Performance assessment of the FAO AquaCrop model for irrigated cabbage on farmer plots in a semi-arid environment
Wellens, Joost × Raes, Dirk Traore, Farid Denis, Antoine Djaby, Bakary Tychon, Bernard #
Elsevier Scientific Pub. Co.
Agricultural Water Management vol:127 pages:40-47
tSmall-scale irrigation initiatives are expanding rapidly in Burkina Faso. However, in many cases optimalyields are not being obtained despite the available water and the required nutrient applications. Localstakeholders need an easy-to-use decision-support tool to assess irrigation water use and its impact onyield. In this study, a water-driven crop model, AquaCrop, developed by FAO, was adapted for cabbage(Brassica oleracea L.) using a limited dataset and leave-one-out cross-validation (LOOCV). The experimentwas conducted in south-western Burkina Faso on small irrigated farmer plots, where optimal managerialconditions could not always be guaranteed. Statistical indicators – normalized root mean square error(nRMSE) and index of agreement (d) – suggested that the model is very reliable for simulating cabbagebiomass yield and soil water content (low nRMSE and d-index near 1). The relationship between observedand simulated yield produced a d-index of 0.99 and an nRMSE of 1.39% (or 0.59 ton/ha). The comparisonbetween observed and modeled soil water content gave a d-index of 0.90 and an nRMSE of 4.38% (or9.13 mm). Also of interest was the indirect link between plant density and yield via maximum canopycover, which can considerably simplify yield estimation. It was concluded that AquaCrop was a veryuseful tool for enabling local end-users to evaluate and optimize cabbage yield and irrigation water use.