Actual evapotranspiration (Eta) is one of the largest components of the hydrologic budget and accounts for a majority of water lost from a watershed. It is primarily controlled by soil water availability, which is largely controlled by rainfall, and atmospheric demand (potential evapotranspiration). Consequently, Eta is sensitive to changes in meteorologic conditions. Understanding the relationship between Eta and controlling meteorologic variables across time and space is important for future predictions of Eta under a changing climate, especially in California where demand for surface and groundwater is high. A regression modeling approach was used to (1) determine the relative control of rainfall, rainfall intensity, and potential evapotranspiration (Etp) over annual and long-term mean annual Eta across watersheds in western California, and (2) quantify the sensitivity of watershed annual Eta to changes in these variables. Annual Eta data for 20 snow-free California watersheds was derived using the water balance method for hydrologic years 1982-2011. Independent variables examined in this study were annual rainfall, rainfall intensity, and potential evapotranspiration. These quantities were obtained or calculated from daily PRISM rainfall and temperature datasets. Results indicated that rainfall was the dominant control over variations in mean annual Eta across the study region (Adj. R2 0.935) and was the primary control over interannual variations in Eta for 15 out of 17 study watersheds. Rainfall intensity was a significant but weaker predictor of mean annual Eta (adj. R2 0.833) and was a significant predictor of annual variations in Eta for 12 out of 17 watersheds. A weak relationship between Etp and Eta was observed across the study region (adj. R2 = 0.660) and the relationship was found to be negative. Etp was a significant, though weak, predictor of annual Eta for 8 out of 17 watersheds. The amount of variance in annual Eta explained by rainfall, rainfall intensity, and Etp decreased with increasing watershed wetness (P/Etp). Slope coefficients in the sensitivity analysis indicated that Eta was sensitive to rainfall in 15 out of 17 watersheds, while 12 and 8 watersheds were sensitive to changes in rainfall intensity and Etp, respectively. Watershed Eta was most sensitive to changes in Etp (avg. _ = -5.46), moderately sensitive to rainfall intensity (avg. _ = 1.08), and least sensitive to rainfall (avg. _ = 0.59). Inland, water-limited watersheds exhibited the greatest sensitivity to changes in meteorologic variables affecting Eta, which has implications for potential changes in future climate.