Assessing the trends and uncertainty of maize net irrigation water requirement estimated from climate change projections for Zimbabwe

Imperfect knowledge arising from unpredictable climate leads inexorably to uncertainty in decision making in agriculture. Uncertainty estimates are needed in descriptive and prescriptive analysis and in validation studies at both farm level and policy level agricultural decision making. This study assessed the trends and uncertainty of the global climate model (GCM) derived net irrigation water requirement (NIR) predictions estimated by the Food and Agriculture CROPWAT model for Natural Agro-ecological Region II in Zimbabwe. The trends and uncertainty were assessed with the aid of descriptive statistics and exploratory data analysis on the simulated climate parameters, crop water requirement and NIR. The temperature and reference evapotranspiration were projected to increase while rainfall was predicted to decrease in the future periods. NIR was projected to increase by an average (and range) of 33% (−22–92%), 66% (15–168%) and 99% (17–205%) in the 2020s, 2050s and 2090s time slices respectively from a baseline of 67 mm. The NIR relative standard deviation and 95% confidence interval averaged 22% and 10% respectively in the 2020s time slice and were projected to increase in the future. Further quantitative expert elicitation that identifies the divergence of opinion within the subject of uncertainty and advocates for policies on a combination of climate mitigation and adaptation measures is recommended.