Irrigation Scheduling using Probabilistic Simulation Optimization Framework and Farm-scale Ecohydrological Model at Daily Scale Using Weather Forecasts

Abstract:

Irrigated agriculture being the world’s largest sector of water consumption, optimization of daily scale irrigation scheduling is an urgent need in a warming world. However, a competent irrigation scheduling system should not negatively impact crop yield and profitability. With this motivation, we have developed an irrigation water use optimization model, which incorporates the short to medium range weather forecasts to deliver the probability distribution of the rainfall and consequently the soil moisture, to match farmers’ requirements of irrigation. The objective function being the minimization of irrigation water use, the model ensures avoidance of crop water stress beyond a prescribed threshold. A daily timescale has been chosen for the ecohydrological model to compute evapotranspiration, runoff and leakage based on the stochastic rainfall component derived from forecasts, and the ecohydrological model itself is a component of the optimization algorithm, solved by Monte Carlo simulations. The model has been tested in two representative grape farms in Nashik, in the state of Maharashtra, India after being calibrated with the past data procured from the farmers. Based on the analysis for irrigation once in 1 day, 3 days and 7 days, respectively, we find that the water use can be minimized by 10 - 45% depending on the reliability factor and interval of irrigation considered in the model. Moreover, water use minimization achieved at daily timescale irrigation scheduling is higher than in once in 3 days or 7 days irrigation. This tool can be extremely valuable for decision support system in optimum irrigation prescription and practice all over the world.