Designing on-farm irrigation ponds for high and stable yield for different climates and risk-coping attitudes

In many regions precipitation does not reliably meet crop water demands - a situation that climate change will likely exacerbate. Supplemental irrigation can help enhance and stabilize crop yields, but the need of water for irrigation has often led to groundwater over-exploitation. On-farm ponds can provide a more sustainable water source. Their use has often been promoted by local authorities, but, by converting a portion of cultivated area to water storage and reducing water availability downstream, on-farm ponds also imply constraints and trade-offs. For an effective exploitation of their potential benefits, they must be carefully designed and managed based on the local edaphic and climate conditions - a non trivial, task because of the cascading effects of rainfall unpredictability. Here we identify the most suitable on-farm pond size, according to two criteria: maximization of average yield (i.e., production maximization) and achievement of a minimum acceptable yield (i.e., risk minimization, accounting for the farmer's risk aversion). To this aim, we develop a minimalist model, requiring few, physically based parameters, coupling crop biomass, soil moisture, and water stored in the pond. While general, the model is here applied to a case-study in the Lower Mississippi River Basin (USA). Simulations show that yield maximization and risk minimization are goals hard to reconcile, regardless of climatic conditions, soil type and irrigation strategy, with smaller ponds allowing the maximum average yield at the cost of reducing its stability from year to year. Stress avoidance irrigation ensures higher yields than deficit irrigation, even if it implies a faster use of the stored water. Future, more extreme climates will result in lower maximum average yields and narrower ranges of pond sizes ensuring desirable minimum yields.