Updates to a Groundwater Flow Model to Facilitate Decision Support and Uncertainty Analysis for the Mississippi Alluvial Plain

Author(s): Peterson, S.; Fienen, M.; Clark, B.; White, J.

A recently-updated groundwater model of the Mississippi Embayment has previously provided key information and decision support for stakeholders and serves as the foundation for the Mississippi Alluvial Plain project for continued decision support tool development. The MODFLOW-2005 groundwater model was updated to use the Newton-Raphson solver, which provided an improved ability to simulate the nonlinear unconfined aquifer. The streamflow routing network was updated with a more detailed and comprehensive representation of surface water features. Recharge was updated using outputs from a Soil-Water-Balance model generated through a companion study. The MODFLOW-2005 model also was converted to MODFLOW-6 to take advantage of the most-current (2018) software improvements available and to facilitate future refinement. Existing calibration targets were retained with the exception of stream base flow, which also was updated through a companion study. Forecast uncertainty for stakeholder-identified scenarios is being evaluated through history matching and uncertainty analysis using PEST and a newly developed iterative ensemble smoother.

In addition, the value of potential future data is assessed through using the updated model to quantify their role in reducing uncertainty of stakeholder-identified forecasts. This type of analysis, known as "data worth" analysis, is being used to prioritize geophysical mapping and hydrogeologic monitoring efforts so that the data collected is most valuable to model-simulated future conditions. Such analyses can be repeated as necessary as new societal concerns arise and as new understanding of the MAP system is gained.

Go back


Past Conference Archive