THE TESTING OF A MACROPHYSICAL CLIMATE MODEL WITH FIELD DATA IN WISCONSIN AND THE UPPER MIDWEST

Lake-based proxy records of Late-Pleistocene and Holocene climate change, as well as data from archaeological and geological sites throughout Wisconsin and the upper Midwest, are compared to site-specific “archaeoclimate” models of temperature, precipitation and evapotranspiration. Unlike a GCM, the archaeoclimate model is hierarchical – based on the boundary conditions that drive weather, not on the initial values for weather phenomena. External, or macrophysical, inputs used to force the model include Milankovitch parameters, global energy measurements, and past volcanism. These inputs permit calculation of global ice volume and the attenuation of solar radiation by aerosols, which in-turn allow calculation of global temperature back through time. Zonal temperature gradients can be derived from global temperature measurements and used to model the position of atmospheric features such as the jet stream, the subtropical highs, and the intertropical convergence. These atmospheric features are synoptically related to local weather patterns. At sites in the study area, greater than 95% of the local variability in monthly precipitation and temperature can be related to the position of circulation features. Assuming that these relationships have held constant through time, determination of the past position of these atmospheric features allows derivation of past values of temperature and precipitation for a specific location.

Broad-scale trends are observed in both the models and the field data. These include cooling around the Younger Dryas, a return to cool conditions or increased precipitation circa 9 ka, reduced summer precipitation and increased summer temperatures between 9 ka and 5 ka, and generally cooler and moister conditions in the last 5000 years punctuated by shorter-lived but regionally synchronous climatic events. One of these events occurs circa 4200 yr BP and another about 2100 yr BP. The models suggest cool and damp conditions regionally; however, field data are not in complete agreement, with evidence of drought at some sites. Model results also show that 2000 years ago sites in southern and eastern Wisconsin were wetter than present while those in northwestern Wisconsin were the same or slightly drier, a pattern consistent with the proxy record.