Search for: All records

Innovative groundwater management strategies are needed to preserve aquifers for crop irrigation. For sustainability to be lasting, any strategy must balance environmental goals with the economic aims of farmers. These tradeoffs are difficult to manage due to the inherent uncertainty in farming. To address these challenges, we develop a transferable two‐stage stochastic modeling framework to support optimal multi‐year crop and irrigation planning under groundwater pumping restrictions and uncertain precipitation. This modular framework is broadly applicable to regions facing groundwater overuse, helping to balance aquifer sustainability and farmer profitability under uncertainty. We illustrate the model using a case study from western Kansas, USA, where irrigators self‐imposed 5‐year groundwater pumping limits to extend the aquifer's lifespan. While these multi‐year allocation periods offer flexibility, they introduce a temporal dimension to decision‐making beyond typical annual planning. Optimal cropping and irrigation strategies from the stochastic model significantly outperform observed farmer behavior during the first two 5‐year allocation periods (2013–2022), and outperform a deterministic model assuming long‐term average precipitation during dry conditions. We show that optimal crop choices shift from corn to sorghum under more stringent pumping restrictions. Under these constraints, irrigators benefit by conserving water in earlier years and using more in later years, whereas the reverse holds under more lenient restrictions. Extending the allocation window further enhances profitability, though marginal gains diminish beyond 7 years. This modeling framework offers insights for agricultural regions seeking to improve long‐term groundwater management through strategies that support both economic resilience and hydrologic sustainability. 

Machine learning provides valuable information for data-driven decision-making. However, real-world problems commonly include uncertainties and the features needed to generate the prediction outputs are random variables. Even the most reliable machine learning models may not be helpful for decision-makers when the decisions must be taken before the values of features used in machine learning models are realized. To support decision-making under uncertainty, we propose a scenario generation procedure for stochastic programs that incorporates the uncertainties in both prediction features and the machine learning model prediction error. A statistical test is implemented to assess the reliability of the scenario sets by comparison with corresponding historical observations. We test the whole procedure in a case study for crop yield in Midwest.