Pervasive climatemediated changes in western forests N. Stephenson, J. Littell, P. van Mantgem, D. Peterson, and D. McKenzie Part 1: Defining the issues and setting the stage Part 2: FOREST DEMOGRAPHY: • Ongoing changes • Possible future changes Part 3: FOREST GROWTH: • Ongoing changes • Possible future changes Part 4: A call to action DEFINING THE ISSUES AND SETTING THE STAGE Climate Fire Forest Climate Fire Forest Climate Fire Forest Already, climate has been linked to episodes of broad-scale forest die-back in the mountainous West Drought Warming (e.g., southern Calif.) (e.g., British Columbia) Credit: USFS Credit: BC Ministry of Forests and Range But what’s happening – and likely to happen – in the bulk of mountain forests in the West, which have not experienced extensive die-back? Credit: Nate Stephenson Why care? Recent studies of “healthy” forests in the American tropics show that substantial directional changes are in progress, with potentially profound consequences. Credit: Nate Stephenson Tropical forest COMPOSITION is changing (e.g., lianas [woody vines] are increasing) Credit: Yadvinder Malhi Phillips et al., Nature, 2002 Tropical forest DYNAMICS are changing (e.g., recruitment, growth, and mortality rates are increasing) Recruitment Mortality Phillips et al., Phil. Trans. B, 2004 Tropical forest STRUCTURE AND FUNCTION are changing (e.g., aboveground biomass, hence C storage, may be increasing) Basal area gain Basal area loss Difference Lewis et al., Phil. Trans. B, 2004 Are similar changes underway in our Western forests? Are similar changes underway in our Western forests? We simply don’t know. No one has been doing the necessary systematic analyses. Our questions: In the bulk of mountain forests in the West (which have not experienced extensive die-back): • Are climatically-driven changes in progress? • What might we expect for the future? Our questions: In the bulk of mountain forests in the West (which have not experienced extensive die-back): • Are climatically-driven changes in progress? • What might we expect for the future? Our approach: Here, we will address each question separately for: • Forest demography • Forest growth DEMOGRAPHY determines NUMBERS of trees (birth, natality, recruitment & death, mortality) GROWTH determines SIZES of trees TOGETHER, demography and growth rates give us a forest (structure, composition, productivity, and dynamics) DEMOGRAPHY: ongoing changes • Data from dozens of permanent forest plots show that over the last few decades, in the otherwise undisturbed old-growth forests of California, Oregon, and Washington, tree mortality rates have been increasing. • However, unlike the tropics, recruitment rates have NOT been increasing. Possible cause: Maximum snow water content Summers are getting longer and drier. Snowpack has been decreasing over most of the West in recent decades … Mote et al., BAMS, 2005 … and spring streamflow has been arriving earlier. Spring-pulse dates Stewart et al., 2004 Evidence from California’s Sierra Nevada: Mortality rate (% yr-1) 2.5 280 2.0 1.5 240 1.0 200 0.5 160 0.0 1985 1990 1995 2000 Summer water deficit (mm) (3-yr running mean) • Summer drought (water deficit) is increasing, due to increasing temperature (not decreasing precipitation). • Increasing tree mortality rates are being driven by increasing deaths due to insects, pathogens, and stress. 2005 Year van Mantgem & Stephenson, in prep. DEMOGRAPHY: possible futures Some water-limited forests may be primed for a southwestern-style die-back Credit: Craig Allen & NSF Annual temperature (˚C) Annual precipitation (mm) The recent southwestern drought was not exceptional (it was wetter than the1950s drought), but the temperature was higher Year Breshears et al., PNAS, 2005 But what about forests that are not primed for a similar die-back? Specifically, • Forests that are currently temperature-limited, not waterlimited (e.g., high-elevation forests, coastal rain forests). • Forests that may currently be water-limited, but that will experience substantially increased precipitation. We can get hints from natural productivity gradients. • Globally, forests of productive environments have higher turnover rates (mortality and recruitment) ... • ... at least partly because environments that favor tree growth also favor the organisms that kill trees. 5 -1 3 2 1 Tropical (Amazonia) Gy mn os pe xe d rm 0 Mi 0 4 e rm 1 27 27 Temperate only sp 2 30 g io 3 84 46 T R ic r o p i h e ca l rs o il s PToem orep resra oitle s 4 158 50 An Forest turnover (% yr ) Forest turnover (% yr-1) 5 Stephenson & van Mantgem, Ecol. Lett., 2005 Temperate (global) Consequences? 3 y = 2.76 - 0.00066 x -1 Forest turnover (% yr ) In the coniferous forests of the Sierra Nevada: a 4○ C increase in temperature is associated with a 0.5 %yr -1 increase in population turnover rate, potentially reducing average tree age by one third. r 2 = 0.49, P < 0.001 2 1 0 1500 2000 2500 3000 3500 Elevation (m) Stephenson & van Mantgem, Ecol. Lett., 2005 A possible future scenario ... “Benign” climatic changes (e.g., warmer and wetter) Increased forest turnover rates Smaller, younger trees Cascading effects on wildlife and biodiversity Reduced forest carbon storage A possible future scenario ... “Benign” climatic changes (e.g., warmer and wetter) Increased forest turnover rates Smaller, younger trees Cascading effects on wildlife and biodiversity Reduced forest carbon storage ... but what about changing growth rates? Tree and Forest Growth • Compared to biogeography, we know relatively little about long-term, broad-scale climatic controls on lifehistory processes of trees • Especially true in non-plantation, mountain ecosystem settings • Growth is an indicator of environmental factors influencing species and may be a surrogate for establishment. GROWTH: Ongoing Changes, 1850-1980 Factor 1 Factor 1= 27 Chronologies w/ high loadings Factor 2 Factor 2 = 12 Chronologies w/ high loadings 185 Tree-ring chronologies, traditional detrending, factor analysis Factor 1 = Drought-sensitive tree-ring collections Factor 2 = High elevation, maritime, high-latitude tree-ring collections ~ 146 chronologies don’t load highly on either continental pattern McKenzie et al. 2001. Can. J. For. Res. 31: 526–538. Recent growth of conifer species of western North America: assessing spatial patterns of radial growth trends GROWTH: Ongoing Changes Are Location and Species Dependent High Lat. & Low Elev. Low Lat. & High Elev. These patterns point to three kinds of growth limitation by climate: • Water limitation • Energy limitation • Some combination of water and energy These patterns primarily represent sites dendroclimatologists would choose. What about the rest of the forests in the West? How do we go from reconstruction-grade sites that tell us about the most sensitive trees … to more mechanistic responses that allow inferences for large areas of forests? GROWTH: Within a species’ range across biogeographical space, climate impacts depend on elevation PDO Peterson and Peterson. 2001. Ecology 82: 3330-3345. Mountain hemlock growth responds to climatic variability at annual and decadal time scales. GROWTH: Across forest types and species within a mountain range, climate impacts depend on physiography Nakawatase and Peterson. 2006. Can. J. For. Res. 36: 77-91. Spatial variability in forest growth – climate relationships in the Olympic Mountains, Washington. GROWTH: Within a watershed, for the same species, elevation affects growth-climate relationships Low elevation vs. Max. Sum. T High elevation vs. Prior PDO Case and Peterson. 2005. Can. J. For. Res. 35. Fine-scale variability in growth–climate relationships of Douglas-fir, North Cascade Range, Washington. Growth-limiting factors are not really elevation, latitude, physiography, or even biotic. These are all surrogates for different scales of climatic (water or energy) limitation, and point to the need for MULTI-SCALE, GRADIENT-BASED studies of climatic limitation of growth 411 cm = ANN PPT Quinault North (ONP) 219cm Thornton North (NCNP) 72 cm Robinson South (IPNF) 170cm Belly River South (GNP) Highest Elevation Climate Change Local climate Climate Variability Physiography North Topography Lowest Elevation South Climate Dimensions of the PSME Transect Mean Climate Data: DAYMET 1981-1997 Climatic Niche Dimensions: Thompson et al. 2001 ONP Standard Chronology (mod. Z index) NCNP IPNF GNP Within each park, the variability in treegrowth is similar across low, middle, and high elevations. Climate-Growth Correlations: Temperature VIC Climate Climate-Growth Relationships: Hydrological Variables Divisional Climate Summary: Growth-Climate Relationships • Most frequent patterns of correlations point to combined influence of (-) temperature and (+) precipitation during summer • Underscored by PDSI (+) and water balance deficit (-), esp. in IPNF and GNP. • Some cool season (+) temp. and (-) snow relationships, primarily in ONP and NCNP. Bonsai PSME, Saint Mary, Glacier National Park The magnitude of the correlation between seasonal hydrological variables and tree-growth depends on the position of the plot along a gradient of surplus water in the environment. GROWTH: Possible Futures Depends on species, climate regime, and changes in water vs. energy. If we had results for most western conifers, we could estimate responses. But we don’t. Yet. McCabe and Wolock. 2002. Clim. Res. 20: 19–29, 2002 A CALL TO ACTION We do a pretty good job of monitoring weather, snow, and hydrology. We need a complementary network of forest “gauging stations.” This network of forest “gauging stations” will have two primary goals: (2) Developing a mechanistic understanding (otherwise we are lost) Mortality rate (1) Change detection (complementary to remote sensing) h Hig Low In d iv gro idual t High w th re rate e w Lo e S it vity i t c du pro Stephenson et al., in prep. Such a network is taking shape in CORFOR (the Cordillera Forest Dynamics Network) http://mri.scnatweb.ch/content/view/88/30/ PLEASE JOIN US! (... if you have the right kind of data.)