National Park Service
U.S. Department of the Interior
Natural Resource Program Center
Climate and Ecosystem Change in the U.S. Rocky Mountains and Upper Columbia Basin: Historical and Future Perspectives for Natural Resource Management
1David B. McWethy, 2Stephen T. Gray, 3Philip E. Higuera, 4Jeremy S. Littell, 5Gregory T. Pederson, 6Andrea J. Ray, 1Cathy Whitlock
1Department of Earth Sciences, Montana State University, Bozeman, MT 59717
2Water Resource Data System, University of Wyoming, Laramie, WY 82071
3Department of Forest Resources, University of Idaho, Moscow, ID 83844
4Climate Impacts Group, University of Washington, Seattle, WA 98195
5U.S. Geological Survey, Northern Rocky Mountain Science Center, Bozeman, MT 59715
6NOAA Earth System Research Lab, Boulder, CO 80305
This information is distributed solely for the purpose of pre-dissemination peer review under applicable information quality guidelines. It has not been formally disseminated by the National Park Service. It does not represent and should not be construed to represent any National Park Service determination or policy.
McWethy D. B., S. T. Gray, P. E. Higuera, J. S. Littell, G. S. Pederson, A. J. Ray, and C. Whitlock, 2010. Climate and Ecosystem Change in the U.S. Rocky Mountains and Upper Columbia Basin: Historical and Future Perspectives for Natural Resource Management. Report for the National Park Service, Great Northern Landscape Conservation Cooperative and the Greater Yellowstone Coordinating Committee. XX p.
Executive summary 6
Climate controls and variability at different spatio-temporal scales 10
How can understanding climatic variability inform management? 14
The last 20,000 years of environmental change in the western United States 18
Drivers of Millennial-Scale Climate Variation 18
Glacial–Holocene transition (ca. 16,000-11,000 cal yr BP) 19
Mid-Holocene transition (ca. 7000-4000 cal yr BP) 21
Southern Canadian and Northern U.S. Rocky Mountains 21
Central U.S. Rocky Mountains and the Greater Yellowstone Area 24
Cygnet Lake 25
Southern U.S. Rocky Mountains 25
Upper Columbia Basin 26
What do paleoenvironmental records tell us about millennial scale climate variations? 28
The last 2000 years of environmental change 30
Primary drivers of change 30
Biophysical conditions 30
Case studies 33
Northern U.S. Rocky Mountains 33
Drought variability and ecosystem dynamics in Glacier National Park 33
Central U.S. Rocky Mountains and the Greater Yellowstone Area (CR-GYA) 35
Changing distributions of Utah juniper in response to climatic variability 35
Precipitation variability in Wyoming’s Green River Basin 36
Southern U.S. Rocky Mountains 36
Changing pinyon pine distribution in response to climatic variability 37
Upper Columbia Basin 38
Climate variation and fire-related sedimentation 38
What can we learn from the last 2000 years about decadal and centennial scale climate change? 40
20th century climate change and the instrumental record 42
Surface hydrology 44
Ocean-atmosphere interactions 45
Changes in storm track and circulation patterns 47
Ecological impacts 47
Northern U.S. Rocky Mountains 48
Surface hydrology 56
Ocean-atmosphere interactions 56
Central U.S. Rocky Mountains and the Greater Yellowstone Area 57
Ocean-atmosphere interactions 58
Southern U.S. Rocky Mountains 58
Surface hydrology 60
Ocean-atmosphere interactions 60
Upper Columbia Basin 60
Surface hydrology 62
Ocean-atmosphere interactions 62
What can we learn from 20th century observations? 63
Small changes can have large impacts 63
Shifting distributions and new norms 65
What can we expect in the future? 66
GCM projections for North America 66
Westwide climate: Statistically downscaled projections 68
Climate projections downscaled to specific alpine sites 71
Model projections of future climatic and hydrologic conditions 75
Downscaled model methodology 75
Climate projections for the western United States 77
Climate conditions 77
Surface hydrology 77
Extreme conditions: droughts, floods, heat waves 78
Productivity and phenology 78
Planning for the future 79
Summary conclusions 81
Literature Cited 82
Paleoenvironmental records provide critical information on past climates and the response of ecosystems to climatic variability. Ecosystems have changed in a variety of ways as a result of past climate change, and they will continue to do so in the future. At large scales, climate governs the distribution of vegetation across the landscape and acts as a strong control of important biophysical conditions (e.g., extent of mountain glaciers) and ecosystem processes (e.g., area burned by fire). Paleoenvironmental and instrumental records from throughout the western United States suggest that vegetation response to climate change varies along a hierarchy of temporal and spatial scales, and the responses range from wholesale shifts in biomes to small adjustments in forest density or structure. Anticipating how ecosystems may respond to ongoing and future climate change requires an understanding of the climate–ecological linkages on all these scales as well as cross-scale interactions that lead to abrupt responses and regime shifts.
Vegetation changes occurring on millennial time scales are related to changes in the seasonal cycle of solar radiation and its attendant effect on atmospheric circulation patterns and surface energy balances. After the ice sheets and local glaciers retreated 17,000–12,000 cal yr BP (calibrated years before present, equal to the number of calendar years before 1950), paleoenvironmental records from the Pacific Northwest and Rocky Mountains reveal a sequence of vegetation changes as a result of increasing temperatures and more effective precipitation. Initially, deglaciated regions were colonized by tundra communities and the climate was colder and probably drier than at present. After 14,000 cal yr BP, warmer and wetter conditions allowed present-day conifer taxa to expand first in open parkland and later as closed forest communities. By 11,000 cal yr BP, closed subalpine forests were widespread. The early Holocene (11,000–7000 cal yr BP) was a time of warmer and drier summer conditions than at present. Warmer temperatures led to an upslope shift in conifer ranges and xerothermic shrub communities occupied valley bottoms. Summer drought led to higher fire frequencies than at present in many ecosystems. A gradual cooling and increase in effective moisture in the mid-Holocene (7000–4000 cal yr BP) was followed by the cool, moist conditions (4000 cal yr BP–present).
Embedded with these millennial scale changes (largely related to variations in the seasonal cycle of solar radiation), are centennial climate variations such as the Medieval Climate Anomaly (ca. 950–1250 AD, AD = Anno Domini) and the Little Ice Age (ca. 1400–1700 AD). These variations had less dramatic impacts on vegetation, but records describe shifts in ecotone positions, including the upper and lower treeline, and disturbance regimes. Gridded tree-ring networks suggest that within these intervals were multidecadal “megadroughts” associated with tree mortality and fire. On annual to decadal scales, climate variations have led to disturbance events and shorter droughts. These events have shaped successional pathways, tree growth and mortality, and community structure. Case studies investigating ecological response to these changes provide important lessons for understanding how ecosystems may respond to ongoing and future environmental change:
The last century is an inadequate reference period for considering future climate change because it does not capture the range of natural climate variability that vegetation responds to or the magnitude of climate change projected for the near future. For example, managers often rely on the last several decades of fire occurrence as a baseline for managing different ecosystems in the West even though, because of fire suppression policies and fire elimination, fire activity of the last century is atypical of long-term historical patterns in much of the West and is unlikely to represent future conditions. Many vegetation types have evolved under a wide range of fire frequencies and intensities, calling into question the value of a static view of a fire regime, e.g., characterized by a mean fire return interval. To understand the full range of conditions that may be important for sustaining ecosystems in the future, long-term records of fire (e.g., centuries to millennia) provided by tree-ring and lake sediment data are essential.
Rapid climate transitions have occurred in the past and will likely occur in the future. In the past, the response of vegetation has been highly variable, suggesting an equally complex response to future climate change. Among the likely outcomes will be a highly individualistic response by different species, a reorganization of plant communities, and the likelihood of differential lags in the ability of species to stay in equilibrium with climate change.
Climate variability at large scales is often expressed in complex and asynchronous patterns across the U.S. Northern Rockies and Upper Columbia Basin, largely because of interactions with topography and other sources of spatial heterogeneity. This heterogeneity can result in nearby communities showing different directions of change, (e.g., depending on the location of shifts and changes in the intensity of precipitation regimes changing in opposite directions at different elevations).
Many western ecosystems represent assemblages that formed as a result of a specific sequence of climate conditions during the last several thousand years. In particular, many middle and late-successional communities were established during the colder, wetter Little Ice Age and would be unlikely to form under current climate and land use conditions. Likewise, restoration to historical baselines is at best challenging and in most cases impossible.
Temperatures are projected to warm 1–5˚C for much of the West by 2100, accompanied by declines in snowpack, earlier spring snowmelt, and reduced late-summer flows. While projections for future precipitation are less certain, increased precipitation is unlikely to offset increased evapotranspiration associated with even modest warming (e.g., 1–2˚C), particularly during the summer. Consequently, drought is projected to increase in frequency and intensity over the next several decades, particularly in the Southwest and southern Rockies.
Investigations of the past suggest that we should expect dynamic and rapid ecosystem response to changing climate conditions, but in ways that may be difficult to predict. Paleoenvironmental records illustrate that while existing biomes have experienced distribution shifts, they have been resilient to climatic change across multiple time scales (e.g., decadal to millennial). They also suggest, however, that we should anticipate increased extreme and unforeseen disturbance synergisms; increased tree mortality, shifts in treeline position, and non-native plant invasions; and ultimately changes in plant community composition, structure, and function that may constitute novel vegetation assemblages. This poses significant challenges for developing management plans, but should not deter an adaptive management approach that allows for reassessment and modification of management strategies in the coming decades. By using scenario planning, managers can consider a wide range of possible future conditions to examine potential trajectories of ecological change. Managing for future conditions will, at the least, involve a continuation of well-established resource management and conservation practices.
We would like to acknowledge all the participants of the 2009 Rocky Mountain Inventory & Monitoring Technical Committee Meeting. A NPS steering committee including Judy Visty, Kathy Tonnessen, Lisa Garrett, Tom Rodhouse, Tom Olliff, David McWethy, Stacey Ostermann-Kelm, Bruce Bingham, and Penny Latham provided useful comments throughout the process. An interagency committee organized through the Great Northern Landscape Conservation Cooperative also provided guidance and advice throughout the writing and development process. This committee included: Yvette Converse (USFWS), Mike Britten (NPS), Tom Olliff (NPS), Molly Cross (Wildlife Conservation Society), Steve Gray (WY state climatologist), Beth Hahn (USFS), Virginia Kelly (Greater Yellowstone Coordinating Committee), Tim Mayer (USFWS), Jim Morrison (USFS), Stacey Ostermann-Kelm (NPS), Greg Pederson (USGS), David Wood (BLM), Andrea Ray (NOAA), and Lou Pitelka (NEON).
Funding was provided by the Greater Yellowstone Coordinating Committee and the NPS Inventory & Monitoring Program. Conceptual models were developed and created by Robert Bennetts. Maps were created by Meghan Lonneker.
At large spatial and temporal scales, climatic conditions act as primary controls shaping the structure and distribution of ecosystems and the species they support. Changes in climate have dramatically altered ecosystem dynamics by shifting plant communities, creating opportunities for recruitment of new species, and restructuring land-surface processes and nutrient cycles (Solomon et al. 2007). The controls of climate vary at different time scales (i.e., millennial, centennial, multidecadal, decadal, annual and interannual), and the ecological response to climate change varies accordingly. Paleoecological data show that ecosystems in the West have undergone significant and sometimes rapid changes since the Last Glacial Maximum (ca. 20,000 years ago), and the biotic assemblages observed today are relatively recent phenomena (Thompson et al. 1993: Whitlock and Brunelle 2007; Jackson et al. 2009a). The future is also likely to be characterized by rapid biotic adjustment, including the possibility of novel assemblages as species respond individualistically to climate change (MacDonald et al. 2008, Williams et al. 2007). Understanding the drivers and rates of past climate change and the sensitivity of ecosystems to such changes provides critical insight for assessing how ecological communities and individual species will respond to future climate change (MacDonald et al. 2008; Shafer et al. 2005; Whitlock et al. 2003; Overpeck et al. 2003; Swetnam and Betancourt 1999).
The purpose of this report is to provide land and natural resource managers with a foundation of both climate and ecosystem response information that underpins management-relevant biophysical relationships likely to play an important role over the coming decades. We begin by synthesizing the climate and vegetation history over the last 20,000 years following the retreat of late Pleistocene glaciers. This time span provides examples of ecosystem responses to long-term (e.g., millennial) climate warming as well as several well-known periods of rapid climate change (i.e., substantial decadal to centennial scale climate perturbations). To contextualize past climate and ecosystem changes, and to provide a best estimate of future climate conditions, we also report on the most current statistically and dynamically downscaled Global Climate Model (GCM) projections of future changes in key climate variables (e.g., precipitation, temperature, snow water equivalent, water deficit [CIG 2010]). Overall, our objective is to use the past to highlight a range of climate-driven biophysical responses to illustrate potential system trajectories and associated uncertainties under future climate conditions.
To meet the requested needs of the National Park Service (NPS), the U.S. Forest Service and the U.S. Fish and Wildlife Service, the geographic scope of this report encompasses the core regions of the Great Northern Landscape Conservation Cooperatives and the NPS high-elevation parks of the Rocky Mountains. For organizational purposes, the report divides the study area into four regions: the northern U.S. Rocky Mountains, the central U.S. Rocky Mountains and Greater Yellowstone Area, the southern U.S. Rocky Mountains, and the Upper Columbia Basin (fig. 1).
The synthesis is organized into four sections: (1) biophysical responses and drivers of climate changes occurring on multi-centennial to millennial scales during the last 20,000 years; (2) biophysical responses and drivers of climate change on annual to centennial scales over the last 2000 years; (3) the last century of climate and ecosystem change as observed by high-resolution instrumental records; and (4) the next century of likely future climate and ecosystem changes under a range of greenhouse gas emission scenarios. For each section, we present the large-scale and regional drivers of climate change as inferred from GCMs and paleoenvironmental data. We then detail the associated biophysical and ecological responses documented in both modern and paleoecological proxy data with a focus on the implications for maintaining key resources in the face of changing conditions. The synthesis concludes with a discussion of challenges in planning for future conditions where there is high uncertainty about climatic change and ecological response, and provides a planning approach designed to address a wide range of potential conditions. The purpose of this review is to highlight important climate and ecosystem linkages using records relevant to the regions of great conservation and natural resource management value shown in figure 1. Accordingly, this document does not necessarily provide a comprehensive review of the suite of available climate and ecosystem-related research available for the entire study region.
Figure 1. Location of study area. Climate region boundaries modified from Littell et al. (2009a), Kittell et al. 2002, and Bailey’s (1995) ecoprovince boundaries. Climate regions represent coarse aggregations of biophysical constraints on modern ecological assemblages and the interaction between climate, substrate, elevation, and other conditions. (Figure created by M. Lonneker, NPS, 2010.)
Climate controls and variability at different spatio-temporal scales
The influence of variation in climate on ecosystems changes at different spatial and temporal scales. Understanding the potential influence of climate on biophysical processes often requires local to synoptic (regional or larger) information on the type and magnitude of forcings (mechanisms driving changing conditions), e.g., relative humidity-related changes versus changes in solar radiative output, long-lived greenhouse gasses, and ocean-atmosphere interactions and local albedo (low albedo = low reflectance or dark ground surface such as pavement; high albedo = high reflectance and lighter surface such as snow) along with an understanding of the sensitivity of the ecosystem property that is being measured (Overpeck et al. 2003; Webb and Bartlein 1992) (Table 1). It is also important to recognize the hierarchical nature of climate variation and change, and that short-term (i.e., daily to interannual) events are superimposed on longer ones, amplifying or dampening the magnitude of the underlying physical controls that influence ecosystem dynamics.
Table 1. Climatic variation at different time scales and biotic response (modified from Overpeck et al. 2003).
Frequency and Scale of Variation (years)
Kind of Variation
Deglacial and postglacial variations
Ice sheet size, insolation, trace gases, regional ocean-atmosphere-ice interactions
Species migration, range expansion and contraction; community reorganization and establishment, species extirpation/extinction
Decadal and centennial anomalies
Internal variations in the climate system, solar variability, volcanism
Shifts in relative abundance and composition of different taxa through recruitment, mortality, and succession
Storms, droughts, ENSO events
Internal variations in the climate system, solar variability, volcanism
Adjustments in physiology, life history strategy, and succession following disturbance
At the longer temporal range of major changes in Earth’s climate system, variations on scales of 10,000 to 100,000 years are attributed to slowly varying changes in Earth’s orbit known as Milankovitch Cycles (i.e., changes in Earth’s precession, tilt, and obliquity [Hays et al. 1976, Berger 1978, Berger and Loutre 1991], fig. 2). The net result of slowly varying changes in Earth’s orbit have included multiple glacial and inter-glacial cycles driven by changes in global average temperatures over the past several million years. These temperature changes are initiated by changes in the seasonal and annual cycle of insolation (incident solar radiation; the amount of solar radiation received on a given surface area during a given time) over the high latitudes of the Northern Hemisphere that result in substantial positive feedbacks from changing concentrations of atmospheric greenhouse gasses (Vettoretti and Peltier 2004; Bond et al. 2001; Kutzbach and Guetter 1986; Kutzbach et al. 1998, 1993).
Figure 2. Primary drivers of climate and resulting climate variations at millennial, centennial, and interannual scales. (A) Temperature reconstruction from the central Greenland (the GISP2) ice core record and the forcing mechanisms thought to influence variation during the glacial period (ca. 49,000–12,000 cal yr BP). The Laurentide ice sheet in North America began to recede and climate warming commenced ca. 17,000 cal yr BP. Gray bands indicate two abrupt climate changes: the Younger Dryas Chronozone ca. 12,900–11,600 BP, and the 8200 cal yr BP cool event (8.2 ka event); (B) Temperature anomalies for the Northern Hemisphere based on multiple proxy data (e.g., ice core, ice borehole, lake sediment, pollen, diatom, stalagmite, foraminifera, and tree-ring records) from Moberg 2005 (black line), Mann and Jones (2003, red line), and the instrumental record (blue) for the past 2000 years (Viau et al. 2006). Continental patterns of drought and interannual and decadal climate variability are associated with the Medieval Climate Anomaly (ca. 950–1250) and fewer fires during the Little Ice Age (ca. 1400–1700). (Time spans from Mann et al. 2009 but varies by region [Cook et al. 2004, MacDonald et al. 2008, Bradley et al. 2003, Carrara 1989]); (C) Recent global temperature anomalies (black line, based on 1900–2000 mean) from HadCRUT3v instrumental reconstruction (Brohan et al. 2006) and ocean-atmosphere variability. Magenta line represents the Multivariate El Niño–Southern Oscillation (ENSO) Index, which is based on six observed ocean-atmosphere variables. Positive values of the index depict El Niño events (Wolter and Timlin 1993, 1998). Source: NOAA Earth System Research Laboratory.
On multi-millennial scales (here specific to the last 20,000 years), the presence or absence of the large North American ice sheets, particularly the Laurentide ice sheet results in ocean-ice-atmosphere interactions that drive changes in atmospheric circulation patterns (i.e., the position of westerlies and preferential positioning of storm tracks [fig. 2a]), resulting in major changes in ecosystem distribution and structure. For example, as Northern Hemisphere summer insolation increased and the ice sheets and glaciers began to retreat, seasonal storm tracks shifted north? and paleoecological records show widespread reorganization of plant communities throughout the West (e.g., MacDonald et al. 2008; Thompson et al. 1993; Bartlein et al. 1998; Whitlock and Brunelle 2007; Jackson et al. 2005; Betancourt et al. 1990). These large-scale and long-term changes in insolation and ice cover? are important features of Earth’s climate dynamics because they influence the persistence and strength of storm tracks, subtropical high-pressure systems, ocean-land temperature gradients, and consequently interannual to decadal scale drivers of climate variability such as the El Niño–Southern Oscillation (ENSO). For example, higher-than-present summer insolation in the Northern Hemisphere during the early Holocene (ca. 11,000–7000 cal yr BP) led directly to increased summer temperatures and indirectly to a strengthening of the Pacific subtropical high-pressure system off the northwestern United States, effectively intensifying summer drought in the region (Bartlein et al. 1998). Records of early Holocene glacier dynamics, lake levels, aeolian activity (blowing dust), vegetation, and fire show the ecological effects of this increased summer insolation at local to subcontinental scales (e.g., Whitlock and Brunelle 2007; Whitlock et al. 2008; Jackson et al. 2009a; Luckman and Kearney 1986; Osborne and Gerloff 1997; Rochefort et al. 1994; Graumlich et al. 2005; Fall 1997; Booth et al. 2005; Dean et al. 1996; Dean 1997).
On shorter and perhaps more management-relevant time scales, climate variations at interdecadal to centennial scales are related to changes in solar activity, volcanism, sea-surface temperature, and pressure anomalies in both the Atlantic and Pacific oceans. More recently, important contributions arise from rapidly increasing atmospheric greenhouse gas concentrations (Barnett et al. 2008; fig. 2b,c). The Medieval Climate Anomaly could be considered an example of centennial-scale climate variation due to the relatively warm and dry conditions that prevailed across the western United States from approximately 900 to 1300. The West experienced substantially reduced streamflows (Meko et al. 2007), shifts in the upper treeline (Rochefort et al. 1994; Fall 1997), and increased fire activity (Cook et el. 2004). Though the exact causes of the MCA are still debated, the prevailing evidence suggests that it was driven by changes in solar activity, volcanism, and perhaps sustained La Niña-like conditions in the tropical Pacific (Mann et al. 2009). At decadal to interdecadal scales, sustained sea surface temperature anomalies in the north Pacific and Atlantic oceans appear to be important drivers of climate variability across western North America (e.g., McCabe et al. 2004; Einfeld et al. 2001). The major indices that capture these modes of interdecadal variability include the Pacific Decadal Oscillation (PDO; see Mantua et al. 1997) and the Atlantic Multidecadal Oscillation (AMO; see Enfield et al. 2001, Regonda et al. 2005). Decadal climate shifts associated with changes in the PDO and AMO are well expressed in 20th century records of drought and winter precipitation (e.g., Cayan et al. 1998; McCabe et al. 2004), as well as in proxy-based reconstructions of precipitation and streamflow (fig. 2c; e.g., Gray et al. 2003). These events are often regional to subcontinental in scale and initiate and terminate within years, but often have widespread physical and ecological effects (e.g., Allen and Breshears 1998, Bitz and Battisti 1999, Pederson et al. 2004, Watson and Luckman 2004). Examples include widespread bark beetle outbreaks, increased forest fire activity and stress-related tree mortality, and rapid changes in glacier mass balance, snowpack, and streamflow.
The El Niño–Southern Oscillation (ENSO) is a major global control of both temperature and moisture patterns (fig. 2c). ENSO events are defined by changes in atmospheric pressure gradients across the tropical Pacific that are related to patterns of warming (El Niño) and cooling (La Niña) in the central and eastern equatorial Pacific which typically last 6–18 months and reoccur every 2–7 years (Bjerknes 1969; Cane and Zebiak 1985; Graham and White 1988; Horel and Wallace 1981; Philander 1990; Chang and Battisti 1998). The magnitude of these sea surface temperature anomalies varies, but they typically exert a substantial influence on regional temperature and precipitation patterns (McCabe et al. 2004, Einfeld et al. 2001). For example, El Niño events typically result in warm dry conditions across the northwestern United States and a southerly displacement of the winter storm track (i.e., the jet stream), resulting in cool wet conditions across the Southwest (McCabe et al. 2004; Enfield et al. 2001). The inverse is typically true for La Niña-like conditions, but in all cases this mode of climate variability appears to exert its strongest influence across the Southwest (e.g., Swetnam and Betancourt 1998), with an important but somewhat attenuated spatial signature across the Northwest (e.g., Dettinger and Ghil 1998).
In summary, climate-ecosystem linkages are evident across many time scales, from individual records as well as regional and global compilations. Shifts in species distributions and abundance are a response to climate variations occurring over seasons to millennia. Knowledge of climate drivers on all time scales is necessary to identify temporal and spatial dimensions of future changes and possible ecological responses to those changes.