VALIDATION OF AQUACROP MODEL FOR SIMULATION OF RAINFED BULB ONION (ALLIUM CEPA L.) YIELDS IN WEST UGENYA SUB-COUNTY, KENYA

Abstract

Background. The precision of crop growth simulation models is a paramount facet in their use for evaluating on-field management practices to improve crop yields. Objective. To validate the accuracy of AquaCrop model in simulating onion yields and canopy cover in the sub humid environment of West Ugenya Sub County, Kenya. Methodology. Aqua Crop model version 5.0 was evaluated for experimental yields of bulb onion (Allium cepa L.) grown in West Ugenya Sub County, Kenya for two seasons (March to May long rains season, and October to December short rains season) on soil integrated with organic and inorganic fertilizers i.e. 5 Mega grams ha-1 cattle manure combined with inorganic fertilizers containing 28 kg P Ha-1 and 30 kg N ha-1. The model was calibrated based on its conservative parameters for C3 crops under the Growing Degree Days (GDD) mode. Results. Statistical comparison of the model’s simulated yields versus experimental yields gave RMSE (Root Mean Square Error) values of 0.22 and 0.61 in season I and season II, respectively, which are generally closer to zero, indicating average to high model precision. Modified Willmott index of agreement (d mod) was 0.44 (season I) and 0.69 (season II), while for Nash and Sutcliffe coefficient (E), 0.85 (season I) and 0.14 (season II). The constant for d mod and E indicates high model accuracy if value is close to one. The values from the simulations were detached, generally indicating in both cases average to high model performance. The canopy cover development from germination to the crop’s 150 days to physiological maturity gave a Pearson correlation coefficient (r) that averaged 0.9. The r-values were close to one, indicating a positive linear relationship between simulated and experimental canopy cover. Conclusion. Overall, the model provided acceptable simulation of onion crop yield and canopy cover.