An evaluation of EPIC soil water and yield components in the gully region of Loess Plateau, China

The Erosion and Productivity Impact Calculator (EPIC) has been used to determine the effect of different cropping systems and management practices on soil productivity in the Loess Plateau of China. However, its crop growth and soil water balance submodels have not been verified in this region. The objective of the present study was to evaluate the ability of EPIC to estimate soil water content (theta in m(3)/m(3)), seasonal evapotranspiration (ET in mm/season) and crop yield (Y in t/ha) for winter wheat and maize. A 20-year field experiment was conducted at the Changwu Agro-ecological Experimental Station of the Loess Plateau, and divided into a calibration period and a validation period. Data from calibration (1984-94) were used to optimize the four most sensitive parameters of the EPIC crop yield submodel, whereas data from 1994 to 2004 were used for validation. For both crops, there were no significant differences between measured and estimated long-term means of the three variables (P=0(.)05) for either the calibration or validation periods. EPIC estimated all three variables with a small relative root mean square error (RRMSE), i.e. the ratio of root mean square error to the mean value. For wheat and maize, the calibration period resulted in respective RRMSE values of 0(.)112 and 0(.)100 for theta, 0(.)121 and 0(.)116 for ET, and 0(.)135 and 0(.)147 for Y. During the validation period, the RRMSE values obtained were 0(.)090 and 0(.)085 for theta, 0(.)129 and 0(.)135 for ET, and 0(.)169 and 0(.)149 for Y, for wheat and maize, respectively. The performance of EPIC in estimating annual values of theta, ET and Y was variable. For validation, EPIC explained 65, 79 and 64% of the measured variations of theta, ET and Y, respectively, for wheat, and 60, 70 and 67% for maize. The EPIC-estimated long-term average values of the three variables were not significantly different from measured values for winter wheat and maize during the calibration and validation periods. It can therefore be used in the gully region of the Loess Plateau to define alternative cropping systems and management practices.