Hydrologic patterns are imprinted in the geologic record and play a prominent role in the investigations of Pangean climate. However, the hydrologic aspects of climate are complex (1) many variables are required to analyze hydrology, (2) hydrologic processes act on spatial scales that are smaller than GCM (General Circulation Model) grid spacing, and (3) observational data bases for calibrating modern climate simulations are few. This study, which uses a GCM to depict arid climates of the past and present, is designed to evaluate a variety of simulated hydrologic variables that are becoming increasingly available for paleoclimate model/data comparison studies.

Simulations of the current climate show that there are quantifiable levels of precipitation (P), soil moisture, and surface runoff that delineate the locations of modern deserts. However, in a Pangean simulation, using the same model, the threshold values of these variables reveal disparate views of desert extent. Altered boundary conditions and atmospheric circulation can invalidate thresholds based on modern climatology. The credibility of these variables thus rests heavily on the accuracy of reconstructed boundary conditions. A more fundamental meteorological approach defines deserts as regions where the atmospheric demand for moisture (potential evapotranspiration, E_p) exceeds the supply (precipitation, P). GCM values of E_p are overestimated; thus simulated P–E_p fields are unusable. However, simulated Ep values can be derived independently using GCM-produced temperature values in empirically derived equations (E_Tp ). Similar methods, using observed temperatures, are employed by atmospheric scientists to construct global and regional E_p data sets. The results of this study indicate that P–E_Tp is more accurate than other individual variables for calculating moisture balance and is more useful for paleoclimate model/data comparisons because of its fundamental meteorological basis and lack of dependence on prescribed vegetation and soil conditions. Ultimately, hydrological misinterpretations can be avoided by understanding the limitations of GCM parameterizations and by strictly cross-checking variables with each other and with available paleoclimate data.