In 2010, a two-dimensional hydrodynamic and water-quality model (CE-QUAL-W2) of Lake Ashtabula, North Dakota, was developed by the U.S. Geological Survey in cooperation with the North Dakota State Water Commission to understand the dynamics of chemical constituents in the reservoir and to provide a tool for the management and operation of the Devils Lake State Outlet in meeting the water-quality standards downstream from Baldhill Dam. The Lake Ashtabula model was calibrated for hydrodynamics, sulfate concentrations, and total dissolved-solids concentrations to ambient conditions from June 2006 through June 2010. The calibrated model then was used to simulate four scenarios that represent various Devils Lake outlet options that have been considered for reducing the water levels in Devils Lake. Simulated water temperatures compared well with measured temperatures and differences varied spatially in Lake Ashtabula from June 2006 through June 2010. The absolute mean error ranged from 0.7 degrees Celsius to 1.0 degrees Celsius and the root mean square error ranged from 0.7 degrees Celsius to 1.1 degrees Celsius. Simulated sulfate concentrations compared well with measured concentrations in Lake Ashtabula. In general, simulated sulfate concentrations were slightly overpredicted with mean differences between simulated and measured sulfate concentrations ranging from -2 milligram per liter to 18 milligrams per liter. Differences between simulated and measured sulfate concentrations varied temporally in Lake Ashtabula from June 2006 through June 2010. In 2006, sulfate concentrations were overpredicted in the lower part of the reservoir and underpredicted in the upper part of the reservoir. Simulated total dissolved solids generally were greater than measured total dissolved-solids concentrations in Lake Ashtabula from June 2006 through June 2010. The mean difference between simulated and measured total dissolved-solids concentrations ranged from -3 milligrams per liter to 15 milligrams per liter, the absolute mean error ranged from 58 milligrams per liter to 100 milligrams per liter, and the root mean square error ranged from 73 milligrams per liter to 114 milligrams per liter. Simulated sulfate concentrations from four scenarios were compared to simulated ambient concentrations from June 2006 through June 2009. For scenario 1, the same location, outflow capacity, and sulfate concentration as the current (2010) Devils Lake State Outlet were assumed. The increased flow and sulfate concentration in scenario 1, beginning on May 31 and extending to October 31 each year, resulted in an increase in sulfate concentrations to greater than 450 milligrams per liter in the reservoir at site 7T (approximately the middle of the reservoir), starting July 5 in 2006, July 28 in 2007, and July 15 in 2008. Sulfate concentrations increased to greater than 450 milligrams per liter considerably later at site 1T (near the dam), starting October 8 in 2006, October 29 in 2007, and October 3 in 2008. For scenario 2, the same Devils Lake State Outlet sulfate concentration as scenario 1 was assumed, but the flow through the Devils Lake State Outlet was doubled, which resulted in a more rapid increase in sulfate concentrations in the lower part of the reservoir and slightly greater values at all four sites compared to scenario 1. Sulfate concentrations increased to greater than 450 milligrams per liter 61 days earlier in 2006, 67 days earlier in 2007, and 41 days earlier in 2008 at site 1T. For scenarios 3 and 4, possible increases in flow and concentration from the current outlet location (from the West Bay of Devils Lake) and from a proposed outlet from East Devils Lake were simulated. Conditions for scenario 3 resulted in a relatively rapid increase in sulfate concentrations in the reservoir, and concentrations were greater than 750 milligrams per liter in most years at all four sites. As expected, scenario 4 resulted in greater sulfate concentr