Modeling orbital forcing of lake level change: Lake Gosiute (Eocene), North America

Abstract

The sedimentary record of Lake Gosiute, a lake that existed in southwestern Wyoming during the Eocene, contains evidence for lake level fluctuations thought to be caused by the earth's precession cycle. However, it is not clear how the effects of these orbital variations were transferred through the climate system and into the sedimentary record. We carry out a series of experiments using a general circulation model (GCM), a lake energy balance model and a lake water balance model to better understand the processes by which these orbital variations could have altered lake evaporation, on-lake precipitation and runoff from the lake's catchment. GCM simulations indicate significant differences in surface incident shortwave radiation between the two end-members of the precession cycle. These differences cause lake evaporation to be ∼25% higher when perihelion occurs at the summer solstice. GCM simulations also indicate significant seasonal changes in the amount of precipitation between the two end-members, but no change in the annual mean precipitation. Preliminary experiments with a lake water balance model show that local effects such as changes in vegetation, in snowmelt runoff, or in the area of mudflats surrounding the lake could have a large impact on lake level. However, more data need to be collected to determine the importance of these effects. Our results challenge previous interpretations of paleoclimate that were based on geologic data and simple assumptions regarding the effects of orbital variations on the water balance of the lake. In particular, we find that (1) changes in shortwave radiation may have been more important than changes in temperature or moisture in causing lake level fluctuations and (2) changes in catchment and lake characteristics should be further examined. In order to make an accurate reconstruction of past climatic change from a lake level record, climate system processes and local non-climatic variables must be considered explicitly.