Sensitivity analysis of 18 different potential evapotranspiration models to observed climatic change at German climate stations

Potential evapotranspiration models very often are important part of hydrological catchment models to calculate potential evapotranspiration (PET) which then is used to estimate actual evapotranspiration considering the soil moisture status. As many different approaches exist, the question arises in which way the choice of the PET model affects the impact of climate change on the calculated water balance? Therefore, 18 different PET models were compared with respect to their sensitivity to observed climate change. Long-term climate data of six German climate stations were used to identify changes in the climate data itself and changes in the calculated PET. The results show that all investigated PET models are sensitive to significant trends in climate data. However, it is also shown that all models show different sensitivities, and that the sensitivities cannot be grouped in terms of different types of PET models such as the aerodynamic concept, radiation or temperature based approaches and combination equations. Predominantly, the variability within a group of models of the same type is comparable to the variability between different model types. Therefore it can be concluded that PET models should be validated in a regional context before they are applied to a certain region within a climate change study despite the poor availability of long-term PET measurements.