An experimental detrending approach to attributing change of pan evaporation in comparison with the traditional partial differential method

Abstract

In predicting how droughts and hydrological cycles might change in a changing climate, change of pan evaporation (Epan) is one crucial element to be understood. The derived partial differential (PD) form of the PenPan equation is a prevailing attribution approach worldwide. However, small biases exist and the application of PD method is limited within the derivation of partial differential form of the equation, which impede the attribution analysis in hydrology. Here we designed a series of numerical experiments by detrending each climatic variable, i.e., an experimental detrending (ED) approach, to attribute changes of Epan over China. We compared the attribution results using these two methods and further analyzed the plausible advantages of ED approach. The comparison shows that both ED approach and PD method perform well in attributing changes of Epan and to the input meteorological variables in China over 1960–2017. The first advantage of ED approach is that it can help make robust adjustment for the PD method in attribution analysis. Another advantage lies in its ability to attribute to the observed meteorological variables in China, when the PD method fails to quantify the contribution of relative humidity and air temperature in net radiation in Epan attribution analysis. We highlight that the ED approach is recommended in attribution analysis for hydrologic research. Together with the adjusted PD method, both methods can assist a better understanding and prediction of water-energy cycles change in a changing climate.