Rise and fall of late Pleistocene pluvial lakes in response to reduced evaporation and precipitation: Evidence from Lake Surprise, California

Widespread late Pleistocene lake systems of the Basin and Range Province indicate substantially greater moisture availability during glacial periods relative to modern times, but the climatic factors that drive changes in lake levels are poorly constrained. To better constrain these climatic factors, we present a new lacustrine paleoclimate record and precipitation estimates for Lake Surprise, a closed basin lake in northeastern California. We combine a detailed analysis of lake hydrography and constitutive relationships describing the water balance to determine the influence of precipitation, evaporation, temperature, and seasonal insolation on past lake levels. At its maximum extent, during the last deglaciation, Lake Surprise covered 1366 km(2) (36%) of the terminally draining Surprise Valley watershed. Using paired radiocarbon and Th-230-U analyses, we dated shoreline tufa deposits from wave-cut lake terraces in Surprise Valley, California, to determine the hydrography of the most recent lake cycle. This new lake hydrograph places the highest lake level 176 m above the present-day playa at 15.19 +/- 0.18 calibrated ka (C-14 age). This significantly postdates the Last Glacial Maximum (LGM), when Lake Surprise stood at only moderate levels, 65-99 m above modern playa, similar to nearby Lake Lahontan. To evaluate the climatic factors associated with lake-level changes, we use an oxygen isotope mass balance model combined with an analysis of predictions from the Paleoclimate Model Intercomparison Project 3 (PMIP3) climate model ensemble. Our isotope mass balance model predicts minimal precipitation increases of only 2%-18% during the LGM relative to modern, compared to an similar to 75% increase in precipitation during the 15.19 ka highstand. LGM PMIP3 climate model simulations corroborate these findings, simulating an average precipitation increase of only 6.5% relative to modern, accompanied by a 28% decrease in total evaporation driven by a 7 degrees C decrease in mean annual temperature. LGM PMIP3 climate model simulations also suggest a seasonal decoupling of runoff and precipitation, with peak runoff shifting to the late spring-early summer from the late winter-early spring. Our coupled analyses suggest that moderate lake levels during the LGM were a result of reduced evaporation driven by reduced summer insolation and temperatures, not by increased precipitation. Reduced evaporation primed Basin and Range lake systems, particularly smaller, isolated basins such as Surprise Valley, to respond rapidly to increased precipitation during late-Heinrich Stadial 1 (HS1). Post-LGM highstands were potentially driven by increased rainfall during HS1 brought by latitudinally extensive and strengthened midlatitude westerly storm tracks, the effects of which are recorded in the region's lacustrine and glacial records. These results suggest that seasonal insolation and reduced temperatures have been under-investigated as long-term drivers of moisture availability in the western United States.