Overview Forest ecosystem services and climate change in mountain regions • Forest ecosystem services – a European perspective? • Tools to assess future forest dynamics • Uncertainties • Two case studies with three models and four scenarios • Adaptive forest management Harald Bugmann, Christof Bigler, Sebastian Leuzinger, Annett Wolf, Che Elkin, Alvaro Gutierrez, Christian Temperli, Corina Manusch & Livia Rasche Forest Ecology, ETH Zürich, Switzerland • Conclusions 2 Ecosystem services framework Forest ecosystem services Resources 3 Millennium Ecosystem Assessment (2005) 4 Conservation Water Forest ecosystem services Natural hazards Carbon storage Overview (Tourism) • Forest ecosystem services – a European perspective? • Tools to assess future forest dynamics • Uncertainties • Two case studies with three models and four scenarios • Adaptive forest management • Conclusions 5 6 Forest succession models: approach Forests don’t fit into greenhouses • Quantitative description of tree population dynamics: – Establishment – Growth – Mortality Images by speedtree Spatial scale (m)! • Concept of small-scale mosaic of successional patches (Gleason, Botkin, Shugart): so-called „Gap models“ • Sensitive to climatic factors 7 Bugmann et al. (2000), Clim Change 8 • Concept underlying most current dynamic models of (potentially) uneven-aged stands Bugmann (2001), Clim Change 1. FORCLIM model 2. LPJ-GUESS model • Volume change of a tree: Fagus 40 silvatica T Hm 30 H (m) dV/dt = r · L – m · V 20 CO2 H 2O measurement 10 • Allometric relationships Eq. 1 (parabolic) Eq. 2 (asymptotic) (D = tree diameter at breast height): L = ƒ1(D) V = ƒ2(H,D) H = ƒ3(D) 0 0 10 20 30 D (cm) 40 50 • Carbon & water balance • Vegetation distribution at several spatial & temporal scales: “from leaf to ecosystem” Dm 60 CO2 H 2O H 2O H 2O CO2 CO2 • ...from which follows (after some math): dD = g · D · (1 – H ) · 1 · ƒ(e) dt Hmax b(D) exponential asymptote Soil processes allometry environment Moore (1989), Ecol Modelling 9 3. LandClim model LandClim Overview Climate • Forest ecosystem services – a European perspective? Fire Vegetation (Gap model) Recruitment Growth Mortality • Tools to assess future forest dynamics Bark beetles • Uncertainties Windthrow Cell Management Dispersal • Two case studies with three models and four scenarios Landscape • Adaptive forest management Spatially explicit input • • • • 11 Smith et al. (2001), Glob Ecol Biogeogr 10 Elevation Aspect Slope Soil AWC • • • • • Conclusions Monthly climate Browsing mask Management masks Land type mask Schumacher et al. (2006), Landsc Ecol! 12 Types of uncertainty Climate scenarios • Scenarios of climate change • Understanding of forest ecosystems (model formulations) • Derivation of indicators of ecosystem goods and services • Future management approaches (adaptive?) … and others (with which I do not deal today) 13 14 “Grow fast, die young” • Study of growth rate vs. longevity: – 3 species – 2 continents • Data on maximum growth rate (at young age) and maximum longevity of 141 temperate & boreal species 72 species, eastern North America Maximum life expectancy (yrs) • Generalizing the finding Negative exponential relationships: slope α in range [-0.35…-0.64] Slope A of the relationship 600 y = 499.7e-0.0048x 500 R2 = 0.70 400 300 200 100 0 • 15 Implications? e.g. CO2 fertilization, long-term forest dynamics & biomass? 0 50 100 150 200 250 300 350 400 450 500 Maximum growth rate (cm/yr) • Bigler & Veblen (2009), Oikos 16 Slope α (scaled to Bigler & Veblen units): [-0.31...-0.61] Bugmann & Bigler (2011), Oecologia Forest succession models: mortality Exploring the effect using FORCLIM • Combination of • • – “background“ mortality that is constant across tree life, tied to maximum tree age kAm): small fraction of trees survives to kAm (“age-independent” mortality = AIM) – growth-related mortality (“stress-related” mortality = SM) Net effect of growth stimulation vs. reduced longevity unknown Simulation study at 6 sites along climate gradient Davos Change in total biomass (%) Davos, change in total aboveground biomass Mortality probability (yr-1) ! • Overall effect: A005 A004 A003 G+10 • CO2 fertilization: G+20 slope of growth-age relationship A006 A000 G+30 growth rate – Higher AIM (reduced longevity) – Reduced SM (higher growth rate) 17 Bugmann (2001), Clim Change Bugmann & Bigler (2011), Oecologia 18 Exploring the effect using FORCLIM 21.0%-24.0% 18.0%-21.0% 15.0%-18.0% 12.0%-15.0% 9.0%-12.0% 6.0%-9.0% 3.0%-6.0% 0.0%-3.0% -3.0%-0.0% -6.0%--3.0% -9.0%--6.0% -12.0%--9.0% -15.0%--12.0% -18.0%--15.0% -21.0%--18.0% -24.0%--21.0% Taken together... • Lack of growth-longevity relationship & emphasis on “source limitation” (not shown here) … …explain strong CO2 effect in “mechanistic” Dynamic Global Vegetation Models (DGVMs) • Results averaged over all sites (multi-species case): All sites, multi-species, Δbiomass! most probable value of slope 19 Bugmann & Bigler (2011), Oecologia 20 (Short term) reality – (long-term) artefact? Cramer et al. (2001), GCB Indicators of ecosystem goods & services • Forest diversity Shannon’s tree species diversity Example: Protection from rockfall • Forest protection against gravitational hazards Shannon’s structural diversity • • • • General protection Avalanche Rockfall Landslip Stand Maturity index Habitat quality (Umbrella species) 21 22 Overview Case study areas Saas Valley • Forest ecosystem services – a European perspective? • Tools to assess future forest dynamics • Uncertainties Dischma Valley • Two case studies with three models and four scenarios • Adaptive forest management • Conclusions 23 24 Forest models and management scenarios Forest ecosystem services • Forest models: – FORCLIM: forest stand dynamics • • • • (e.g., Rasche et al. 2011, J Appl Ecol) – LPJ-GUESS: biogeochemistry (e.g., Wolf et al. 2012, Clim Change) – LANDCLIM: landscape dynamics • Net Ecosystem Exchange (NEE) and Net Primary Productivity (NPP) • Ecosystem carbon storage • Runoff (pixel-based, no explicit water routing) (e.g., Temperli et al. 2012, Ecol Appl) • Wildfire risk and impacts • Protection from rockfall and avalanches • Landscape diversity • Management scenarios: – Continuation of current management (past-WWII) – Management parameters calibrated across models to yield comparable results Rasche et al. (2011), J Appl Ecol 25 26 Climate scenarios 27 Forest biomass Forest composition Diversity (species, structure) Harvestable timber Results: Forest biomass 28 LPJ-GUESS Results: Forest composition Results: Soil carbon FORCLIM 29 Results: Rockfall protection ACQWA LPJ-GUESS 30 Results: Harvestable biomass +2° A2 Saas valley, Year 2100 31 Strong decrease Strong increase LANDCLIM 32 FORCLIM Case study northern Black Forest Stem number >16 cm DBH /ha 0 100 200 Overview • Forest ecosystem services – a European perspective? • Tools to assess future forest dynamics • Uncertainties • Two case studies with three models and four scenarios Mixed forest (equilibrium) Even-aged spruce forest (age 100) DBH classes [cm] DBH classes [cm] • Adaptive forest management 2 km • Conclusions Elevation: 250 – 1050 m a.s.l. 33 Local stand data courtesy of FVA German national forest inventory from Yousefpour et al. (2010), Env. Management 34 Adaptive forest management Adaptive forest management Management types 1. No-change management: Traditions Management type Even-aged Norway spruce No-change (Business-As-Usual scenario in mid and upper part) Uneven-aged Managament decisions based on traditions Does not consider climate change or uncertainties in general 2. Reactive management: Current knowledge Considers observed forest state and current knowledge 3. Trend-adaptive management: Expert predictions Takes observed forest state and expert predictions of climate change into account 3 subtypes: 1. Adapts at one decision point 2. Adapts at multiple decision points 3. Considers change in climate change prediction Mixed forest Reactive (Business-As-Usual scenario in lower part) Natural Vegetation – Douglas/silver fir Simple trend-adaptive (Subtype 1) Mixed oak forest Simple trend-adaptive (Subtype 1) Species adaptation No species adaptation 4. Forward-looking adaptive management: Inherent Uncertainty Douglas-fir / silver fir Even-aged spruce Timber production Solves a stochastic dynamic programming problem to identify the optimal decision 35 Management scenario ! Thorsen et al. (2010), MOTIVE Working Paper Mixed oak forest Mixed forest Natural veg. Biodiversity promotion Ministerium Ländlicher Raum Baden-Württemberg, 1999, Richtlinien landesweiter Waldentwicklungstypen, Stuttgart Dunker et al. 2007, Definition of forest management alternatives, EFORWOOD D2.1.3, Albert-Ludwigs-Universität, Freiburg 36 Spiecker et al. 2004, Norway Spruce Conversion: Options and Consequences, Brill Academic Publishers, Köln Climatic change 1 0 1 Ecosystem services Harvested volume and biodiversity (normalized) Forest state Simulation results 0 Simulation results Management intensity Temperli et al. (2012), Ecol Appl 37 ! Overview Utility function: ! • Forest ecosystem services – a European perspective? UTot = wT·UT + wEco·(UH + USMI)/2 wT = 1 – wEco • Tools to assess future forest dynamics • Uncertainties Utility Timber production Trade-offs over time wT Biodiversity promotion EGS 39 ! wT ! Temperli et al. (2012), Ecol Appl 38 • Two case studies with three models and four scenarios • Adaptive forest management ! wT • Conclusions 40 Conclusions • Dynamic models are essential tools for assessing future ecosystem service provision from mountain forests, but large uncertainties remain Forest ecosystem services and climate change in mountain regions • The issue may be less whether processes are formulated “statistically” or “mechanistically”, but rather whether we consider all relevant processes • Ensemble climate scenarios may smooth out important ecosystem responses, and imply a false sense of certainty for impact models • Collapsing ecosystem states into indicators of ecosystem service provisions is necessary step, but a slippery slope • For some ecosystem services in some parts of some landscapes, strongly negative effects under a 2° scenario – but: also positive effects of CC! Thank you for your attention! • Not even a 2° world is generally “safe” for mountain forests and the multiple services they provide => Worrying! High sensitivity of mountain forests to climate change 41 http://www.fe.ethz.ch 42