Uncertainties in Projecting Stream flows in Two Watersheds Under 2xC02 Climate Conditions

Abstract

Two surface water runoff models (SWAT and MINRUN96) were utilized to project changes in streamflows in two watersheds under a 2xC02 climate scenario, One is the Baptism River watershed (363 km2), in northern Minnesota, mainly forested, with a cold and humid climate, and the other is the Little Washita River watershed, in Oklahoma (538 km\ an agricultural watershed, with a warm and seasonally dry climate, 2xC02 climate conditions were obtained from the Canadian Climate Center General Circulation Model. No agreement was evident in the streamflows projected by the two watershed models for the Baptism River. Snow accumulation and snowmelt were the main differences between the two models. MINRUN96 simulated the past streamflow more accurately than SWAT. SWAT includes a biomass submodel which incorporates the effects of vegetation changes under 2xC02 climate conditions. SWAT also takes into account the effects of carbon dioxide in the estimation of evapotranspiration, while MINRUN96 does not contain any algorithm projecting the changes in vegetation or evapotranspiration due to doubling of atmospheric carbon dioxide. Both models projected more runoff in the Baptism River watershed in winter due to reduced snowfall under the 2xC02 climate scenario; however, the magnitudes of the projected increases were an order of magnitude different. SWAT was developed for agricultural watersheds and is probably not ,suited for application to forested watersheds at this time. For the Little Washita River watershed, the streamflows projected by the two models were in agreement. Both models showed an increase in fall and spring runoff and a decrease in summer runoff. The two models also proj ected a significant increase in the mmual streamflow of the Little Washita River under a 2xC02 climate scenario. The magnitUdes of the projected increases were comparable.