Modeling daily reference ET in the karst area of northwest Guangxi (China) using gene expression programming (GEP) and artificial neural network (ANN)

Nonlinear complexity is a characteristic of hydrologic processes. Using fewer model parameters is recommended to reduce error. This study investigates, and compares, the ability of gene expression programming (GEP) and artificial neural network (ANN) techniques in modeling ET0 by using fewer meteorological parameters in the karst area of northwest Guangxi province, China. Over a 5-year period (2008-2012), meteorological data consisting of maximum and minimum air temperature, relative humidity, wind speed, and sunshine duration were collected from four weather stations: BaiSe, DuAn, HeChi, and RongAn. The ET0 calculated by the FAO-56 PM equation was used as a reference to evaluate results for GEP, ANN, and Hargreaves models. The coefficient of determination (R (2)) and the root mean square error (RMSE) were used as statistical indicators. Evaluations revealed that GEP, and ANN, can be used to successfully model ET0. In most cases, when using the same input variables, ANN models were superior to GEP. We then established ET0 equations with fewer parameters under various conditions. GEP can produce simple explicit mathematical formulations which are easier to use than the ANN models.