MODELING EFFECTIVENESS OF LEVEE SET-BACKS USING COMBINED 1D HYDRAULIC MODELING AND FLOOD-LOSS SIMULATIONS

Levees have traditionally been the central element of U.S. flood protection, although levee impacts and their effectiveness have been debated periodically. One potential compromise between local flood protection and safe conveyance of flood flows involves set-backs of current levee systems. To test the effectiveness of set-backs, hydraulic modeling and flood-loss modeling software was used to assess three levee scenarios on the Middle Mississippi River (MMR): 1) current levee configuration (protection for ≤ 50-year flood), 2) no levees, and 3) a 1,524 meter (5000 ft) set-back of current levees.

The study area is a 70 km reach of the MMR that includes the floodplain through Union and Jackson Counties, IL, an area for which we have detailed building inventory data and assessed values. In order to quantify the potential flood damages for each of the scenarios, we used HEC-RAS 1D hydraulic modeling software and FEMA’s HAZUS-MH (Hazards U.S. Multi-Hazard) flood-loss software package.

Comparing flood losses for the three scenarios revealed that the 1,524 m levee set-back would decrease flood losses by 42% (~$25 million), and complete removal of the levee system would decrease flood losses by 24% (~14 million). It is counterintuitive that removal of levees or reduction in levee extent would reduce flood losses; however, higher flood waters caused by reduced conveyance in the current levee configuration cause more damage, at least for floods that overtop the levees. Comparing losses for the levee set-back scenario and the no-levee scenario for the 50-year flood, the 1,524 m levee set-back resulted in 20% (~8 million) less damage than with no levees present. These results suggest that levees increase damages for 50 and 100-year floods on the MMR that overtop the levees in the study area. Flood-damage totals can be reduced by levee set-backs, which reduce flood levels. Flood losses from smaller, more frequent floods (e.g. 10-year event, etc.) that would not overtop the levees, however, need to be modeled in order to create a cumulative probability-based accounting of the positive and negative effects of levees.