Assessment of the validity of using mean potential evaporation in computations of the long-term soil water balance

Multi-decadal modelling of the soil water balance is often limited by the availability of data needed to calculate potential evaporation (PE). A solution sometimes adopted is to use available precipitation data but substitute a climatological representation of PE, the rationale being that PE is an inherently conservative variable. Convincing support for such an approach can be found in the literature but there are grounds for suspecting significant bias under extreme conditions. The validity of using climatological PE in long-term water balance studies is investigated by comparing soil water deficit time series simulated using a daily water balance model. This involves comparing a baseline run of the model using actual PE (based on observations) with alternative simulations using various climatology-based estimates of PE. The modelling experiments are over 13 years for one site (Auckland, New Zealand) with emphasis on comparing model performance during wet and dry years. Results show that substituting climatological PE estimates into the daily water balance produces a soil water regime very similar to that derived using actual PE values, including over relatively extreme periods. Best results were obtained where a PE reduction was applied to account for PE suppression on raindays, but the improvement in performance was minor compared to cruder methods, including PE estimates based on monthly data.