. Annu Rev Plant Physiol Plant Mol Biol 31:491-543. doi:10.1146/annurev.pp.31.060180.002423 Bertin RI (2009) Plant phenology and distribution in relation to recent climate change. The Journal of the Torrey Botanical Society 135 (1):126-146 Bindi M, Fibbi L, Gozzini B, Orlandini S, Miglietta F (1996) Modelling the impact of future climate scenarios on yield and yield variability of grapevine. Clim Res 7 (3):213-224. doi:10.3354/Cr007213 Bindon K, Dry P, Loveys B (2008) Influence of partial rootzone drying on the composition and accumulation of anthocyanins in grape berries (Vitis vinifera cv. Cabernet Sauvignon). Aust J Grape Wine R 14 (2):91-103. doi:10.1111/j.17550238.2008.00009.x Blanco-Ward D, Queijeiro JMG, Jones GV (2007) Spatial climate variability and viticulture in the Mino River Valley of Spain. Vitis 46 (2):63-70 Bock A, Sparks T, Estrella N, Menzel A (2011) Changes in the phenology and composition of wine from Franconia, Germany. Clim Res 50 (1):69-81. doi:10.3354/Cr01048 Böhm J (2010) Portugal vitícola: o grande livro das castas. C. Ferreira, Lisbon, Portugal Böhm U, Kücken M, Ahrens W, Block A, Hauffe D, Keuler K, Rockel B, Will A (2006) CLM- The Climate Version of LM: Brief Description and Long- Term Applications. COSMO Newsletter 6:225-235 Branas J (1974) Viticulture. Dehan, Montpellier, France Branas J, Bernon G, Levadoux L (1946) Éléments de viticulture générale. impr. Delmas, Montpellier, France Brewer CA (1992) Responses by stomata on leaves to microenvironmental conditions. Proceedings of the 13th Workshop/Conference of the Association for Biology Laboratory Education (ABLE), 191 pages (C A Goldman, Editor) 13:67-77 Bruinsma J (2009) The Resource Outlook to 2050: By how much do land, water and crop yields need to increase by 2050? Food and Agriculture Organization, Rome. Bureau SM, Razungles AJ, Baumes RL (2000) The aroma of Muscat of Frontignan grapes: effect of the light environment of vine or bunch on volatiles and glycoconjugates. J Sci Food Agric 80 (14):2012-2020. doi:10.1002/1097-0010(200011)80:14<2012::AidJsfa738>3.0.Co;2-X Burg JP (1975) Maximum entropy spectral analysis. Stanford University, Stanford, California Butt TM, Copping LG (2000) Fungal biological control agents. Pestic Outlook 11 (5):186-191 Büttner G, Feranec G, Jaffrain G (2006) EEA CORINE Land Cover Nomenclature Illustrated Guide - Addendum 2006. Buttrose MS, Hale CR, Kliewer WM (1971) Effect of Temperature on Composition of Cabernet Sauvignon Berries. Am J Enol Vitic 22 (2):71-75 Caetano M, Mata F, Freire S (2006) Accuracy assessment of the Portuguese CORINE land cover map. Global Developments in Environmental Earth Observation from Space:459-467 Caffarra A, Eccel E (2011) Projecting the impacts of climate change on the phenology of grapevine in a mountain area. Aust J Grape Wine R 17 (1):52-61. doi:10.1111/j.17550238.2010.00118.x 158 Camps JO, Ramos MC (2012) Grape harvest and yield responses to inter-annual changes in temperature and precipitation in an area of north-east Spain with a Mediterranean climate. Int J Biometeorol 56 (5):853-864. doi:10.1007/s00484-011-0489-3 Carbonneau A (2003) Ecophysiologie de la vigne et terroir. In: Terroir, zonazione, viticoltura. Trattato internazionale. Phytoline, pp 61-102 Cardoso RM, Soares PMM, Miranda PMA, Belo-Pereira M (2013) WRF high resolution simulation of Iberian mean and extreme precipitation climate. Int J Clim 33 (11):25912608. doi:10.1002/joc.3616 Carey VA, Archer E, Barbeau G, Saayman D (2008) Viticultural Terroirs in Stellenbosch, South Africa. Ii. The Interaction of Cabernet-Sauvignon and Sauvignon Blanc with Environment. J Int Sci Vigne Vin 42 (4):185-201 Celette F, Gaudin R, Gary C (2008) Spatial and temporal changes to the water regime of a Mediterranean vineyard due to the adoption of cover cropping. Eur J Agron 29 (4):153-162. doi:10.1016/j.eja.2008.04.007 Centeno A, Baeza P, Lissarrague JR (2010) Relationship between Soil and Plant Water Status in Wine Grapes under Various Water Deficit Regimes. Horttechnology 20 (3):585-593 Chalmers Y, Downey M, Krstic M, Loveys B, Dry PR (2010) Influence of sustained deficit irrigation on colour parameters of Cabernet Sauvignon and Shiraz microscale wine fermentations. Aust J Grape Wine R 16 (2):301-313. doi:10.1111/j.17550238.2010.00093.x Chaves MM, Santos TP, Souza CR, Ortuno MF, Rodrigues ML, Lopes CM, Maroco JP, Pereira JS (2007) Deficit irrigation in grapevine improves water-use efficiency while controlling vigour and production quality. Ann Appl Biol 150 (2):237-252. doi:10.1111/j.1744-7348.2006.00123.x Chaves MM, Zarrouk O, Francisco R, Costa JM, Santos T, Regalado AP, Rodrigues ML, Lopes CM (2010) Grapevine under deficit irrigation: hints from physiological and molecular data. Ann Bot 105 (5):661-676. doi:10.1093/aob/mcq030 Christensen JH, Christensen OB, Lopez P, van Meijgaard E, Botzet M (1996) The HIRHAM 4 regional atmospheric climate model. DMI Scientific Report 96-4, Copenhagen, Denmark Christensen JH, Hewitson B, Busuioc A, Chen A, Gao X, Jones R, Kolli RK, Kwon WT, Laprise R, Magaña Rueda V, Menéndez CG, Räisänen J, Rinke A, Sarr A, Whetton P (2007) Regional Climate Projections. In: Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change [Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M. Tignor and H.L. Miller (eds.)]. In. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA., pp 847-940 Christensen JH, Kjellstrom E, Giorgi F, Lenderink G, Rummukainen M (2010) Weight assignment in regional climate models. Clim Res 44 (2-3):179-194. doi:10.3354/Cr00916 Chuine I, Kramer K, Hanninen H (2003) Plant development models. In: Phenology - An Integrative Environmental Science, Tasks for vegetation science 39. Ed. Schwartz M.D., Kluwer Academic Publishers. p. 217-235. Chuine I, Yiou P, Viovy N, Seguin B, Daux V, Ladurie EL (2004) Historical phenology: Grape ripening as a past climate indicator. Nature 432 (7015):289-290. doi:10.1038/432289a Cifre J, Bota J, Escalona JM, Medrano H, Flexas J (2005) Physiological tools for irrigation scheduling in grapevine (Vitis vinifera L.) An open gate to improve water-use efficiency? Agric Ecosyst Environ 106 (2-3):159-170. doi:10.1016/j.agee.2004.10.005 159 Collins M, Booth BB, Bhaskaran B, Harris GR, Murphy JM, Sexton DMH, Webb MJ (2011) Climate model errors, feedbacks and forcings: a comparison of perturbed physics and multi-model ensembles. Clim Dynam 36 (9-10):1737-1766. doi:10.1007/s00382-0100808-0 Coombe BG (1987) Influence of temperature on composition and quality of grapes. Acta hort 206:23-36 Costa AC, Santos JA, Pinto JG (2012) Climate change scenarios for precipitation extremes in Portugal. Theor Appl Clim 108 (1-2):217-234. doi:10.1007/s00704-011-0528-3 Crespy A (1987) Viticulture d'aujourd'hui. 2 edn. Lavoisier, Cachan, France Dalla Marta A, Grifoni D, Mancini M, Storchi P, Zipoli G, Orlandini S (2010) Analysis of the relationships between climate variability and grapevine phenology in the Nobile di Montepulciano wine production area. J Agric Sci 148:657-666. doi:10.1017/S0021859610000432 Daux V, Garcia de Cortazar-Atauri I, Yiou P, Chuine I, Garnier E, Le Roy Ladurie E, Mestre O, Tardaguila J (2011) An open-database of Grape Harvest dates for climate research: data description and quality assessment. Clim Past Discuss 7 (6):3823-3858. doi:10.5194/cpd-7-3823-2011 de Cortazar-Atauri IG, Brisson N, Gaudillere JP (2009) Performance of several models for predicting budburst date of grapevine (Vitis vinifera L.). Int J Biometeorol 53 (4):317326. doi:DOI 10.1007/s00484-009-0217-4 Denman KL, Brasseur G, Chidthaisong A, Ciais P, Cox PM, Dickinson RE, Hauglustaine D, Heinze C, Holland E, Jacob D, Lohmann U, Ramachandran S, da Silva Dias PL, Wofsy SC, Zhang X (2007) Couplings Between Changes in the Climate System and Biogeochemistry. In: Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change [Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M. Tignor and H.L. Miller (eds.)]. In. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA., pp 499-587 Deser C, Phillips A, Bourdette V, Teng HY (2012) Uncertainty in climate change projections: the role of internal variability. Clim Dynam 38 (3-4):527-546. doi:10.1007/s00382010-0977-x Dessai S, Hulme M (2007) Assessing the robustness of adaptation decisions to climate change uncertainties: A case study on water resources management in the East of England. Global Environ Chang 17 (1):59-72. doi:10.1016/j.gloenvcha.2006.11.005 Doll P (2002) Impact of climate change and variability on irrigation requirements: A global perspective. Clim Change 54 (3):269-293 Downey MO, Dokoozlian NK, Krstic MP (2006) Cultural practice and environmental impacts on the flavonoid composition of grapes and wine: A review of recent research. Am J Enol Vitic 57 (3):257-268 Drake BG, GonzalezMeler MA, Long SP (1997) More efficient plants: A consequence of rising atmospheric CO2? Annu Rev Plant Physiol Plant Mol Biol 48:609-639 Du TS, Kang SZ, Zhang JH, Li FS, Yan BY (2008) Water use efficiency and fruit quality of table grape under alternate partial root-zone drip irrigation. Agric Water Manag 95 (6):659-668. doi:10.1016/j.agwat.2008.01.017 Duchene E, Butterlin G, Dumas V, Merdinoglu D (2012) Towards the adaptation of grapevine varieties to climate change: QTLs and candidate genes for developmental stages. Theor Appl Genet 124 (4):623-635. doi:10.1007/s00122-011-1734-1 Duchene E, Huard F, Dumas V, Schneider C, Merdinoglu D (2010) The challenge of adapting grapevine varieties to climate change. Clim Res 41 (3):193-204. doi:10.3354/Cr00850 160 Duchene E, Schneider C (2005) Grapevine and climatic changes: a glance at the situation in Alsace. Agron Sustain Dev 25 (1):93-99. doi:10.1051/Agro:2004057 Due G, Morris M, Pattison S, Coombe BG (1993) Modeling Grapevine Phenology against Weather - Considerations Based on a Large Data Set. Agric For Meteorol 65 (1-2):91106. doi:Doi 10.1016/0168-1923(93)90039-K During H (1986) ABA and water stress in grapevines. Acta hort (179):413-420 Easterling WE, Aggarwal PK, Batima P, Brander KM, Erda L, Howden SM, Kirilenko A, Morton J, Soussana J-F, Schmidhuber J, Tubiello FN (2007) Food, fibre and forest products. Climate Change 2007: Impacts, Adaptation and Vulnerability. Contribution of Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, M.L. Parry, O.F. Canziani, J.P. Palutikof, P.J. van der Linden and C.E. Hanson [eds.]. In. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA., pp 273-312 EEA (2002) CORINE Land Cover update, I&CLC2000 project, Technical Guidelines. Eichorn KW, Lorenz DH (1977) Phaenologische Entwicklungsstadien der Rebe. Nachrichtembl. astsch. Pflanzenschutzdienstes. Brauschweig 29:119-120 Eitzinger J, Kubu G, Formayer H, Gerersdorfer T (2009) Climatic Wine Growing Potential under Future Climate Scenarious in Austria. Sustainable Development and Bioclimate: Reviewed Conference Proceedings:146-147 Elguindi N, Bi X, Giorgi F, Nagarajan B, Pal J, Solmon F, Rauscher S, Zakey A (2007) RegCM version 3.1 user’s guide. PWCG Abdus Salam ICTP, FAO (2006) World reference base for soil resources 2006, a framework for international classification, correlation and communication. World soil resources reports 103, Food and Agriculture Organization of the United Nations, Rome. FAO/IIASA/ISRIC/ISSCAS/JRC (2012) Harmonized World Soil Database (version 1.2). FAO, Rome, Italy and IIASA, Laxenburg, Austria. Ferreira MI, Silvestre J, Conceição N, Malheiro AC (2012) Crop and stress coefficients in rainfed and deficit irrigation vineyards using sap flow techniques. Irrig Sci 30 (5):433447. doi:10.1007/s00271-012-0352-2 Field SK, Smith JP, Holzapfel BP, Hardie WJ, Emery RJN (2009) Grapevine Response to Soil Temperature: Xylem Cytokinins and Carbohydrate Reserve Mobilization from Budbreak to Anthesis. Am J Enol Viticult 60 (2):164-172 Fischer G, Tubiello FN, van Velthuizen H, Wiberg DA (2007) Climate change impacts on irrigation water requirements: Effects of mitigation, 1990–2080. Technol Forecast Soc Change 74 (7):1083-1107. doi:10.1016/j.techfore.2006.05.021 Flexas J, Galmes J, Galle A, Gulias J, Pou A, Ribas-Carbo M, Tomas M, Medrano H (2010) Improving water use efficiency in grapevines: potential physiological targets for biotechnological improvement. Aust J Grape Wine R 16:106-121. doi:10.1111/j.17550238.2009.00057.x Fraga H, Malheiro AC, Moutinho-Pereira J, Jones GV, Alves F, Pinto JG, Santos JA (2014a) Very high resolution bioclimatic zoning of Portuguese wine regions: present and future scenarios. Reg Environ Change 14 (1):295-306. doi:10.1007/s10113-013-0490y Fraga H, Malheiro AC, Moutinho-Pereira J, Santos JA (2012a) An overview of climate change impacts on European viticulture. Food and Energy Security 1 (2):94-110. doi:10.1002/fes3.14 Fraga H, Malheiro AC, Moutinho-Pereira J, Santos JA (2013) Future scenarios for viticultural zoning in Europe: ensemble projections and uncertainties. Int J Biometeorol 57 (6):909-925. doi:10.1007/s00484-012-0617-8 161 Fraga H, Malheiro AC, Moutinho-Pereira J, Santos JA (2014b) Climate factors driving wine production in the Portuguese Minho region. Agric For Meteorol 185:26-36. doi:10.1016/j.agrformet.2013.11.003 Fraga H, Santos JA, Malheiro AC, Moutinho-Pereira J (2012b) Climate Change Projections for the Portuguese Viticulture Using a Multi-Model Ensemble. Ciência Téc Vitiv 27 (1):39-48 Gaal M, Moriondo M, Bindi M (2012) Modelling the Impact of Climate Change on the Hungarian Wine Regions Using Random Forest. Appl Ecol Environ Res 10 (2):121140 Garcia-Mozo H, Mestre A, Galan C, Contributors D (2010) Phenological trends in southern Spain: A response to climate change. Agric For Meteorol 150 (4):575-580. doi:DOI 10.1016/j.agrformet.2010.01.023 Garcia JG, Martinez-Cutillas A, Romero P (2012) Financial analysis of wine grape production using regulated deficit irrigation and partial-root zone drying strategies. Irrig Sci 30 (3):179-188. doi:10.1007/s00271-011-0274-4 Genovese G, Vignolles C, Negre T, Passera G (2001) A methodology for a combined use of normalised difference vegetation index and CORINE land cover data for crop yield monitoring and forecasting. A case study on Spain. Agronomie 21 (1):91-111 Gibelin AL, Deque M (2003) Anthropogenic climate change over the Mediterranean region simulated by a global variable resolution model. Clim Dynam 20 (4):327-339. doi:10.1007/s00382-002-0277-1 Glenn DM, Cooley N, Walker R, Clingeleffer P, Shellie K (2010) Impact of Kaolin Particle Film and Water Deficit on Wine Grape Water Use Efficiency and Plant Water Relations. HortScience 45 (8):1178-1187 Gobin A, Campling P, Feyen J (2001) Logistic modelling to identify and monitor local land management systems. Agric Sys 67 (1):1-20. doi:10.1016/S0308-521X(00)00043-3 Goncalves B, Falco V, Moutinho-Pereira J, Bacelar E, Peixoto F, Correia C (2009) Effects of elevated CO2 on grapevine (Vitis vinifera L.): volatile composition, phenolic content, and in vitro antioxidant activity of red wine. J Agric Food Chem 57 (1):265-273. doi:10.1021/jf8020199 Gouveia C, Liberato MLR, DaCamara CC, Trigo RM, Ramos AM (2011) Modelling past and future wine production in the Portuguese Douro Valley. Clim Res 48 (2):349-362. doi:10.3354/cr01006 Greer DH, Weedon MM, Weston C (2011) Reductions in biomass accumulation, photosynthesis in situ and net carbon balance are the costs of protecting Vitis vinifera 'Semillon' grapevines from heat stress with shade covering. AoB plants 2011:plr023. doi:10.1093/aobpla/plr023 Grifoni D, Carreras G, Zipoli G, Sabatini F, Dalla Marta A, Orlandini S (2008) Row orientation effect on UV-B, UV-A and PAR solar irradiation components in vineyards at Tuscany, Italy. Int J Biometeorol 52 (8):755-763. doi:10.1007/s00484-008-0168-1 Grifoni D, Mancini M, Maracchi G, Orlandini S, Zipoli G (2006) Analysis of Italian wine quality using freely available meteorological information. Am J Enol Vitic 57 (3):339346 Hall A, Jones GV (2009) Effect of potential atmospheric warming on temperature-based indices describing Australian winegrape growing conditions. Aust J Grape Wine R 15 (2):97-119. doi:10.1111/j.1755-0238.2008.00035.x Hall A, Jones GV (2010) Spatial analysis of climate in winegrape-growing regions in Australia. Aust J Grape Wine R 16 (3):389-404. doi:10.1111/j.17550238.2010.00100.x 162 Hann Jv (1883) Handbuch der Klimatologie. Bibliothek geographischer handbücher. J. Engelhorn, Stuttgart, Germany Hanson CE, Palutikof JP, Livermore MTJ, Barring L, Bindi M, Corte-Real J, Durao R, Giannakopoulos C, Good P, Holt T, Kundzewicz Z, Leckebusch GC, Moriondo M, Radziejewski M, Santos J, Schlyter P, Schwarb M, Stjernquist I, Ulbrich U (2007) Modelling the impact of climate extremes: an overview of the MICE project. Clim Change 81:163-177. doi:10.1007/s10584-006-9230-3 Harbertson JF, Keller M (2012) Rootstock Effects on Deficit-Irrigated Winegrapes in a Dry Climate: Grape and Wine Composition. Am J Enol Vitic 63 (1):40-48. doi:10.5344/ajev.2011.11079 Hardie WJ, Considine JA (1976) Response of Grapes to Water-Deficit Stress in Particular Stages of Development. Am J Enol Vitic 27 (2):55-61 Hardie WJ, Martin SR (2000) Shoot growth on de-fruited grapevines: a physiological indicator for irrigation scheduling. Aust J Grape Wine R 6 (1):52-58. doi:10.1111/j.1755-0238.2000.tb00162.x Hay LE, Wilby RJL, Leavesley GH (2000) A comparison of delta change and downscaled GCM scenarios for three mountainous basins in the United States. J Am Wat Resour Ass 36 (2):387-397. doi:10.1111/j.1752-1688.2000.tb04276.x Haykin SS (1991) Adaptive filter theory. Prentice Hall information and system sciences series, 2nd edn. Prentice Hall, Englewood Cliffs, NJ Haylock MR, Hofstra N, Klein Tank AMG, Klok EJ, Jones PD, New M (2008) A European daily high-resolution gridded data set of surface temperature and precipitation for 1950–2006. J geophys Res 113 (D20):D20119. doi:10.1029/2008jd010201 Hedberg PR, Mcleod R, Cullis B, Freeman BM (1986) Effect of Rootstock on the Production, Grape and Wine Quality of Shiraz Vines in the Murrumbidgee Irrigation Area. Aust J Exp Agric 26 (4):511-516. doi:10.1071/Ea9860511 Hegerl GC, Zwiers FW, Braconnot P, Gillett NP, Luo Y, Marengo Orsini JA, Nicholls N, Penner JE, Stott PA (2007) Under-standing and Attributing Climate Change. In Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change [Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M. Tignor and H.L. Miller (eds.)]. In. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA., Heinrich G, Gobiet A (2011) The future of dry and wet spells in Europe: A comprehensive study based on the ENSEMBLES regional climate models. Int J Clim:n/a-n/a. doi:10.1002/joc.2421 Hidalgo L (2002) Tratado de viticultura general. Mundi-Prensa Libros Spain, Hijmans RJ, Cameron SE, Parra JL, Jones PG, Jarvis A (2005) Very high resolution interpolated climate surfaces for global land areas. Int J Clim 25 (15):1965-1978. doi:10.1002/joc.1276 Hofstra N, Haylock M, New M, Jones PD (2009) Testing E-OBS European high-resolution gridded data set of daily precipitation and surface temperature. J Geophys Res 114 (D21). doi:10.1029/2009jd011799 Holton JR (2004) An introduction to dynamic meteorology. 4 edn. Elsevier Academic Press, Huete A, Didan K, Miura T, Rodriguez EP, Gao X, Ferreira LG (2002) Overview of the radiometric and biophysical performance of the MODIS vegetation indices. Remote Sens Environ 83 (1-2):195-213. doi:10.1016/S0034-4257(02)00096-2 163 Huglin P (1978) Nouveau mode d'évaluation des possibilités héliothermiques d'un milieu viticole. Comptes Rendus de l’Académie d’Agriculture. Académie d'agriculture de France, Paris, France Hulme M, Mahony M (2010) Climate change: What do we know about the IPCC? Prog phys Geogr 34 (5):705-718. doi:10.1177/0309133310373719 Intrieri C, Poni S, Rebucci B, Magnanini E (1998) Row orientation effects on whole-canopy gas exchange of potted and field-grown grapevines, vol 37. vol 4. JKI, Siebeldingen, Germany IPCC (2001) Climate Change 2001: Synthesis Report. A Contribution of Working Groups I, II, III to the Third Assessment Report of the Intergovernmental Panel on Climate Change, R.T.Watson and the Core Team, Eds. Cambridge University Press, Cambridge, UK IVV (2011) Vinhos e Aguardentes de Portugal, Anuário 2010/2011. Ministério da Agricultura, do Desenvolvimento Rural e das Pescas: Instituto da Vinha e do Vinho, Lisboa Jackson D (2001) Climate, monographs in cool climate viticulture - 2. Daphne Brasell NZ, Jackson DI, Lombard PB (1993) Environmental and Management-Practices Affecting Grape Composition and Wine Quality - a Review. Am J Enol Viticult 44 (4):409-430 Jacob D (2001) A note to the simulation of the annual and inter-annual variability of the water budget over the Baltic Sea drainage basin. Meteorol Atmos Phys 77 (1-4):61-73 Jacob D, Podzun R (1997) Sensitivity studies with the regional climate model REMO. Meteorol Atmos Phys 63 (1-2):119-129 Jaeger EB, Anders I, Luthi D, Rockel B, Schar C, Seneviratne SI (2008) Analysis of ERA40driven CLM simulations for Europe. Meteorol Z 17 (4):349-367. doi:10.1127/09412948/2008/0301 Jakob Themeßl M, Gobiet A, Leuprecht A (2011) Empirical-statistical downscaling and error correction of daily precipitation from regional climate models. Int J Clim 31 (10):1530-1544. doi:10.1002/joc.2168 Johnson LF, Bosch DF, Williams DC, Lobitz BM (2001) Remote sensing of vineyard management zones: Implications for wine quality. Appl Eng Agric 17 (4):557-560 Johnson LF, Roczen DE, Youkhana SK, Nemani RR, Bosch DF (2003) Mapping vineyard leaf area with multispectral satellite imagery. Comput Electron Agric 38 (1):33-44. doi:10.1016/S0168-1699(02)00106-0 Jones GV (2003) Winegrape phenology. In: Phenology - An Integrative Environmental Science, Tasks for vegetation science 39. Ed. Schwartz M.D., Kluwer Academic Publishers. p. 523-539. Jones GV (2004) Making wine in a changing climate. Geotimes 49 (8):24-28 Jones GV (2005) Climate change in the western united states grape growing regions. Acta Hortic 689:41-59 Jones GV (2006) Climate and Terroir: Impacts of Climate Variability and Change on Wine In Fine Wine and Terroir - The Geoscience Perspective. Macqueen, R.W., and Meinert, L.D., (eds.). Geoscience Canada, Geological Association of Canada, Newfoundland, Canada Jones GV (2012) A Climate Assessment for the Douro Wine Region: An Examination for the Past, Present and Future Conditions for Wine Production. ADVID, Peso da Régua, Portugal Jones GV, Alves F (2012a) Impact of climate change on wine production: a global overview and regional assessment in the Douro Valley of Portugal. Int J of Global Warming 4 (3/4):383-406 164 Jones GV, Alves F (2012b) Spatial Analysis of Climate in Winegrape Growing Regions in Portugal. Proceedings of the 9th International Terroir Congress, pp 3:1-4 25-29 June, 2012, Burgundy and Champagne, France Jones GV, Davis RE (2000a) Climate influences on grapevine phenology, grape composition, and wine production and quality for Bordeaux, France. Am J Enol Vitic 51 (3):249261 Jones GV, Davis RE (2000b) Using a synoptic climatological approach to understand climateviticulture relationships. Int J Clim 20 (8):813-837 Jones GV, Duchêne E, Tomasi D, Yuste J, Braslavska O, Schultz HR, Martinez C, Boso S, Langellier F, Perruchot C, Guimberteau G (2005a) Changes in European winegrape phenology and relationships with climate. Paper presented at the Proc. XIV GESCO Symposium, Geisenheim, Germany, August23–26 Jones GV, Duff AA, Hall A, Myers JW (2010) Spatial Analysis of Climate in Winegrape Growing Regions in the Western United States. Am J Enol Vitic 61 (3):313-326 Jones GV, Goodrich GB (2008) Influence of climate variability on wine regions in the western USA and on wine quality in the Napa Valley. Clim Res 35 (3):241-254. doi:10.3354/Cr00708 Jones GV, Snead N, Nelson P (2004a) Geology and wine 8. Modeling viticultural landscapes: A GIS analysis of the terroir potential in the Umpqua Valley of Oregon. Geosci Can 31 (4):167-178 Jones GV, White MA, Cooper O (2004b) Climate change and global wine quality. Bull Am met Soc 85 (4):504-504 Jones GV, White MA, Cooper OR, Storchmann K (2005b) Climate Change and Global Wine Quality. Clim Change 73 (3):319-343. doi:10.1007/s10584-005-4704-2 Judit G, Gabor Z, Adam D, Tamas V, Gyorgy B (2011) Comparison of three soil management methods in the Tokaj wine region. Mitt Klosterneuburg 61 (4):187-195 Juma NG (1999) Introduction to soil science and soil resources, Salman Productions, vol 1. 335pp. Junquera P, Lissarrague JR, Jimenez L, Linares R, Baeza P (2012) Long-term effects of different irrigation strategies on yield components, vine vigour, and grape composition in cv. Cabernet-Sauvignon (Vitis vinifera L.). Irrig Sci 30 (5):351-361. doi:10.1007/s00271-012-0348-y Keller M (2010a) Managing grapevines to optimise fruit development in a challenging environment: a climate change primer for viticulturists. Aust J Grape Wine R 16:5669. doi:10.1111/j.1755-0238.2009.00077.x Keller M (2010b) The Science of Grapevines: Anatomy and Physiology. Elsevier, Inc. 400pp. Keller M, Tarara JM (2010) Warm spring temperatures induce persistent season-long changes in shoot development in grapevines. Ann Bot 106 (1):131-141. doi:Doi 10.1093/Aob/Mcq091 Kenny GJ, Harrison PA (1992) The effects of climate variability and change on grape suitability in Europe. J Wine Res 3 (3):163-183. doi:10.1080/09571269208717931 Kistler R, Kalnay E, Collins W, Saha S, White G, Woollen J, Chelliah M, Ebisuzaki W, Kanamitsu M, Kousky V, van den Dool H, Jenne R, Fiorino M (2001) The NCEPNCAR 50-year reanalysis: Monthly means CD-ROM and documentation. Bull Am met Soc 82 (2):247-267 Kjellström E, Bärring L, Gollvik S, Hansson U, Jones C, Samuelsson P, Rummukainen M, Ullerstig A, WillØn U, Wyser K (2005) A 140-year simulation of European climate with the new version of the Rossby Centre regional atmospheric climate model 165 (RCA3). SMHI, Reports Meteorology and Climatology, 108, SMHI, SE-60176 Norrköping, Sweden Klein RJT, Huq S, Denton F, Downing TE, Richels RG, Robinson JB, Toth FL (2007) Interrelationships between adaptation and mitigation. Climate Change 2007: Impacts, Adaptation and Vulnerability. Contribution of Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, M.L. Parry, O.F. Canziani, J.P. Palutikof, P.J. van der Linden and C.E. Hanson [eds.]. In. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA., pp 745-777 Kliewer WM (1977) Effect of High-Temperatures during Bloom-Set Period on Fruit-Set, Ovule Fertility, and Berry Growth of Several Grape Cultivars. Am J Enol Vitic 28 (4):215-222 Kliewer WM, Soleiman A (1972) Effect of Chilling on Budbreak in Thompson Seedless and Carignane Grapevines. Am J Enol Vitic 23 (1):31-34 Kliewer WM, Torres RE (1972) Effect of Controlled Day and Night Temperatures on Grape Coloration. Am J Enol Vitic 23 (2):71-77 Knutti R, Allen MR, Friedlingstein P, Gregory JM, Hegerl GC, Meehl GA, Meinshausen M, Murphy JM, Plattner GK, Raper SCB, Stocker TF, Stott PA, Teng H, Wigley TML (2008) A review of uncertainties in global temperature projections over the twentyfirst century. J Clim 21 (11):2651-2663. doi:10.1175/2007jcli2119.1 Knutti R, Furrer R, Tebaldi C, Cermak J, Meehl GA (2010) Challenges in Combining Projections from Multiple Climate Models. J Clim 23 (10):2739-2758. doi:10.1175/2009jcli3361.1 Kottek M, Grieser J, Beck C, Rudolf B, Rubel F (2006) World map of the Koppen-Geiger climate classification updated. Meteorol Z 15 (3):259-263. doi:10.1127/09412948/2006/0130 Koundouras S, Tsialtas IT, Zioziou E, Nikolaou N (2008) Rootstock effects on the adaptive strategies of grapevine (Vitis vinifera L. cv. Cabernet-Sauvignon) under contrasting water status: Leaf physiological and structural responses. Agric Ecosyst Environ 128 (1-2):86-96. doi:10.1016/j.agee.2008.05.006 Koundouras S, Van Leeuwen C, Seguin G, Glories Y (1999) Influence of water status on vine vegetative growth, berry ripening and wine characteristics in mediterranean zone (example of Nemea, Greece, variety Saint-George, 1997). J Int Sci Vigne Vin 33:149160 Kroodsma DA, Field CB (2006) Carbon sequestration in California agriculture, 1980-2000. Ecol Appl 16 (5):1975-1985 Latif M, Anderson D, Barnett T, Cane M, Kleeman R, Leetmaa A, O'Brien J, Rosati A, Schneider E (1998) A review of the predictability and prediction of ENSO. Journal of Geophysical Research: Oceans 103 (C7):14375-14393. doi:10.1029/97JC03413 Lavee S, May P (1997) Dormancy of grapevine buds - facts and speculation. Aust J Grape Wine R 3 (1):31-46. doi:10.1111/j.1755-0238.1997.tb00114.x Le Treut H, Mauritzen C, Mokssit A, Peterson T, Prather M, Somerville R, Cubasch U, Ding Y (2007) Historical Overview of Climate Change. In Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. In. Cambridge University Press, Cambridge, UK and New York, NY, USA, pp 93-127 Lenderink G, van den Hurk B, van Meijgaard E, van Ulden A, Cuijpers H (2003) Simulation of present-day climate in RACMO2: first results and model developments. Ministerie van Verkeer en Waterstaat, Koninklijk Nederlands Meteorologisch Instituut, 166 Lobell DB, Field CB, Cahill KN, Bonfils C (2006) Impacts of future climate change on California perennial crop yields: Model projections with climate and crop uncertainties. Agric For Meteorol 141 (2-4):208-218. doi:10.1016/j.agrformet.2006.10.006 Lopes CM, Santos TP, Monteiro A, Rodrigues ML, Costa JM, Chaves MM (2011) Combining cover cropping with deficit irrigation in a Mediterranean low vigor vineyard. Sci hort 129 (4):603-612. doi:10.1016/j.scienta.2011.04.033 Lopes J, Eiras-Dias JE, Abreu F, Climaco P, Cunha JP, Silvestre J (2008) Thermal requirements, duration and precocity of phenological stages of grapevine cultivars of the Portuguese collection. Ciência Téc Vitiv 23 (1):61-71 Lorenz DH, Eichhorn KW, Bleiholder H, Klose R, Meier U, Weber E (1995) Growth Stages of the Grapevine: Phenological growth stages of the grapevine (Vitis vinifera L. ssp. vinifera)—Codes and descriptions according to the extended BBCH scale†. Aust J Grape Wine R 1 (2):100-103. doi:10.1111/j.1755-0238.1995.tb00085.x Lorenzo MN, Taboada JJ, Lorenzo JF, Ramos AM (2012) Influence of climate on grape production and wine quality in the Rías Baixas, north-western Spain. Reg Environ Change:1-10. doi:10.1007/s10113-012-0387-1 Mackenzie DE, Christy AG (2005) The role of soil chemistry in wine grape quality and sustainable soil management in vineyards. Wat Sci Technol 51 (1):27-37 Magalhães N (2008) Tratado de viticultura: a videira, a vinha e o terroir. Chaves Ferreira, Lisboa, Portugal Makra L, Vitanyi B, Gal A, Mika J, Matyasovszky I, Hirsch T (2009) Wine Quantity and Quality Variations in Relation to Climatic Factors in the Tokaj (Hungary) Winegrowing Region. Am J Enol Vitic 60 (3):312-321 Malheiro AC, Campos R, Fraga H, Eiras-Dias J, Silvestre J, Santos JA (2013) Winegrape phenology and temperature relationships in the Lisbon Wine Region, Portugal. J Int Sci Vigne Vin 47 (4):287-299 Malheiro AC, Santos JA (2011) Macrozonagem vitivinícola da Península Ibérica face a cenário de alteração climática. In: Atlas das castas da Península Ibérica - história, terroir, ampelografia. Coordenador Jorge Böhm, Dinalivro., pp 155 -159 Malheiro AC, Santos JA, Fraga H, Pinto JG (2010) Climate change scenarios applied to viticultural zoning in Europe. Clim Res 43 (3):163-177. doi:10.3354/cr00918 Malheiro AC, Santos JA, Fraga H, Pinto JG (2012) Future scenarios for viticultural climatic zoning in Iberia. Acta hort 931 (931):55-61 Mandelli F, Tonietto J, Hasenack H, Weber EJ (2005) Zoneamento Climático para a produção de uvas para vinhos de qualidade. Proceeding of Congresso Brasileiro de Agrometeorologia, 14, Campinas Manica I, Pommer CV (2006) Uva: do plantio a produção, pós-colheita e mercado. Porto Alegre: Continente. McCarthy MG (1999) Weight loss from ripening berries of Shiraz grapevines (Vitis vinifera L. cv. Shiraz). Aust J Grape Wine R 5 (1):10-16. doi:10.1111/j.17550238.1999.tb00145.x McKenzie N, Grundy M, Webster R, Ringrose-Voase A (2008) Guidelines for Surveying Soil and Land Resources, CSIRO PUBLISHING 2ed, 576pp. Mearns LO, Hulme M, Carter TR, Leemans M, Lal M, Whetton PH (2001) Climate scenario development. Climate Change 2001: The Scientific Basis. Contribution of Working Group I to the Third Assessment Report of the Intergovernmental Panel on Climate Change. In: Houghton JT, Ding Y, Griggs DJ et al. (eds). Cambridge University Press, Cambridge, UK ; New York, pp 739-768 167 Meehl GA, Stocker TF, Collins WD, Friedlingstein P, Gaye AT, Gregory JM, Kitoh A, Knutti R, Murphy JM, Noda A, Raper SCB, Watterson IG, Weaver AJ, Zhao Z-C (2007) Global Climate Projections. In: Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change [Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M. Tignor and H.L. Miller (eds.)]. In. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA., pp 747-845 Menzel A, Estrella N, Testka A (2005) Temperature response rates from long-term phenological records. Clim Res 30 (1):21-28. doi:Doi 10.3354/Cr030021 Menzel A, Seifert H, Estrella N (2011) Effects of recent warm and cold spells on European plant phenology. Int J Biometeorol 55 (6):921-932. doi:10.1007/s00484-011-0466-x Metzger MJ, Rounsevell MDA (2011) A need for planned adaptation to climate change in the wine industry PERSPECTIVE. Environ Res Lett 6 (3). doi:10.1088/17489326/6/3/031001 Metzger MJ, Schroter D, Leemans R, Cramer W (2008) A spatially explicit and quantitative vulnerability assessment of ecosystem service change in Europe. Reg Environ Change 8 (3):91-107. doi:10.1007/s10113-008-0044-x Meyer A, Mart ne -Casasnovas JA (1999) Prediction of existing gully erosion in vineyard parcels of the NE Spain: a logistic modelling approach. Soil Tillage Res 50 (3–4):319331. doi:10.1016/S0167-1987(99)00020-3 Monteiro A, Lopes CM (2007) Influence of cover crop on water use and performance of vineyard in Mediterranean Portugal. Agric Ecosyst Environ 121 (4):336-342. doi:10.1016/j.agee.2006.11.016 Morgan DC, Stanley CJ, Warrington IJ (1985) The Effects of Simulated Daylight and ShadeLight on Vegetative and Reproductive Growth in Kiwifruit and Grapevine. J Hortic Sci 60 (4):473-484 Mori K, Sugaya S, Gemma H (2005) Decreased anthocyanin biosynthesis in grape berries grown under elevated night temperature condition. Sci hort 105 (3):319-330. doi:10.1016/j.scienta.2005.01.032 Moriondo M, Bindi M (2007) Impact of climate change on the phenology of typical mediterranean crops. Italian Journal of Agrometeorology 3:5-12 Morlat R, Jacquet A (1993) The Soil Effects on the Grapevine Root-System in Several Vineyards of the Loire Valley (France). Vitis 32 (1):35-42 Morlat R, Jacquet A (2003) Grapevine root system and soil characteristics in a vineyard maintained long-term with or without interrow sward. Am J Enol Vitic 54 (1):1-7 Moss RH, Edmonds JA, Hibbard KA, Manning MR, Rose SK, van Vuuren DP, Carter TR, Emori S, Kainuma M, Kram T, Meehl GA, Mitchell JFB, Nakicenovic N, Riahi K, Smith SJ, Stouffer RJ, Thomson AM, Weyant JP, Wilbanks TJ (2010) The next generation of scenarios for climate change research and assessment. Nature 463 (7282):747-756. doi:10.1038/Nature08823 Moutinho-Pereira J, Goncalves B, Bacelar E, Cunha JB, Coutinho J, Correia CM (2009) Effects of elevated CO2 on grapevine (Vitis vinifera L.): Physiological and yield attributes. Vitis 48 (4):159-165 Moutinho-Pereira J, Magalhães N, Gonçalves B, Bacelar E, Brito M, Correia C (2007) Gas exchange and water relations of three Vitis vinifera L. cultivars growing under Mediterranean climate. Photosynthetica 45 (2):202-207. doi:10.1007/s11099-0070033-1 168 Moutinho-Pereira JM, Correia CM, Goncalves BM, Bacelar EA, Torres-Pereira JM (2004) Leaf gas exchange and water relations of grapevines grown in three different conditions. Photosynthetica 42 (1):81-86 Mullins MG, Bouquet A, Williams LE (1992) Biology of the grapevine. Cambridge University Press, Cambridge, UK Nakićenović N, Alcamo J, Davis G, de Vries HJM, Fenhann J, Gaffin S, Gregory K, Grubler A, Jung TY, Kram T, La Rovere EL, Michaelis L, Mori S, Morita T, Papper W, Pitcher H, Price L, Riahi K, Roehrl A, Rogner H-H, Sankovski A, Schlesinger M, Shukla P, Smith S, Swart R, van Rooijen S, Victor N, Dadi Z (2000) Emissions scenarios. A special report of Working Group III of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA. Nascimbene J, Marini L, Ivan D, Zottini M (2013) Management Intensity and Topography Determined Plant Diversity in Vineyards. Plos One 8 (10):7. doi:10.1371/journal.pone.0076167 Neethling E, Barbeau G, Bonnefoy C, Quenol H (2012) Change in climate and berry composition for grapevine varieties cultivated in the Loire Valley. Clim Res 53 (2):89101. doi:10.3354/Cr01094 Nemani RR, White MA, Cayan DR, Jones GV, Running SW, Coughlan JC, Peterson DL (2001) Asymmetric warming over coastal California and its impact on the premium wine industry. Clim Res 19 (1):25-34 Nendel C (2010) Grapevine bud break prediction for cool winter climates. Int J Biometeorol 54 (3):231-241. doi:DOI 10.1007/s00484-009-0274-8 Neumann PA, Matzarakis A (2011) Viticulture in southwest Germany under climate change conditions. Clim Res 47 (3):161-169. doi:10.3354/cr01000 OIV (2010) RESOLUTION OIV/VITI 333/2010, DEFINITION OF VITIVINICULTURAL “TERROIR”, Tbilisi, 25th June 2010. OIV (2012) Statistical Report on World Vitiviniculture. OIV, Paris, France, 32 pp. Ojeda H, Deloire A, Carbonneau A (2001) Influence of water deficits on grape berry growth. Vitis 40 (3):141-145 Olesen JE, Bindi M (2002) Consequences of climate change for European agricultural productivity, land use and policy. Eur J Agron 16 (4):239-262. doi:10.1016/s11610301(02)00004-7 Olesen JE, Trnka M, Kersebaum KC, Skjelvåg AO, Seguin B, Peltonen-Sainio P, Rossi F, Kozyra J, Micale F (2011) Impacts and adaptation of European crop production systems to climate change. Eur J Agron 34 (2):96-112. doi:10.1016/j.eja.2010.11.003 Oliveira M (1998) Calculation of budbreak and flowering base temperatures for Vitis vinifera cv. Touriga Francesa in the Douro Region of Portugal. Am J Enol Vitic 49 (1):74-78 Orduna RM (2010) Climate change associated effects on grape and wine quality and production. Food Res Int 43 (7):1844-1855. doi:10.1016/j.foodres.2010.05.001 Orlandini S, Di Stefano V, Lucchesini P, Puglisi A, Bartolini G (2009) Current trends of agroclimatic indices applied to grapevine in Tuscany (Central Italy). Idojaras 113 (12):69-78 Orlandini S, Grifoni D, Mancini M, Barcaioli G, Crisci A (2005) ANALYSIS OF METEOCLIMATIC VARIABILITY EFFECTS ON QUALITY OF BRUNELLO DI MONTALCINO WINE. Rivista Italiana di Agrometeorologia (2):37-44 Osorio ML, Osorio J, Pereira JS, Chaves MM (1995) Responses of photosynthesis to water stress under field conditions in grapevines are dependent on irradiance and temperature. Photosynthesis: From Light to Biosphere, Vol Iv:669-672 169 Overman AR, Scholtz RV, Martin FG (2003) In defense of the extended logistic model of crop production. Commun Soil Sci Plant Anal 34 (5-6):851-864. doi:10.1081/css120018979 Ozden M, Vardin H, Simsek M, Karaaslan M (2010) Effects of rootstocks and irrigation levels on grape quality of Vitis vinifera L. cv. Shiraz. Afr J Biotechnol 9 (25):38013807 Pal JS, Giorgi F, Bi XQ, Elguindi N, Solmon F, Gao XJ, Rauscher SA, Francisco R, Zakey A, Winter J, Ashfaq M, Syed FS, Bell JL, Diffenbaugh NS, Karmacharya J, Konare A, Martinez D, da Rocha RP, Sloan LC, Steiner AL (2007) Regional climate modeling for the developing world - The ICTP RegCM3 and RegCNET. Bull Am met Soc 88 (9):1395-1409. doi:10.1175/Bams-88-9-1395 Paranychianakis NV, Aggelides S, Angelakis AN (2004) Influence of rootstock, irrigation level and recycled water on growth and yield of Soultanina grapevines. Agric Water Manag 69 (1):13-27. doi:10.1016/j.agwat.2004.03.012 Parker AK, de Cortazar-Atauri IG, van Leeuwen C, Chuine I (2011) General phenological model to characterise the timing of flowering and veraison of Vitis vinifera L. Aust J Grape Wine R 17 (2):206-216. doi:DOI 10.1111/j.1755-0238.2011.00140.x Pavlousek P (2011) Evaluation of drought tolerance of new grapevine rootstock hybrids. J Environ Biol 32 (5):543-549 Peel MC, Finlayson BL, McMahon TA (2007) Updated world map of the Koppen-Geiger climate classification. Hydrol Earth Syst Sci 11 (5):1633-1644 Peixoto JP, Oort AH (1992) Physics of climate. American Institute of Physics, New York Pelaez H, Rubio JA, Robredo LM, Baeza P, Yuste J (2000) Canopy radiation balance and dry matter production in different training systems. Relationship with geometry and foliar development in irrigated and non-irrigated vines. Proceedings of the Fifth International Symposium on Grapevine Physiology (526):381-389 Pellegrino A, Lebon E, Voltz M, Wery J (2004) Relationships between plant and soil water status in vine (Vitis vinifera L.). Pl Soil 266 (1-2):129-142 Pennec A, Gond V, Sabatier D (2011) Tropical forest phenology in French Guiana from MODIS time series. Remote Sens Lett 2 (4):337-345. doi:10.1080/01431161.2010.507610 Petrie PR, Sadras VO (2008) Advancement of grapevine maturity in Australia between 1993 and 2006: putative causes, magnitude of trends and viticultural consequences. Aust J Grape Wine R 14 (1):33-45. doi:10.1111/j.1755-0238.2008.00005.x Pieri P, Gaudillere JP (2003) Sensitivity to training system parameters and soil surface albedo of solar radiation intercepted by vine rows. Vitis 42 (2):77-82 Plaut G, Simonnet E (2001) Large-scale circulation classification, weather regimes, and local climate over France, the Alps and Western Europe. Clim Res 17 (3):303-324. doi:10.3354/Cr017303 Poni S, Bernizzoni F, Civardi S (2007) Response of "Sangiovese"' grapevines to partial rootzone drying: Gas-exchange, growth and grape composition. Sci hort 114 (2):96-103. doi:10.1016/j.scienta.2007.06.003 Porter JR, Semenov MA (2005) Crop responses to climatic variation. Philos Trans R Soc Lond B Biol Sci 360 (1463):2021-2035. doi:10.1098/rstb.2005.1752 Rabbinge R, Vanlatesteijn HC, Goudriaan J (1993) Assessing the Greenhouse-Effect in Agriculture. Ciba F Symp 175:62-79 Ramirez-Villegas J, Jarvis A (2010) Downscaling global circulation model outputs: The Delta method; Decision and Policy Analysis Working Paper No 1: CIAT-CGIAR. 170 Ramos MC, Jones GV, Martinez-Casasnovas JA (2008) Structure and trends in climate parameters affecting winegrape production in northeast Spain. Clim Res 38 (1):1-15. doi:10.3354/Cr00759 Reidsma P, Ewert F, Lansink AO, Leemans R (2010) Adaptation to climate change and climate variability in European agriculture: The importance of farm level responses. Eur J Agron 32 (1):91-102. doi:10.1016/j.eja.2009.06.003 Reynolds AG, Naylor AP (1994) Pinot-Noir and Riesling Grapevines Respond to WaterStress Duration and Soil Water-Holding Capacity. HortScience 29 (12):1505-1510 Riou C, Carbonneau A, Becker N, Caló A, Costacurta A, Castro R, Pinto PA, Carneiro LC, Lopes C, Clímaco P, Panagiotou MM, Sotez V, Beaumond HC, Burril A, Maes J, Vossen P (1994) Le determinisme climatique de la maturation du raisin: Application au zonage de la teneur en sucre dans la Communauté Européenne. Office des Publications Officielles des Communautés Européennes, Luxembourg, Romero P, Fernandez-Fernandez JI, Martinez-Cutillas A (2010) Physiological Thresholds for Efficient Regulated Deficit-Irrigation Management in Winegrapes Grown under Semiarid Conditions. Am J Enol Vitic 61 (3):300-312 Romero P, Martinez-Cutillas A (2012) The effects of partial root-zone irrigation and regulated deficit irrigation on the vegetative and reproductive development of fieldgrown Monastrell grapevines. Irrig Sci 30 (5):377-396. doi:10.1007/s00271-012-0347z Romero P, Navarro JM, Perez-Perez J, Garcia-Sanchez F, Gomez-Gomez A, Porras I, Martinez V, Botia P (2006) Deficit irrigation and rootstock: their effects on water relations, vegetative development, yield, fruit quality and mineral nutrition of Clemenules mandarin. Tree Physiol 26 (12):1537-1548 Ruml M, Vukovic A, Vujadinovic M, Djurdjevic V, Rankovic-Vasic Z, Atanackovic Z, Sivcev B, Markovic N, Matijasevic S, Petrovic N (2012) On the use of regional climate models: Implications of climate change for viticulture in Serbia. Agric For Meteorol 158:53-62. doi:10.1016/j.agrformet.2012.02.004 Sadras VO, Petrie PR (2011) Climate shifts in south-eastern Australia: early maturity of Chardonnay, Shiraz and Cabernet Sauvignon is associated with early onset rather than faster ripening. Aust J Grape Wine R 17 (2):199-205. doi:10.1111/j.17550238.2011.00138.x Samuelsson P, Jones CG, Willen U, Ullerstig A, Gollvik S, Hansson U, Jansson C, Kjellstrom E, Nikulin G, Wyser K (2011) The Rossby Centre Regional Climate model RCA3: model description and performance. Tellus A 63 (1):4-23. doi:10.1111/j.16000870.2010.00478.x Santos J, Corte-Real J (2006) Temperature extremes in Europe and wintertime large-scale atmospheric circulation: HadCM3 future scenarios. Clim Res 31 (1):3-18 Santos J, Leite S (2009) Long-term variability of the temperature time series recorded in Lisbon. J appl Statist 36 (3):323-337. doi:10.1080/02664760802449159 Santos JA, Corte-Real J, Leite SM (2005a) Weather regimes and their connection to the winter rainfall in Portugal. Int J Clim 25 (1):33-50. doi:10.1002/joc.1101 Santos JA, Corte-Real J, Ulbrich U, Palutikof J (2007) European winter precipitation extremes and large-scale circulation: a coupled model and its scenarios. Theor Appl Clim 87 (1-4):85-102. doi:10.1007/s00704-005-0224-2 Santos JA, Grätsch SD, Karremann MK, Jones GV, Pinto JG (2013) Ensemble projections for wine production in the Douro Valley of Portugal. Clim Change 117 (1-2):211-225. doi:10.1007/s10584-012-0538-x 171 Santos JA, Malheiro AC, Karremann MK, Pinto JG (2011) Statistical modelling of grapevine yield in the Port Wine region under present and future climate conditions. Int J Biometeorol 55 (2):119-131. doi:10.1007/s00484-010-0318-0 Santos JA, Malheiro AC, Pinto JG, Jones GV (2012) Macroclimate and viticultural zoning in Europe: observed trends and atmospheric forcing. Clim Res 51 (1):89-103. doi:10.3354/Cr01056 Santos TP, Lopes CM, Rodrigues ML, Souza CR, Maroco JP, Pereira JS, Silva JR, Chaves MM (2003) Partial rootzone drying: effects on growth and fruit quality of field-grown grapevines (Vitis vinifera). Funct Plant Biol 30 (6):663. doi:10.1071/fp02180 Santos TP, Lopes CMA, Rodrigues ML, de Souza CR, Silva JR, Maroco JP, Pereira JS, Chaves MM (2005b) Effects of partial root-zone drying irrigation on cluster microclimate and fruit composition of field-grown Castelão grapevines. 44 (3):117125 Schaffer B, Andersen PC (1994) Handbook of Environmental Physiology of Fruit Crops. Volume 1. Temperature Crops. CRC Press, Boca Raton. 358 pp. CRC Press, Schultz H (2000) Climate change and viticulture: A European perspective on climatology, carbon dioxide and UV-B effects. Aust J Grape Wine R 6 (1):2-12. doi:10.1111/j.1755-0238.2000.tb00156.x Shahak Y, Ratner K, Giller YE, Zur N, Or E, Gussakovsky EE, Stern R, Sarig P, Raban E, Harcavi E, Doron I, Greenblat-Avron Y (2008) Improving Solar Energy Utilization, Productivity and Fruit Quality in Orchards and Vineyards by Photoselective Netting. Proceedings of the International Symposium on Enhancing Economic and Environmental Sustainability of Fruit Production in a Global Economy (772):65-72 Skamarock WC, Klemp JB, Dudhia J, Gill DO, Barker M, Duda KG, Huang Y, Wang W, Powers JG (2008) A description of the Advanced Research WRF Version 3. doi:citeulike-article-id:8996748 Smart RE, Robinson JB, Due GR, Brien CJ (1985) Canopy Microclimate Modification for the Cultivar Shiraz .2. Effects on Must and Wine Composition. Vitis 24 (2):119-128 Soares PM, Cardoso R, Miranda PA, Medeiros J, Belo-Pereira M, Espirito-Santo F (2012) WRF high resolution dynamical downscaling of ERA-Interim for Portugal. Clim Dynam 39 (9-10):2497-2522. doi:10.1007/s00382-012-1315-2 Solomon A, Goddard L, Kumar A, Carton J, Deser C, Fukumori I, Greene AM, Hegerl G, Kirtman B, Kushnir Y, Newman M, Smith D, Vimont D, Delworth T, Meehl GA, Stockdale T, W UCDP (2011) Distinguishing the Roles of Natural and Anthropogenically Forced Decadal Climate Variability Implications for Prediction. Bull Am met Soc 92 (2):141-156. doi:10.1175/2010bams2962.1 Sousa TA, Oliveira MT, Pereira JM (2006) Physiological Indicators of Plant Water Status of Irrigated and Non-irrigated Grapevines Grown in a Low Rainfall Area of Portugal. Pl Soil 282 (1-2):127-134. doi:10.1007/s11104-005-5374-6 Sparks D, Larsen RP (1966) Effect of Shading and Leaf Area on Fruit Soluble Solids of Concord Grape Vitis Labrusca L. P Am Soc Hortic Sci 89 (Dec):259-294 Spellman G (1999) Wine, weather and climate. Weather 54:230-239 Steppeler J, Doms G, Schattler U, Bitzer HW, Gassmann A, Damrath U, Gregoric G (2003) Meso-gamma scale forecasts using the nonhydrostatic model LM. Meteorol Atmos Phys 82 (1-4):75-96. doi:10.1007/s00703-001-0592-9 Stock M, Gerstengarbe FW, Kartschall T, Werner PC (2005) Reliability of climate change impact assessments for viticulture. Acta Hortic 689:29-39 Stoica P, Moses RL (1997) Introduction to spectral analysis. Prentice Hall, Upper Saddle River, N.J. 172 Storchi P, Costantini EAC, Bucelli P (2005) The influence of climate and soil on viticultural and enological parameters of 'Sangiovese' grapevines under non-irrigated conditions. Acta Hortic 689:333-340 Taylor JA (2004) Digital Terroirs and Precision Viticulture: Investigations Into the Application of Information Technology in Australian Vineyards. Australian Centre for Precision, Agriculture Faculty of Agriculture, Food and Natural Resources, University of Sydney, Sydney, Australia Tello J, Cordero-Bueso G, Aporta I, Cabellos JM, Arroyo T (2012) Genetic diversity in commercial wineries: effects of the farming system and vinification management on wine yeasts. J Appl Microbiol 112 (2):302-315. doi:10.1111/j.13652672.2011.05202.x This P, Lacombe T, Thomas MR (2006) Historical origins and genetic diversity of wine grapes. Trends Genet 22 (9):511-519. doi:10.1016/j.tig.2006.07.008 Tomasi D, Jones GV, Giust M, Lovat L, Gaiotti F (2011) Grapevine Phenology and Climate Change: Relationships and Trends in the Veneto Region of Italy for 1964-2009. Am J Enol Vitic 62 (3):329-339. doi:10.5344/ajev.2011.10108 Tonietto J (1999) Les macroclimats viticoles mondiaux et l’influence du mésoclimat sur la typicité de la Syrah et du Muscat de Hambourg dans le sud de la France: méthodologie de caractérisation. Tonietto J, Carbonneau A (2004) A multicriteria climatic classification system for grapegrowing regions worldwide. Agric For Meteorol 124 (1-2):81-97. doi:10.1016/j.agrformet.2003.06.001 Tramontini S, van Leeuwen C, Domec JC, Destrac-Irvine A, Basteau C, Vitali M, MosbachSchulz O, Lovisolo C (2013) Impact of soil texture and water availability on the hydraulic control of plant and grape-berry development. Pl Soil 368 (1-2):215-230. doi:10.1007/s11104-012-1507-x Trenberth KE, Josey SA (2007) Observations: surface and atmospheric climate change. In Climate Change 2007: The Physical Science Basis: Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. In: Solomon S, Qin D, Manning M et al. (eds). Cambridge University Press, Cambridge, United kingdom., pp 235-336 Trigo RM, DaCamara CC (2000) Circulation weather types and their influence on the precipitation regime in Portugal. Int J Clim 20 (13):1559-1581. doi:10.1002/10970088(20001115)20:13<1559::aid-joc555>3.0.co;2-5 Tubiello FN, Fischer GI (2007) Reducing climate change impacts on agriculture: Global and regional effects of mitigation, 2000-2080. Technol Forecast Soc Change 74 (7):10301056. doi:10.1016/j.techfore.2006.05.027 Ugalde D, Brungs A, Kaebernick M, McGregor A, Slattery B (2007) Implications of climate change for tillage practice in Australia. Soil Tillage Res 97 (2):318-330. doi:10.1016/j.still.2007.09.018 Unwin PTH (1996) Wine & The Vine. Routledge, London, United kingdon Urhausen S, Brienen S, Kapala A, Simmer C (2011) Climatic conditions and their impact on viticulture in the Upper Moselle region. Clim Change 109 (3-4):349-373. doi:DOI 10.1007/s10584-011-0059-z USDA (2006) United States Department of Agriculture, Soil Mechanics Level 1. Module 3USDA Textural Soil Classification Study Guide. Usha K, Singh B (2013) Potential applications of remote sensing in horticulture—A review. Sci hort 153 (0):71-83. doi:10.1016/j.scienta.2013.01.008 173 van der Linden P, Mitchell JFB (2009) ENSEMBLES: Climate Change and its Impacts: Summary of research and results from the ENSEMBLES project. Met Office Hadley Centre, FitzRoy Road, Exeter EX1 3PB, UK. 166 pp. van Leeuwen C, Friant P, Choné X, Tregoat O, Koundouras S, Dubordieu D (2004) Influence of climate, soil, and cultivar on terroir. Am J Enol Vitic 55 (3):207-217 van Leeuwen C, Tregoat O, Chone X, Bois B, Pernet D, Gaudillere JP (2009) Vine Water Status Is a Key Factor in Grape Ripening and Vintage Quality for Red Bordeaux Wine. How Can It Be Assessed for Vineyard Management Purposes? J Int Sci Vigne Vin 43 (3):121-134 Veloso MM, Almandanim MC, Baleiras-Couto M, Pereira HS, Carneiro LC, Fevereiro P, Eiras-Dias J (2010) Microsatellite database of grapevine (Vitis vinifera L.) cultivars used for wine production in Portugal. Ciência Téc Vitiv 25 (2):53-61 von Storch H, Flöser G (1999) Anthropogenic climate change. GKSS School of Environmental Research,. Springer, Berlin ; New York Vrsic S, Vodovnik T (2012) Reactions of grape varieties to climate changes in North East Slovenia. Plant Soil Environ 58 (1):34-41 Webb LB, Whetton PH, Barlow EWR (2007) Modelled impact of future climate change on the phenology of winegrapes in Australia. Aust J Grape Wine R 13 (3):165-175. doi:10.1111/j.1755-0238.2007.tb00247.x Webb LB, Whetton PH, Barlow EWR (2008a) Climate change and winegrape quality in Australia. Clim Res 36:99-111. doi:10.3354/cr00740 Webb LB, Whetton PH, Barlow EWR (2008b) Modelling the relationship between climate, winegrape price and winegrape quality in Australia. Clim Res 36 (2):89-98. doi:10.3354/cr00739 Webb LB, Whetton PH, Barlow EWR (2011) Observed trends in winegrape maturity in Australia. Global Change Biol 17 (8):2707-2719. doi:10.1111/j.13652486.2011.02434.x Webb LB, Whetton PH, Bhend J, Darbyshire R, Briggs PR, Barlow EWR (2012) Earlier wine-grape ripening driven by climatic warming and drying and management practices. Nat Clim Change 2 (4):259-264. doi:10.1038/Nclimate1417 Weigel AP, Knutti R, Liniger MA, Appenzeller C (2010) Risks of Model Weighting in Multimodel Climate Projections. J Clim 23 (15):4175-4191. doi:10.1175/2010jcli3594.1 White MA, Diffenbaugh NS, Jones GV, Pal JS, Giorgi F (2006) Extreme heat reduces and shifts United States premium wine production in the 21st century. Proc Natl Acad Sci U S A 103 (30):11217-11222. doi:10.1073/pnas.0603230103 Wilks DS (2006) Statistical methods in the atmospheric sciences. International geophysics series, vol 91, 2nd edn. Academic Press, Amsterdam ; Boston Wilks DS (2011) Statistical methods in the atmospheric sciences. International geophysics series, vol 91, 2nd edn. Academic Press, Amsterdam ; Boston Winkel T, Rambal S, Bariac T (1995) SPATIAL VARIATION AND TEMPORAL PERSISTENCE OF GRAPEVINE RESPONSE TO A SOIL TEXTURE GRADIENT. Geoderma 68 (1-2):67-78. doi:10.1016/0016-7061(95)00026-k Winkler AJ (1974) General viticulture. University of California Press, California, USA Wramneby A, Smith B, Samuelsson P (2010) Hot spots of vegetation-climate feedbacks under future greenhouse forcing in Europe. J Geophys Res-Atmos 115:D21119. doi:10.1029/2010jd014307 Xi ZM, Zhang ZW, Cheng YF, Li H (2010) The Effect of Vineyard Cover Crop on Main Monomeric Phenols of Grape Berry and Wine in Vitis viniferal L. cv. Cabernet 174 Sauvignon. Agricultural Sciences in China 9 (3):440-448. doi:10.1016/S16712927(09)60115-2 Yau IH, Davenport JR, Rupp RA (2013) Characterizing Inland Pacific Northwest American Viticultural Areas with Geospatial Data. Plos One 8 (4). doi:10.1371/journal.pone.0061994 Zsofi Z, Toth E, Rusjan D, Balo B (2011) Terroir aspects of grape quality in a cool climate wine region: Relationship between water deficit, vegetative growth and berry sugar concentration. Sci hort 127 (4):494-499. doi:10.1016/j.scienta.2010.11.014 175 176 Supplementary material 1 177 Table S1.1. Spatial correlations between the Composite Indices for each pair of the 16 simulations and the ensemble mean. ARP Aladin (CNRM) ARP Aladin (CNRM) ARP HIRHAM (DMI) BCM RCA (SMHI) EH5 HIRHAM (DMI) EH5 RACMO (KNMI) ARP HIRHA M (DMI) BCM RCA (SMHI) EH5 HIRHA M (DMI) EH5 RACMO (KNMI) EH5 RCA (SMHI) EH5 RegCM (ICTP) EH5 REMO (MPI) HC CLM (ETHZ) HC HadRM3 Q0 (HC) HC HadRM3 Q3 (HC) HC HadRM3 Q16 (HC) HC RCA3 (C4I) EH5 CLM (MPI) 0.71 0.61 0.60 0.73 0.70 0.66 0.73 0.74 0.66 0.80 EH5 RCA (SMHI) EH5 RegCM (ICTP) 0.75 0.67 0.68 0.82 0.80 0.69 0.61 0.69 0.79 0.78 0.84 EH5 REMO (MPI) 0.71 0.79 0.61 0.76 0.83 0.74 0.70 HC CLM (ETHZ) 0.65 0.84 0.58 0.64 0.77 0.62 0.59 0.84 0.63 0.83 0.57 0.63 0.74 0.59 0.57 0.81 0.93 0.52 0.73 0.54 0.54 0.67 0.46 0.44 0.76 0.83 0.81 0.67 0.81 0.61 0.68 0.74 0.65 0.60 0.77 0.79 0.81 0.68 HC RCA (SMHI) 0.58 0.76 0.58 0.58 0.72 0.54 0.53 0.77 0.87 0.87 0.86 0.73 HC HadRM3Q0 (HC) HC HadRM3Q3 (HC) HC HadRM3Q16 (HC) HC RCA (SMHI) HC RCA3 (C4I) 0.76 0.78 0.63 0.77 0.77 0.80 0.74 0.78 0.73 0.76 0.55 0.83 0.65 EH5 CLM (MPI) 0.69 0.81 0.58 0.69 0.78 0.66 0.62 0.84 0.85 0.81 0.76 0.78 0.77 0.74 MEAN 0.80 0.90 0.73 0.82 0.89 0.81 0.77 0.91 0.91 0.91 0.81 0.89 0.86 0.88 Key <0.5 0.5>0.7 0.7>0.9 >0.9 178 0.89 Fig. S1.1. Scatter diagrams between the mean and absolute mean bias for the 16 ENSEMBLES regional climate models, relative to the baseline period (1961-2000), in the (a) Huglin Index, (b) Dryness Index and (c) Hydrothermal Index. 179 180 Supplementary material 2 181 Table S2.1. List of the bioclimatic indices computed for this study, their corresponding mathematical definitions, units and classes. Bioclimatic Index Mathematical definition Units Classes Cool Night Index (CI) September average Tmin (°C) °C Very cool nights: < 12 Cool nights: 12 – 14 Temperate nights: 14 – 18 Warm nights: > 18 mm Excessively dry: < 100 Moderately dry: -100 – 50 Sub-Humid: 50 – 150 Humid: > 150 °C Unsuitably Cool: < 900 Too Cool: 900 – 1200 Very Cool: 1200 – 1500 Cool: 1500 – 1800 Temperate: 1800 – 2100 Warm/Temperate: 2100 – 2400 Warm: 2400 – 2700 Very Warm: 2700 – 3000 Too Hot: > 3000 °C h Unsuitable below 2.6 °C Too Cool: <1111 Cool: 1111 – 1389 Temperate: 1389 – 1667 Temperate/Warm: 1667 – 1944 Warm: 1944 – 2222 Very warm: 2222 – 2500 Too Hot: > 2500 – 2778 – – °C Too cool: < 12 Cool: 12 – 15 Intermediate: 15 – 17 Warm: 17 – 19 Hot: 19 – 21 Very Hot: 21 – 22 Too Hot= > 22 °C mm Low risk of contamination: < 2500 Medium risk of contamination: 2500 – 5100 High risk of contamination: 5100 – 7500 Very high risk of contamination: > 7500 mm °C -1 Insufficient hydric regime: < 1 Normal hydric regime: 1 – 3 Excessive hydric regime: > 3 Sept. (Wo P Tv Es ) April Dryness Index (DI) Wo - Initial available soil water reserve (mm) on the first month / DI on the following months; P – Precipitation (mm); Tv - Potential vineyard transpiration (mm); Es - Direct evaporation from the soil (mm) Tv; Es are assessed using the Thornthwaite method (T 10) (Tmax 10) d 2 April Sept . Huglin Index (HI) T - Mean air temperature (°C); Tmax - Maximum air temperature (°C); d - Length of day coefficient, from 1.02 to 1.06 Oct . Branas Heliothermic Index ((T 10) H ) 10 6 April T - Mean air temperature (°C); H – Astronomical insolation (h) Oct . Growing Degree Day (T 10) April T - Mean air temperature (°C); Growing season accumulated precipitation Sept. (P) April P – Precipitation (mm) Oct Growing season mean temperature (T ) April T - Mean air temperature (°C) Aug . (T P) Hydrothermic Index April T - Mean air temperature (°C); P – Precipitation (mm) Sept. Selianinov Index P ( T 10 ) April P – Precipitation (mm); T - Mean air temperature (°C) 182 Table S2.2. Spatial correlation matrix (Pearson correlation coefficients) of nine bioclimatic indices over the Portuguese mainland in values equal to or above 0.95 are highlighted, all correlations are statistically significant at a significance level of 5% (p-level of 0.05). Growing Branas Selianinov Dryness Season Heliothermic Hydrothermal Index Huglin Index Index Index Precipitation Index 1950-2000. Correlations with absolute Cool Night Index Growing Season Temperature Growing Season Precipitation Branas Heliothermic Index -0.84 Selianinov Index 0.89 -0.83 Hydrothermal Index 0.96 -0.72 0.75 Huglin Index -0.81 0.91 -0.88 -0.66 Dryness Index 0.95 -0.83 0.89 0.86 -0.91 Cool Night Index -0.77 0.95 -0.82 -0.63 0.91 -0.81 Growing Season Temperature -0.80 0.89 -0.89 -0.64 0.98 -0.92 0.91 Growing Degree Day -0.81 0.92 -0.88 -0.65 0.98 -0.91 0.93 183 1.00 Growing Degree Day Table S2.3. GCM-RCM chains used in this study, along with their development institutions. GCM ARPEGE RCM Aladin HIRHAM HIRHAM RCA3 HadRM3Q0 Hadley Centre HadRM3Q16 Hadley Centre HadRM3Q3 Hadley Centre RCA RegCM3 REMO CLM RCA HadCM3 Centre National de Recherches Météorologiques Danish Meteorological Institute Danish Meteorological Institute Koninklijk Nederlands Meteorologisch Instituut Swedish Meteorological and Hydrological Institute International Centre for Theoretical Physics Max Planck Institute Eidgenössische Technische Hochschule Zürich Swedish Meteorological and Hydrological Institute C4I Center RACMO2 ECHAM5 Institution 184 Table S2.4. Huglin Index (HI), Dryness Index (DI) and Cool Night Index (DI) minimum, maximum and mean for the 12 Portuguese mainland wine regions shown in Fig. 1b. Highlighted values represent the extremes of each index for the whole country (minimum of the minima, maximum of the maxima and maximum/minimum of the areameans). HI (in C° units) DI (in mm) CI (in C° units) Minimum Maximum Area-Mean Minimum Maximum Area-Mean Minimum Maximum Area-Mean Wine Regions Alentejo Algarve Beira-Atlântico Douro/Porto Lisboa Minho Península-deSetúbal Tejo Terras-da-Beira Terras-de-Cister Terras-do-Dão Trás-os-Montes 19502000 2076 1985 1725 1434 1485 1476 20412070 2729 2531 2297 2078 2041 2086 19502000 2491 2447 2295 2248 2191 1989 20412070 3209 2931 2945 3001 2833 2628 19502000 2282 2285 1944 1853 1822 1749 20412070 2942 2772 2561 2518 2368 2407 19502000 -21.7 -23.3 51.3 13.9 26.4 65.5 20412070 -127 -98 -25 -60 -43 -20 19502000 25.6 8.5 93.0 101.5 67.0 124.6 20412070 -64 -48 26 20 14 39 19502000 -0.2 -8.0 74.4 56.7 49.5 97.0 20412070 -95 -79 4.0 -19 -10 4.0 19502000 15.4 15.6 14.1 11.3 14.5 12.7 20412070 18.6 18.6 17.0 14.7 17.2 15.9 19502000 16.8 17.7 15.3 14.7 16.7 15.1 20412070 20.0 20.5 18.1 18.1 19.5 18.0 19502000 16.1 16.8 14.6 13.2 15.8 13.7 20412070 19.3 19.6 17.5 16.5 18.4 16.8 2057 2515 2465 3102 2319 1881 1461 1445 1523 1056 2457 2138 2075 2261 1825 2428 2418 1995 1995 2076 3115 3148 2626 2768 2762 2180 1956 1571 1810 1623 2910 -5.4 -92 31.8 -21 10.3 -64 16.1 18.5 17.0 19.8 16.9 19.7 2802 2667 2210 2566 2316 10.1 -16.0 62.8 53.2 35.9 -76 -119 -17 -32 -38 48.0 85.2 99.4 87.0 105.7 -17 13 21 -2.0 30 26.9 31.8 82.6 68.2 63.2 -49 -51 6.0 -15 -11 15.3 11.8 11.8 12.7 9.8 18.3 15.1 15.0 16.4 13.1 17.0 16.7 13.6 14.3 13.7 19.9 20.1 16.8 17.9 17.1 16.4 14.5 12.5 13.8 12.0 19.4 17.9 15.7 17.1 15.3 185 Table S2.5. Comparison between the Multi-criteria Climatic Classification (MCC) and the Categorized Index (CatI) for the 12 Portuguese Wine Regions in 1950-2000. All calculation are done using area averages of the composing indices (HI, DI and CI) in each region. Region MCC CatI Alentejo HI+1,DI+1,CI-1 10 Algarve HI+1,DI+1,CI-1 10 Beira-Atlântico HI-1,DI-1,CI-1 8 Douro/Porto HI-1,DI-1,CI+1 7 Lisboa HI-1,DI+1,CI-1 6 Minho HI-2,DI-1,CI+1 7 Península-de-Setúbal HI+1,DI+1,CI-1 10 Tejo HI+1,DI+1,CI-1 10 Terras-da-Beira HI-1,DI+1,CI-1 6 Terras-de-Cister HI-2,DI-1,CI+1 7 Terras-do-Dão HI-1,DI-1,CI+1 7 Trás-os-Montes HI-2,DI-1,CI+2 7 186 Fig. S2.1. Differences between the mean patterns of the Cool Night Index (CI; in ºC), Dryness Index (DI; in mm) and Huglin Index (HI; in °C) calculated using the E-OBS and WorldClim datasets for the same period (1950-2000). The WorldClim dataset was bilinearly interpolated to the same grid as the E-OBS observational dataset in these computations. 187 Fig. S2.2. Mean patterns of the Cool Night Index (CI; in ºC) for (a) 1950-2000 (WorldClim baseline) and (b) 2041-2070 (13-member ensemble mean for the A1B scenario) over Portugal 188 Fig. S2.3. Mean patterns of the Dryness Index (DI; in mm) for (a) 1950-2000 (WorldClim baseline) and (b) 2041-2070 (13-member ensemble mean for the A1B scenario) over Portugal 189 Fig. S2.4. Mean patterns of the Huglin Index (HI; in C° units) for (a) 1950-2000 (WorldClim baseline) and (b) 2041-2070 (13-member ensemble mean for the A1B scenario) over Portugal 190 Fig. S2.5. Percentages of category changes in CatI between 1950-2000 and 2041-2070 (A1B scenario) over Portugal. Black (white) bar for changes to (from) category 0. Only category transitions above 1% are represented 191 Supplementary material 3 192 Table S3.1 - The Cool Nigh, Dryness and Huglin indices, along with their mathematical definition, units and classes. Index Mathematical definition Units Classes Cool Night Index (CI) September average Tmin (°C) (North hemisphere) °C Very cool nights: < 12 Cool nights: 12 – 14 Temperate nights: 14 – 18 Warm nights: > 18 mm Excessively dry: <-100 Moderately dry: -100 – 50 Sub-Humid: 50 – 150 Humid: > 150 °C Unsuitably Cool: < 900 Too Cool: 900 – 1200 Very Cool: 1200 – 1500 Cool: 1500 – 1800 Temperate: 1800 – 2100 Warm/Temperate: 2100 – 2400 Warm: 2400 – 2700 Very Warm: 2700 – 3000 Too Hot: > 3000 Sept. (Wo P Tv Es) April Dryness Index (DI) Wo - Initial available soil water reserve (mm) on the first month / DI on the following months; P – Precipitation (mm); Tv - Potential vineyard transpiration (mm); Es - Direct evaporation from the soil (mm) Tv; Es are assessed using the Thornthwaite method (T 10) (Tmax 10) d 2 April Sept . Huglin Index (HI) T - Mean air temperature (°C); Tmax - Maximum air temperature (°C); d - Length of day coefficient, from 1.02 to 1.06 193 Table S3.2 - Spearman ranked correlation coefficient between the EVI, CatI, SoilT, elevation, aspect and solar radiation in all of Iberia. EVI CatI SoilT Elevation Aspect Radiation EVI CatI SoilT -0.35 0.12 0.02 0.01 -0.29 0.04 -0.66 0.01 0.61 -0.10 0.03 0.13 Elevation Aspect -0.01 -0.28 Radiation -0.01 194 Table S3.3 - Area (km2) of each DO region in Iberia. # Region 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 Alella Alenquer Alentejo Alicante Almansa Arlanza Arribes Arruda Bairrada Beira Interior Bierzo Binissalem Bucelas Bullas Calatayud Campo de Borja Carcavelos Cariñena Chacolí de Álava Chacolí de Guetaria Chacolí de Vizcaya Cigales Colares Conca de Barberà Condado de Huelva Costers del Segre Dão Douro Empordà Encostas d'Aire Jerez Area (km2) 140 270 4873 2645 2069 2231 1780 128 1034 5639 1497 156 46 5012 1608 624 38 829 324 88 1517 625 118 427 2638 1791 3859 2555 1070 1819 2409 # 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 Area (km2) 2507 23569 693 987 608 345 5184 6059 196 2555 926 772 2667 593 3755 437 295 1653 1121 1782 177 180 2065 2120 400 3125 16353 415 3715 2846 724 Region Jumilla La Mancha Lafões Lagoa Lagos Lourinhã Málaga & Sierras de Málaga Manchuela Manzanilla Méntrida Mondéjar Monterrei Montilla-Moriles Montsant Navarra Obidos Palmela Penedès Pla de Bages Pla i Llevant Portimão Priorat Rías Baixas Ribeira Sacra Ribeiro Ribera del Duero Ribera del Guadiana Ribera del Júcar Rioja Rueda Setúbal 195 # 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 Region Somontano Tarragona Tavira Távora-Varosa Tejo Terra Alta Tierra de León Tierra del Vino de Zamora Toro Torres Vedras Trás-os-Montes Uclés Utiel-Requena Valdeorras Valdepeñas Valencia Vinho Verde Vinos de Madrid Yecla Area (km2) 2261 1623 454 419 5459 766 3294 1828 719 273 3566 2103 1775 670 1789 3432 7926 3054 630
