Artificial Intelligence in Water Resource Management: The Past, Present and Opportunities thereof

Abstract

Understanding hydrologic systems at the scale of interconnected watersheds and associated river basins is critically important for developing and developed nations alike, when faced with extreme weather events, often affecting water security and quality, and ultimately affecting civilizations as a whole. Current state hydrological modeling has been largely deterministic and simplified to linear or linear like models, often being unable to respond to intra- and inter-basin variations in topography, weather patterns, land cover, soil drainage capacity, and other associated secondary factors. The need of the hour is for a quantum shift in the practice of hydrological modeling with the ability to handle data driven stochastic variables leveraging the availability of real-time data and capable of establishing explicit relationships between input variables and phenomena with little to no physical modeling. The aim of this survey paper is to present a review of current artificial intelligence (AI) applications in the water resource management domain with an emphasis on surface hydrology. For the purpose of this paper, we will focus on what the US Defense Advanced Research Projects Agency (DARPA), one of the major public funders of AI research in the past decades, has recently labeled the second and third wave of AI. (DARPA, 2021) classifies the second and third wave of AI techniques as intelligent systems with the ability to learn and respond to a changing environment.

In the subsequent sections, the paper describes the review methodology adopted to identify relevant literature, outlines the DARPA second and third wave of AI application frameworks, and describes two specific applications of AI frameworks in water resource management, followed by concluding remarks.