National Park Service U. S. Department of the Interior Natural Resource Program Center

Knowledge of natural variations in climate at different scales provides a context for understanding how communities and individual species might respond to current and future rates of change (Table 2). For example, past changes in fire regimes have been largely driven by large-scale climate changes on millennial, centennial, decadal and annual scales (Whitlock et al. 2003; Kitzberger et al. 2007; Littell et al. 2009a; Westerling et al. 2006). In ecosystems where fire regimes are expected to change with future climate conditions, management efforts should focus on the ecological response to rapidly changing conditions as opposed to maintaining current or past conditions (Whitlock et al. 2010). Additionally, paleoenvironmental records showing evidence of rapid changes in climate and attendant ecological responses suggest that even small changes in climate can have large consequences and provide an important context for anticipating ecosystem response to future climate change (Whitlock and Brunelle 2007; Gray et al. 2006, 2004; MacDonald et al. 2008; Lyford et al. 2003).

Reconstructing past climatic conditions and the associated ecological response involves a number of direct and indirect measurements. Direct measurements include ground temperature variations, gas content in ice core air bubbles, ocean sediment pore-water change, and glacier extent changes. Indirect measurements or paleoclimate proxy typically come from organisms that respond to changes in climate through changes in their growth rate, abundance, or distribution as recorded in living or fossil specimens or assemblages of organisms. Each proxy indicates past change at different spatial and temporal scales and resolutions.

L
Obtaining lake sediment core, Rocky Mountain National Park. Photo courtesy P. Higuera.
ake sediment, tree-ring cores, and packrat middens are three of the primary proxies used for reconstructing past conditions for the western United States (Whitlock and Larsen 2001; Fritts and Swetnam 1989; Betancourt 1990). Lake sediment cores often provide some of the longest records of vegetation and fire through analysis of pollen, plant macrofossils, and charcoal particles at intervals throughout the core. Most lakes in the Northwest were formed during deglaciation, and therefore provide a sedimentary record spanning the last 15,000 years or longer, depending on the time of ice retreat. Sediment cores are retrieved from modern lakes and wetlands using anchored platforms in summer or from the ice surface in winter. Samples for pollen and charcoal analyses are removed from the cores at intervals (e.g., every 0.5–1 cm) that depend on the detail and temporal resolution required. The pollen extracted from the sediment is chemically treated, identified under the microscope, and tallied for each sediment level sampled. Pollen counts are converted to percentages of terrestrial pollen and plotted as a diagram.

The reconstruction of past vegetation and climate from pollen percentages rests on the relationship between modern pollen rain and present-day vegetation and climate. Modern pollen samples have been collected at lakes throughout North America, and this information is calibrated to modern vegetation and climate. Past fire activity is inferred from the analysis of particulate charcoal, which is extracted and tallied from the sediment cores (Whitlock and Bartlein 2004). High-resolution charcoal analysis involves extraction of continuous samples from the core such that each sample spans a decade or less of sediment accumulation. These samples are washed through sieves and the charcoal residue is tallied under a microscope. Examining these relatively large particles enables a local fire reconstruction because large particles do not travel far from a fire. Charcoal counts are converted to charcoal concentration (particles/cm³), which is then divided by the deposition time of each sample (yr/cm) to yield charcoal accumulation rates (particle/cm²/yr). Detection of fire events involves identification of charcoal accumulation rates above background levels.

T

M

The last 20,000 years of environmental change in the western United States

The climate variations of the last 20,000 years occurring on millennial scales are best understood through model-based simulations that look at the regional response to large-scale climate changes and paleoenvironmental data that measure specific components of climate change. Broad-scale climate variations were described by Bartlein et al. (1998) using the NCAR Community Climate Model (CCM1; 4.4º latitude by 7.5º longitude spatial resolution, mixed-layer ocean, crude depiction of western cordillera topography). The simulation provided estimates of climatic conditions over six discrete periods (21,000 cal yr BP=full glacial period with full-sized ice sheets; 16,000 to 14,000 cal yr BP =late Glacial period with shrinking ice sheets; 11,000 cal yr BP = early Holocene insolation maximum; 6000 cal yr BP= mid-Holocene transition; and present). Other recent regional-scale modeling studies have provided better temporal and spatial resolution with more realistic topography (Hostetler 2009; Hostetler et al. 2003). Results from these efforts are summarized below.

At the time of the Last Glacial Maximum (ca. 21,000 cal yr BP), the large Laurentide and Cordilleran ice sheets strongly influenced climatic conditions in the western United States (Bartlein et al. 1998; fig. 3), depressing temperatures approximately 10ºC in areas south of the ice sheets and steepening the latitudinal temperature gradient. The presence of the large ice sheets also displaced the jet stream south of its present position, greatly reducing winter precipitation in the northwestern United States and Canada while increasing precipitation across the Southwest. Another element of the full-glacial climate was stronger than present surface easterlies related to a strong high-pressure system that persisted over the ice sheets. The presence of this strong high-pressure systemis steepened the latitudinal temperature gradient and weakened the westerly storm tracks, resulting in colder and effectively drier conditions across the Northwest and cold wet conditions in the Southwest (Hostetler et al. 2000; Bartlein et al. 1998; Thompson et al. 1993).

During the Glacial–Holocene transition (16,000–11,000 yr BP), solar insolation over high-latitude Northern Hemisphere landmasses increased, peaking ca. 11,000 yr BP when summer insolation was 8.5% higher and winter insolation was 10% lower than at present at 45ºN latitude. One consequence was a northward shift of winter storm tracks from their full-glacial position, bringing wetter winter conditions to the Northwest while the Southwest became increasingly dry (Bartlein et al. 1998). Increasing summer insolation resulted in warmer growing season temperatures, causing alpine glaciers and ice sheets to melt rapidly. (See fig. 4 for an example of modern air mass circulation.)

At the end of the Pleistocene (the last several 2.5 million years of repeated glaciations prior to the Holocene), much of the Northern Hemisphere experienced an abrupt cooling known as the Younger Dryas Chronozone (YDC, ca. 12,900–11,600 BP) (Alley et al. 1993). This event is clearly registered in the North Atlantic region and across Europe, and is related to changes in ocean circulation during the melting of the Laurentide ice sheet. Evidence of the YDC is less obvious in the western United States, where most paleoenvironmental data show little or no response in terms of glacial activity (Licciardi et al. 2004; Heine 1998), vegetation change (Grigg and Whitlock 2002; Briles et al. 2005; Brunelle et al. 2005; Huerta et al. 2009), or shifts in fire activity (Marlon et al. 2009). In some areas, however, there is evidence of a re-advance of mountain glaciers (Osborn and Gerloff 1997; Reasoner and Huber 1999; Friele and Clague 2002; Menounos and Reasoner 1997) and vegetation changes that indicate cooling, including a lower treeline in central Colorado (Reasoner and Jodry 2000) and changes in the isotopic signature in speleothem data. Temperatures decreased as much as 5–10ºC (9–18°F) in Greenland, but pollen data for the northwestern United States indicate much more moderate cooling (0.4–0.9ºC [0.7–1.6°F]; Reasoner and Jodry 2000). The YDC ended rapidly with warming of ~7˚C (13°F) in Greenland occurring within one to several decades (Alley 2000). Consequently, the period is closely scrutinized as an example of abrupt climate change (Alley et al. 2003).

A similar abrupt cooling occurred ca. 8200 cal yr BP (figs. 2–3) when a large influx of fresh water disrupted circulation in the North Atlantic, causing cooling that lasted several centuries in the North Atlantic region (Le Grande et al. 2006; Alley and Ágústsdóttir 2005). Like the YDC, the 8.2 ka event is not registered at many sites in the western United States, either because the signal is weak or the sampling resolution is inadequate to detect it. Geochemical proxies from lake sediments suggest that it has been associated with drier conditions at Bear Lake, Utah (Dean et al. 2006). The YDC and the 8.2 ka event illustrate how rapid climate changes due to ocean-atmosphere-ice interactions can occur. They also show a variable signal in regions distal to the North Atlantic origin, so that some sites in the western United States show a response while others do not.

Figure 4.

Greater summer insolation (8% above present) and lower winter insolation (8% lower) in the early Holocene (11,000–7000 cal yr BP) profoundly affected the climate and ecosystems of the western United States. Increased summer insolation led to warmer temperatures throughout the region during the growing seasons, while winters were likely colder than at present. Model simulations show that increased insolation led indirectly to a strengthening of the eastern Pacific subtropical high-pressure system which suppressed summer precipitation in much of the Northwest. At the same time, it strengthened inflow of moisture from the Gulf of California to the Southwest and the southern and central Rocky Mountain region (Whitlock and Bartlein 1993), resulting in greater than present summer precipitation. East of the Rockies, increased summer precipitation was likely offset by increased temperatures and rates of evapotranspiration, making conditions effectively drier than at present, which is consistent with low lake levels and dune activation (Shuman et al. 2009; Stokes and Gaylord 1993). Today the West is characterized by summer-dry areas under the influence of the subtropical high (i.e., the Northwest) and summer-wet areas where summer precipitation reflects monsoon activity (i.e., the Southwest). These two precipitation regimes are defined by topography (e.g., Yellowstone Plateau) and the boundary between them is relatively sharp (Whitlock and Bartlein 1993; Gray et al. 2004). The indirect effects of greater-than-present summer insolation strengthened both precipitation regimes, making summer-dry regions drier in the early Holocene and summer-wet regions wetter than at present (Thompson et al. 1993; Whitlock and Bartlein 1993; Bartlein et al. 1998). During the mid-Holocene (ca. 7000–4000 cal yr BP) and the late Holocene (4000 cal yr BP–present), summer insolation decreased and winter insolation increased gradually to present levels. Summer-dry regions became cooler and wetter, and summer-wet regions became cooler and drier than before.

In the Pacific Northwest, including the Columbia Basin, the mid-Holocene was a transition period between the warm, dry, early Holocene and the cooler, wetter, late Holocene. In British Columbia, Hebda and Mathewes (1984) call this period the mesothermal and trace the expansion of hemlock (Tsuga heterophylla) and Douglas fir (Pseudotsuga menziesii) during this period. This signal is also evident in the northern Rockies and perhaps as far south as the Greater Yellowstone Area. In the southern and central Rockies, both summer-wet regions, paleoclimate data suggest several anomalously dry/wet periods during the Holocene (Shuman et al. 2009; Stone and Fritz 2006). Lake level data, for example, indicate that numerous sites throughout the Rockies experienced drier conditions during the mid-Holocene (ca. 6000 cal yr BP) than at present (Shuman et al. 2009), similar to climatic conditions across the Great Plains. Anomalously dry conditions for the central and southern Rockies also contrast with wetter-than-present conditions in the Southwest (Betancourt et al. 1990; Davis and Shafer 1992; Thompson et al. 1993; Fall 1997; Mock and Brunelle-Daines 1999; Harrison et al. 2003). Regional climate simulations from the Colorado Rockies shed some light on the drivers of mid-Holocene aridity; variations in the seasonal insolation cycle imposed local surface feedbacks (e.g., reduced snowpack and soil moisture) that were important drivers of submillennial-scale changes in precipitation and moisture (Shuman et al. 2009; Diffenbaugh et al. 2006).

Prior to 14,000 cal yr BP, much of the southern Canadian and northern U.S. Rocky Mountains were glaciated (fig. 5). The first plant communities to colonize deglaciated regions were alpine in character, composed of shrubs and herbs dominated by sagebrush, grasses, and alder (Artemisia, Gramineae, and Alnus) (Whitlock et al. 2002; MacDonald 1989; Reasoner and Hickman 1989). By the early Holocene, warmer-than-present conditions allowed shrub and herb communities to spread upslope to higher elevations than at present, and forests of pine and alder (Pinus albicaulis/flexilis and Alnus) developed at mid-elevations. The lesser abundances of spruce and pine (Picea and Pinus contorta) suggest warmer growing season conditions than at present. Low-elevation forests were characterized by spruce and lodgepole pine (Pinus contorta), and valley floors supported open grassland with limber pine. The upper treeline was at least 90 meters above present during the early- and mid-Holocene (ca. 8500–3000 4000 cal yr BP). Low and middle elevations experienced the greatest fire activity during the early Holocene (ca. 9000–8000 cal yr BP) (MacDonald 1989) and forest composition resembled modern subalpine spruce-fir (Picea-Abies) forest. In the late Holocene, summer (growing) conditions became cooler, glacial activity increased, and the upper treeline decreased substantially (Reasoner and Hickman 1989). Cooler and moister conditions led to tundra at high elevations, spruce and fir forests at mid-elevations, and spruce, lodgepole pine, and aspen at low elevations.

Ecosystem response to centennial-scale climatic variations is evident from a 3800-year history of climate, vegetation, and ecosystem change inferred from pollen and charcoal concentrations in the lake sediment record from Foy Lake in northwestern Montana. Formed over 13,000 years ago as ice retreated from the Flathead Valley, the lake is situated at the eastern edge of the Salish Mountains, 3 kilometers southwest of Kalispell, Montana (Stevens et al. 2006). Several studies from the site provide historical reconstructions of climate and hydrologic variability and ecosystem response to climate change over the past several millennia (Stevens et al. 2006; Power et al. 2006; Shuman et al. 2009). Paleolimnologic and pollen data indicate that ca. 2700 cal yr BP, an abrupt rise in lake levels coincided with a transition from steppe and pine forest to pine forest-woodland to mixed conifer forest (Power et al. 2006), a transition linked to an increase in effective moisture (winter precipitation) shown in lake level records (Stevens et al. 2006; Shuman et al. 2009). Following the establishment of mixed conifer forests, lake levels decreased from 2200 to 1200 cal yr BP, and increases in grass, pine, and sagebrush and declines in Douglas fir and larch led to the development of a steppe/parkland/forest mosaic ca. 700 cal yr BP (Power et al. 2006; Stevens et al. 2006). Increases in grass and sagebrush in the late 19^th and early 20^th centuries coincided with human activities. Notable climatic events during this period include a long, intense drought ca. 1140 following a wetter period from 1050 to 1100 (Stevens et al. 2006).