The Nonuniform Nature of Climate Change: Some Examples From the Great Basin of the Western United States

L V Benson (U.S. Geological Survey, 3215 Marine St., Boulder, CO

80302-1066; ph. 303-541-3005; fax 303-447-2505; email:; S.P. Lund (Department of Earth Sciences, University of Southern California, Los Angeles, CA 90089; ph. 213-740-5835; fax 213 740 8801; email:

Studies of Greenland ice cores and North Atlantic sediments indicate that centennial-to-millennial-scale climate variability occurred during the last ice age. Hydrologic-balance and glacial records from the Great Basin indicate that the climate of the western United States was also highly unstable during this period. In the Owens Lake basin, the highest amplitude oscillations in hydrologic balance occurred at the beginning (52,000 to 40,000 14C yr B.P.) and end (14,000 to 10,000 14C yr B.P.) of the last Sierran alpine glacial period. In addition, evidence from the Owens and Pyramid Lake basins indicate that at least 15 stade-interstade oscillations occurred in the Sierra Nevada between 52,000 and24,000 14C yr B.P. Can and should these oscillations be correlated with North Atlantic climate events?

Our ability to correlate climate oscillations across the Northern Hemisphere depends on several factors: the resolution of our marine and lacustrine records, age control for those records, the ability of our chosen climate indicators (proxies) to distinguish between the fundamental elements of climate (temperature, humidity, clouds, and wind), and intensity and integrity of the climate signal. Unlike most marine sediments, sample resolution for Great Basin lakes is very high (<10 yr). However, radiocarbon age controls for both lake and marine records are clouded by questions regarding the magnitudes of reservoir effects; and errors in 14C-based age models are often of the same magnitude as the frequencies of the oscillations being correlated. The use of paleomagnetic secular variation for intrahemispheric correlation of lake and marine records holds promise but it rests on the assumption of uniformity of magnetic change across vast distances. Proxies of change in the hydrologic balance of a lake such as 18O are functions of several elements of climate as well as the hydrologic state of the lake (closed or overflowing). And proxies of glacial oscillations such as total organic carbon (TOC), magnetic susceptibility, and lake-sediment chemistry directly or indirectly reflect the production of glacial rock flour and its input to a lake basin, a complex process that cannot be simply linked to a single element of climate.

The intensity and spatial integrity of North Atlantic climate events remain a matter of conjecture. The recent work of Severinghaus et al. (1998)[Nature 391:141] indicates that atmospheric methane concentrations increased within <30 yr after an abrupt temperature increase signaled by 18O in Greenland ice. The ice-core data suggest a synchronous warming of the Northern Hemisphere at the end of the Younger Dryas cold-dry interval. But how strong was that signal in different parts of the northern hemisphere? In the Sierra Nevada, the last substantial glacial oscillation (the Recess Peak) ended before 11,200 14C yr B.P. (Clark and Gillespie, 1997)[Quaternary International 38/39:21], suggesting a warming of the western US during the Younger Dryas chronozone. New data from the Pyramid Lake basin indicates that the last anomalous wet event in the northern Great Basin occurred entirely within the Younger Dryas chronozone. From these data records it would appear that the climate of the northern Great Basin was relatively warm and wet when the climate of the North Atlantic region was cold and dry. To the extent these observations are correct, climate change during one of the best studied of the Dansgaard-Oeschger events, the Younger Dryas, was not uniform across the Northern Hemisphere.

Evidence for regional nonuniformity of climate change is also apparent in climate records from the Great Basin. Oxygen-18 records from Owens, Mono, and Pyramid Lake basins indicate that all three basins witnessed a wet-dry-wet oscillation in climate between 18,000 and 13,000 14C yr B.P. However there was a profound north-south gradient in the effective wetness of the lake basins. For example, Owens Lake completely desiccated ~14,000 14C yr B.P. but the surface of Mono Lake stood higher than at any time in the historical period. In addition, 18O and TOC records

from the Owens Lake basin indicate that alpine glacial oscillations were out of phase with hydrologic balance oscillations between 52,000 and 40,000 14C yr B.P. and during 52,000 to 49,000 and 24,000 to 15,000 14C yr B.P stades, substantial wet-dry oscillations occurred, indicating that precipitation and temperature variations were not

phase locked.

It appears that millennial-scale climate oscillations may not have occurred uniformly across the Northern Hemisphere. A case can be made that certain of the Heinrich events were manifested as relatively dry periods in the Great Basin; however, it remains difficult to demonstrate that low Great Basin lake levels centered at 21,000 and 15,000 14C yr B.P. are uniquely related to Heinrich events 1 and 2. Instead, it can be argued that Milankovitch forcing of Laurentide ice sheet size caused the polar jet stream to shift south of Owens, Mono, and Pyramid Lakes at 21,000 14C yr B.P and north of these lakes at 15,000 14C yr B.P.

Examples of uniform change can be created by comparing the average climate of time slices of rather long duration; e.g., most and perhaps all climate records indicate that Stage 2 was colder than Stage 1 over the entire globe. However, as one begins to compare the synoptic climate of thinner time slices, spatial heterogeneity of precipitation and temperature fields soon become apparent; i.e., the spatial heterogeneity within each time slice begins to approach the weather endmember of the climate spectrum as the time slice approaches annual resolution.

It is clear that in the North Atlantic region, some climate proxies have clearly defined variability/cyclicity with time constants of 102 to 10 3 yr. This variability is correlatable over several thousand km. In the Great Basin, climate proxies also have clearly defined variability with similar (and even smaller) time constants, but the distance over which this variability can be correlated remains unknown. We suggest that there may be some minimal spatial scale (domain) within which climate varies coherently on centennial and millennial scales. The size of that domain will change with location and the time constant of climate forcing. We, therefore, suggest that one of the goals of future research should be to define the space/time structure of climate change with special attention to the millennial scale.