Implications of potential evapotranspiration methods for streamflow estimations under changing climatic conditions

Abstract

ABSTRACTSeveral potential evapotranspiration (PET) estimation methods are commonly used to quantify water and energy budgets. As each PET method can differ, their effects on projected streamflows under changing climatic conditions are critical to quantify the sensitivities of these methods. We used the Coupled Model Intercomparison Project (CMIP5) climate dataset and Hydrological Simulation Program – FORTRAN (HSPF) model to compare the following five PET methods [Hargreaves (HG), Hamon (HM), Thornthwaite (TW), Priestley–Taylor (PT) and Penman–Monteith (PM)] for the Susquehanna River Basin in the northeastern United States. As PET is used as an input, various configurations of HSPF driven by these five PET estimates were used to calibrate HSPF with observed streamflow data from 41 gaging stations. We also used nine global climate model inputs to derive five PET estimates, which were subsequently used as inputs to compute streamflow projections for 2020–2099. An increase in precipitation from 6.2 to 7.2% and an increase in temperature from 1.8 to 2.7 °C were projected, while changes in PET and actual evapotranspiration (AET) were found to substantially differ among the PET methods. The HM method shows an increase in AET of between 14 and 24%, while the other methods show an increase of between 7 and 12%. It is concluded that streamflow projections are sensitive to the selection of the PET methods in the HSPF model; a decrease of up to 5.5% and increase of up to 3.6% are projected for PET levels estimated by using the HM and HG methods, respectively; and both HM and TW are found to be suitable for simple seasonal water balance analyses conducted at regional scales.