We examine the influence of climate parameters on seasonal streamflow in watersheds over a range of elevations in California and Oregon. Effects of precipitation, temperature, and snow water content (SWC) are diagnosed using linear regression models and categoric composites. Most of the models explain over 60--80% of the seasonal streamflow variability. The models and the composites provide insight into the climate influences which drive the individual watersheds. Low (warmer) basins have little snow and little memory of prior seasons' climate variability. High (cooler) basins, with more snow, have longer memories. Precipitation has the greatest influence on streamflow variations in spring. Temperature is important in spring in the mid and high elevations. By late spring, SWC accounts for nearly all of the summer streamflow variation at mid and high elevations, but earlier in the runoff season, precipitation and temperature add variance. The variations in surface climate parameters, including streamflow, are generally controlled by atmospheric circulation anomalies with a spatial scale much larger than those in watersheds. This explains why little skill was lost by broadening the scale of temperature and precipitation predictors to regional climate areas. ¿ American Geophysical Union 1993