Impacts of empirical and physical evaporation methods on changes in hydrological components and drought indices under climate change scenarios

The choice of potential evapotranspiration (PET) methods in hydrological modelling can result in contrasting estimates of meteorological drought and low flows in the context of climate change, but their effects on other drought types and the associated hydrological components are still unknown. This study aims to systematically assess the effects of a temperature-based (T-based) method and a Penman-Monteith (PM) method on 1) historical hydrological simulations, and 2) projected changes in hydrological components, meteorological, agricultural and hydrological droughts under climate scenarios in mainland Norway. The distributed version of the hydrological model HBV with two PET methods was driven by six regional climate projections under three emission scenarios. The results show that the PET methods provide similar historical discharge simulations but different spatial distribution of hydrological components. The T-based method always estimates higher PET and evapotranspiration, lower soil moisture and runoff and longer severe droughts than the PM method under climate scenarios, especially in Eastern, Central and Northern Norway. The discrepancies of projected changes between the two PET methods generally increase linearly with temperature change. Among various drought types, agricultural drought projections are the most sensitive to the choice of PET method.