Semiarid crop production from a hydrological perspective: Gap between potential and actual yields

Rapid population growth in the dry climate regions, arable land scarcity, a nd irrigation expansion limitations direct our interest to possibilities of yield increase in rainfed agriculture. Literature, however, indicates larg e differences between actual and potential yields, and between yields on fa rmers' fields and research stations. This article focuses on the determinan ts of these yield gaps and the windows of opportunity for yield increase on the farmer's field together with the agricultural challenges involved. The study links the conventional approach to estimate crop water requirements and dry spell effects on biomass production to a conceptual Green Water Cro p Model. This model addresses the effects on crop yields of the sequential diversions of infiltrating rainfall (rainwater partitioning into runoff, pl ant available soil water, and deep percolation) and of different relations between nonproductive evaporation flow and productive transpiration flow, d efined together as green water flow. Also, the effects of droughts and dry spells are analyzed. The model is used to demonstrate typical situations fo r semiarid and dry subhumid conditions (lengths of growing period (LGP) of 90 and 179 days, respectively) for maize (Zea mays (L.)) under on-station a gricultural conditions. Based on detailed water flow analysis in a 3-year o n-farm case study in the Sahel on pearl millet (Pennisetum glaucum (L.) Br. ), the model is used to clarify the large scope for improved yield levels, achievable through land and water management securing that runoff losses an d deep percolation are reduced and nonproductive evaporation losses minimiz ed. The analysis indicates that poor rainwater partitioning and low plant w ater uptake capacity alone reduces estimated on-farm grain yields to 1/10th of the potential yields. This suggests that lack of water per se not neces sarily is the primary constraint to crop growth even in drought prone areas of sub-Saharan Africa. The conclusion is that even a doubling of crop yiel ds would he agro-hydrologically possible with relatively small manipulation s of rainwater partitioning in the water balance.