Application of AquaCrop Model Yield Optimization for Prediction of Optimal Sowing Dates under Alternative Soil Moisture Regimes in Laikipia, Kenya.

Abstract

This study presents the use of AquaCrop model in predicting optimal onset dates for wheat crop grown under ASAL conditions in Laikipia County, Kenya. Modeling of onset dates was done through simulation of root zone moisture content to reduce yield losses from crop failure. The optimal sowing date(s) for rain-fed wheat was based on AquaCrop rainfall criterion for generation of onset dates and optimization analysis of AquaCrop simulated grain yields. Cumulative rainfall of 10 mm in 4 successive days’ onset criteria equivalent to the calculated Readily Available Water (RAW) required at 10 cm soil depth was adopted. Based on the onset criteria, a total of 57 simulations (19 years rainfall record and three onsets: early, normal and wet) was run using AquaCrop model and frequency analysis applied to the simulated yields to get the yield levels expected at varying levels of probability of exceedance. The set threshold was incremental at 0.5 ton/ha level from 0-12 ton/ha. The probability of exceedance was zero for threshold mean yields beyond 5.5, 11.0, 11.5 ton/ha for early (SD1), normal (SD2) and late (SD3) onset respectively. At 20%, 50% and 80% probability of exceedance, the average expected yield was less or equal to 4.5, 0.5 and 0.38 ton/ha for SD1, 4.5, 0.75 and 0.38 ton/ha for SD2 and 6.0,0.75, 0.38 ton/ha for SD3 respectively. Results indicated that zero tillage optimized the yield in all the sowing dates selected. In conventional tillage, early onset had high yield advantage for the viable early onset, but the risk of failure was high representing 22.11%. It is recommended that sowing date, which is a technology problem, be given a lot of attention