Service-Driven Modeling Approach to Managing Water Allocation in Priority Doctrine Regions

Abstract

This work focuses on developing methods to better manage significant imbalances between water supply and demand during droughts. A service-driven approach (Model as a Service, or MaaS) is used to couple river modeling services with optimization services for determining optimal water allocation strategies under daily drought scenarios. It demonstrates the promise of coupling simulation-optimization model services to improve real-time water management in a service driven framework, which should be beneficial to many other water resource applications. The approach is implemented using the DataWolf workflow tool and AzureML Cloud machine learning services and applied to an April 2015 drought event in the Upper Guadalupe River Basin, Texas. Weather and water demand uncertainty are considered through scenario-based optimization. The optimization objective is to minimize the daily total curtailment hours across all groups of permit holders. The scenario analysis shows that the current permit grouping system has a significant impact on the optimal water allocation strategy. The scenarios also demonstrate that noncompliance of junior water users is predicted to have a much greater effect on the river system than noncompliance of senior water users. The resulting framework can be deployed for water allocation in any area by updating water user information, water allocation policy constraints, and river data that can be obtained from publicly available sources.