Hydraulic Modeling of The Mississippi and Atchafalaya Rivers To Estimate Flood Risk

Stage-frequency relationships from a calibrated and validated hydraulic model were used to develop site-specific estimates of loadings and overtopping probabilities for locations of interest. The Corps of Engineers New Orleans District (MVN) used a two-dimensional HEC-RAS hydraulics model to develop water levels for flood flows of varying magnitudes, an updated flood flow frequency analysis to inform the annual exceedance probabilities of those flood flows, and a MATLAB-based code to extract water surface elevation results at hundreds of locations where stage-frequency data was needed. This effort simplified the estimation of flood risk at numerous locations for use in ranking and categorizing the highest risk segments in the systems. This procedure represents a significant improvement over processes utilized previously – which may have used a 1-dimensional hydraulic model with inherent simplifications and estimates made to model a system with 2-dimensional flow patterns, or interpolations between locations where stage-flow rating curves could be developed.

Flow frequency analysis

A flood flow frequency analysis was completed using combined discharges for the Atchafalaya and Mississippi Rivers just downstream of the Old River Control Complex for the period 1962-2021 (60 years). Bulletin 17C procedures were used in HEC-SSP to develop the flow frequency curve.

RAS calibration, validation, event modeling An HEC-RAS geometry with a single 2D Area covering nearly all of the New Orleans District limits was used for the hydraulic modeling. The USGS Northern Gulf of Mexico (NGOM) “topobathy” dataset was the base terrain layer, upon which refinements and modifications were made. Levee, floodwall, and structure elevations for Mississippi River and Tributaries (MR&T) systems were burned into the HEC-RAS terrain and those elevations were enforced in the 2D Area mesh using break lines. While the upper half of the Atchafalaya Basin is contained by the Atchafalaya River levees, the lower half is marked by wide floodplains, bifurcations, and small distributaries and tributaries. Where detailed bathymetric data was not available for the smaller but significant channels, cross section dimensions were estimated and burned into the terrain using Terrain Modification features in RAS Mapper.

The model was calibrated to the 2019 flood, primarily by varying the Manning’s n values for the land use types in the model domain. In some instances, the estimated dimensions of smaller channels that had been burned into the terrain were refined to improve the model’s fit to observed data. The model was validated to the 2011 flood.

Once calibrated and validated, flood flows for varying frequencies were simulated in the HEC-RAS model.

Data processing

A MATLAB script was developed to read the HEC-RAS hdf5 output file for any location desired. The maximum WSE for each point was reported in a spreadsheet that could be used for further analysis.

Limitations

The Project Design Flood (3,030,000 million cfs at the latitude of Red River Landing) is larger than the flood of record post-ORCC construction (2,304,000 in 2011) by 726,000,000 cfs. Model calibration is limited to the extent of observed data.