Making the most of data: An information selection and assessment framework to improve water systems operations

Advances in Environmental monitoring systems are making a wide range of data available atincreasingly higher temporal and spatial resolution. This creates an opportunity to enhance real-time under-standing of water systems conditions and to improve prediction of their future evolution, ultimately increas-ing our ability to make better decisions. Yet, many water systems are still operated using very simpleinformation systems, typically based on simple statistical analysis and the operator’s experience. In thiswork, we propose a framework to automatically select the most valuable information to inform water sys-tems operations supported by quantitative metrics to operationally and economically assess the value ofthis information. The Hoa Binh reservoir in Vietnam is used to demonstrate the proposed framework in amultiobjective context, accounting for hydropower production and flood control. First, we quantify theexpected value of perfect information, meaning the potential space for improvement under the assumptionof exact knowledge of the future system conditions. Second, we automatically select the most valuableinformation that could be actually used to improve the Hoa Binh operations. Finally, we assess the eco-nomic value of sample information on the basis of the resulting policy performance. Results show that ourframework successfully select information to enhance the performance of the operating policies withrespect to both the competing objectives, attaining a 40% improvement close to the target trade-offselected as potentially good compromise between hydropower production and flood control.