Mississippi RiverWeb^SM Museum Consortium

Numerical Modeling

One of the primary goals of the Consortium was to leverage advanced computing and visualization technologies "to bring alive the complex interplay of natural and human forces shaping the Mississippi River Basin." (NSF proposal, May1998) Also important to the Consortium was that all visual representations and activities in the DRB should emulate current scientific methodologies and tools as closely as possible and therefore be grounded not on "canned’ but rather on genuine scientific data. We relied on three sources for such data.

• Acquired data (e.g., historic flow, bathymetry, elevation, land cover).
• Pre-constructed geometric models of selected landscape and cultural features (eg., trees, grass cover, bridges, buildings).
• Pre-computed outputs of numerical simulations based on mathematical models of dynamic phenomena, such as channel flow, runoff, erosion, sediment transport, and the effects of levees and dams on river basin hydrology. These simulations provide the data for many of the activities currently available in the DRB.

For example, a two dimensional finite-element grid simulation was used to calculate channel flow velocity, and water surface elevation.

These outputs were fed into other computer models to simulate the location, velocity, and direction of a tow barge in the River Pilot Simulator, to compute sediment transport in conjunction with archival sediment load data, and to compute the numerical data for DRB activities such as Make It Flood, Remove a Dam, and Change a Levee.

Flow velocity and extent ->

Levee in place

Remove Levee

Looking beyond the river channel and onto the adjoining land, yet different models were used to simulate the amount, direction, and velocity of overland flow (rainfall runoff), and the extent of soil erosion or deposition at each location on a grid. These second sets of simulation data underpin the following DRB activities: Make It Rain, Track Sediment, and Examine Erosion.

Hardware constraints dictated that all simulation data be pre-computed at the GMSlab then archived in the DRB database. Nevertheless, all model outputs were visualized in real time in response to visitor requests initiated through the above DRB-supported activities. Future development of the DRB might accommodate real time numerical simulations that afford visitors greater freedom to explore their own "hypotheses" by increasing the number of "what if" scenarios that can be investigated.