To estimate the potential consequences of projected global warming on streamflow, a deterministic model for a monthly timescale has been developed. The model structure is based on the water budget theory and contains deterministic relationships to estimate four components of streamflow: direct runoff, interflow, base flow, and snowmelt. The model inputs are six climate variables for each time step and 10 watershed parameters. The model has four calibration parameters which are related to direct runoff and snowmelt runoff. The model has been applied to the Little Washita River watershed in Oklahoma and the Baptism River watershed in Minnesota. The former is an agricultural watershed with a warm and seasonally dry climate, and the latter is a forested watershed with a cool and humid climate. The model simulates mean monthly streamflow in the Baptism River with a Nash-Sutcliffe coefficient (NSC) of 0.99 and a correlation coefficient of 0.83. For the Little Washita River the same simulation measures are 0.94 and 0.89, respectively. ¿ 1998 American Geophysical Union