Estimation of long-term basin scale evapotranspiration from streamflow time series

We estimated long-term annual evapotranspiration (ETQ) at the watershed scale by combining continuous daily streamflow (Q) records, a simplified watershed water balance, and a nonlinear reservoir model. Our analysis used Q measured from 11 watersheds (area ranged from 12 to 1386 km(2)) from the uppermost section of the Neuse River Basin in North Carolina, USA. In this area, forests and agriculture dominate the land cover and the spatial variation in climatic drivers is small. About 30% of the interannual variation in the basin-averaged ETQ was explained by the variation in precipitation (P), while ETQ showed a minor inverse correlation with pan evaporation. The sum of annual Q and ETQ was consistent with the independently measured P. Our analysis shows that records of Q can provide approximate, continuous estimates of long-term ET and, thereby, bounds for modeling regional fluxes of water and of other closely coupled elements, such as carbon.