Neumann, D., G. Oelsner, J. Prairie, S. Anderholm, and E. Zagona (2013). “Salinity Modeling in the Colorado River and Upper Rio Grande River Basins Using RiverWare,” AWRA Spring Specialty conference on Agricultural Hydrology and Water Quality II, St Louis, Missouri, Mar 25–27, 2013.


Water managers use decision support tools to simulate flow and storage of water in Western U.S. river basins. Some of these basins have high salinity content from both natural and anthropogenic sources. Increasingly, water managers require tools that also simulate salt concentration and/or mass. Although there are many excellent water-quality models, it is not common to incorporate water quality at a scale needed to support planning or decision making. To meet this need, salinity modeling was incorporated into RiverWare, a general river and reservoir operations decision making and planning tool. Currently, salinity modeling in RiverWare has been applied to two basins, the Colorado River and the Upper Rio Grande.

In the Colorado River planning model, flow and volume are simulated at a monthly timestep. A statistical framework is applied to generate salinity concentration consistent with flow. A nonparametric space-time disaggregation model is coupled with a nonparametric regression representing the relationship of natural salt to flow. The framework first simulates annual salinity consistent with annual flow and then transfers the annual data to a monthly timestep for use in the planning model. The movement of salt is tracked using a well-mixed approach. Salt additions are modeled by methods that represent dissolution and application of salt due to irrigation. Water-quality improvement projects remove salt from the system.

In the Upper Rio Grande, several agencies share a suite of models used for planning, forecasting, operations and water accounting called the Upper Rio Grande Water Operations Model (URGWOM). Salinity modeling is currently being tested on the daily timestep planning model. In this model, salinity is simulated in both shallow groundwater storage and surface reservoirs using a two layer approach. In the groundwater objects, a user specified, constant-thickness upper layer represents the portion of the aquifer that actively exchanges water with the surface. Additional processes that affect salinity can also be modeled including inflows from surface water (including point sources), pumping, deep aquifer brine inputs, evaporation, and evapotranspiration. Using previous timestep information, an explicit solution computes the flow of water and salt between adjacent groundwater objects and surface water.