Keywords

agent-based modeling, Bakken shale, CHANS, decentralized optimization, hydraulic fracturing

Start Date

26-6-2018 2:00 PM

End Date

26-6-2018 3:20 PM

Abstract

A water-depot based water allocation system has emerged in western North Dakota to distribute a large quantity of freshwater for shale oil development activities at the Bakken. This novel, multi-agent system of regional water allocation has never been previously examined. An agent-based model was developed by categorizing more than 600 water depots into 9 agents in terms of types of water permit, ownership and water sources. The model was calibrated against annual water use data recorded by a state agency during 2007 to 2014, with R2 values ranging between 0.432 to 0.998 for different agents. The benefit functions also compared favorably against the estimated water sales for the water depot industry. The calibrated model was then used to evaluate the impacts of water policies and to devise effective water management strategies in the Bakken region. Our analysis shows that the authorization of the Western Area Water Supply Project, implementation of the “In-Lieu-Of Irrigation” program, and an accelerated issuance of temporary surface water permits were the most important water policies adopted by the state of North Dakota to manage the limited regional water resources during the recent oil boom. This agent-based model for water allocation and management analysis in the Bakken region will assist and inform other policymakers and water practitioners to develop pertinent water policies and management apparatus to address increased industrial water demands associated with the unconventional oil and gas development in their regions.

Stream and Session

C2

Share

COinS
 
Jun 26th, 2:00 PM Jun 26th, 3:20 PM

A Multi-Agent System of Water Allocation and Management in the Bakken Region

A water-depot based water allocation system has emerged in western North Dakota to distribute a large quantity of freshwater for shale oil development activities at the Bakken. This novel, multi-agent system of regional water allocation has never been previously examined. An agent-based model was developed by categorizing more than 600 water depots into 9 agents in terms of types of water permit, ownership and water sources. The model was calibrated against annual water use data recorded by a state agency during 2007 to 2014, with R2 values ranging between 0.432 to 0.998 for different agents. The benefit functions also compared favorably against the estimated water sales for the water depot industry. The calibrated model was then used to evaluate the impacts of water policies and to devise effective water management strategies in the Bakken region. Our analysis shows that the authorization of the Western Area Water Supply Project, implementation of the “In-Lieu-Of Irrigation” program, and an accelerated issuance of temporary surface water permits were the most important water policies adopted by the state of North Dakota to manage the limited regional water resources during the recent oil boom. This agent-based model for water allocation and management analysis in the Bakken region will assist and inform other policymakers and water practitioners to develop pertinent water policies and management apparatus to address increased industrial water demands associated with the unconventional oil and gas development in their regions.