Deep learning for compute-efficient modeling of BMP impacts on stream- aquifer exchange and water law compliance in an irrigated river basin. (December 2019)

• Record Type: Journal Article
• Authors: Rohmat, Faizal I.W.; Labadie, John W.; Gates, Timothy K.
• Is Part Of: Environmental modelling & software. Volume 122(2019)
• Publication Date: 2019-12
• Subjects: Irrigation -- Stream-aquifer systems -- River basin management -- Groundwater modeling -- Machine learning -- Artificial neural networks -- Geographic information systems -- Water law compliance; Environmental monitoring -- Computer programs -- Periodicals; Ecology -- Computer simulation -- Periodicals; Digital computer simulation -- Periodicals; Computer software -- Periodicals; Environmental Monitoring -- Periodicals; Computer Simulation -- Periodicals; Environnement -- Surveillance -- Logiciels -- Périodiques; Écologie -- Simulation, Méthodes de -- Périodiques; Simulation par ordinateur -- Périodiques; Logiciels -- Périodiques; Computer software; Digital computer simulation; Ecology -- Computer simulation; Environmental monitoring -- Computer programs; Periodicals; Electronic journals; 363.70015118
• DOI: 10.1016/j.envsoft.2019.104529
• Languages: English
• ISSNs: 1364-8152

Abstract

Abstract: Inefficient irrigation practices in the alluvial Lower Arkansas River Basin (LARB) of Colorado are contributing to salinization, waterlogging, reduced crop yields, and harmful concentrations of pollutants in the stream-aquifer system. Intensive data collection and modeling efforts in the LARB over the past 20 years have resulted in development of the GIS-based basin-scale decision support system River GeoDSS. Parallel efforts in regional-scale calibration and application of the MODFLOW-SFR2-RT3D-OTIS stream-aquifer system model permit evaluation of best management practices (BMPs) designed to mollify adverse environmental impacts. Since BMP implementation is allowable only if water laws are not violated, a deep learning model is developed to serve as an accurate, compute-efficient surrogate of MODFLOW-SFR2 and is imbedded in River GeoDSS for assessing basin-scale impacts of BMP implementations on stream-aquifer exchange and water rights. It is shown that BMPs can be implemented while maintaining reasonable water law compliance with development of a new reservoir storage account. Highlights: Inefficient irrigation practices risk Lower Arkansas River Basin sustainability. Data and modeling have resulted in a river basin decision support system. Deep neural network model is developed as a surrogate stream-aquifer model. River basin-scale impacts of best management practices (BMP) are assessed. The BMP assessment ensures protection of water rights and policies in the basin.