PP22A-06
Multi-Record Validation of Centennial-to-Millennial Hydroclimate Trends and Variability During the Holocene in the Northeast United States

Tuesday, 15 December 2015: 11:35
2012 (Moscone West)
Bryan N Shuman, University of Wyoming, Geology & Geophysics, Laramie, WY, United States and Jeremiah Marsicek, University of Wyoming, Laramie, WY, United States
Abstract:
Long-term shifts in baseline hydroclimatic conditions present meaningful risks to ecosystems and society. Multi-decade megadroughts and pluvials illustrate these risks, but much less is known about lower frequency hydroclimatic variability. Here, we build upon previous work demonstrating coherent signals of multi-century lake-level trends and fluctuations over the past >6,000 years in the northeast U.S., which were reconstructed using systematic analyses of past shoreline positions in small ponds in coastal Massachusetts. Based on these records, water levels have risen substantially since 8 ka and repeated multi-century episodes of low water at ca. 4.9-4.6, 4.2-3.9, 2.9-2.1, and 1.3-1.2 ka punctuated the long-term rise. New records from inland Massachusetts, Connecticut, and Pennsylvania extend these patterns to the mid-Atlantic region. By converting the changes in lake volume to estimates of precipitation minus evaporation for each watershed using a simple hydrologic budget model, we compare the paleohydrologic record from lakes with annual precipitation reconstructions derived from fossil pollen data. The fossil pollen data were matched to their nearest modern equivalents, based on a dissimilarity metric, and assigned the annual precipitation associated with their space-for-time analogs. Such an approach accurately reconstructs spatial gradients in annual precipitation today (R2 = 0.70 for North America) with uncertainty <15% (RMSE = 165 mm/yr) for the northeast U.S. Over the Holocene, the mean lake and pollen records correlate well (R2 = 0.96). The comparison indicates that multi-millennial trends and additional multi-century hydroclimate variability impacted both surface hydrology (lakes) and vegetation composition. Because the lake and pollen records derive from independent sediment cores and types of analyses, and show correlated trends across multiple sites, the major features of the record appear robust. Additionally, alkenone-derived sea-surface temperature (SST) estimates from the Atlantic coast predict both the lake-level and pollen-inferred changes. Low SSTs correlate with high effective moisture, and indicate an integrated ocean-atmosphere component to Holocene hydroclimatic changes in the northeast U.S.