THE DEVELOPMENT OF BEST PRACTICES FOR THE USE OF HAZUS-MH TO ESTIMATE EARTHQUAKE LOSSES IN SOUTHERN ILLINOIS

Our research group is currently performing earthquake loss estimates using FEMA's HAZUS-MH (Hazards U.S. Multi-Hazard) software for 17 southern Illinois counties as part of the Southern Illinois Multi-hazard Mitigation Planning Initiative. To date, we have completed earthquake loss estimates for 9 of the 17 counties and found building-related losses range from $1.3 to $445 million dollars.

Here we assess the sensitivity of HAZUS earthquake losses estimates to the selection of model parameters, data resolution, and data quality. Specifically, we have undertaken sensitivity analyses for ground shaking attenuation relationships, hypocenter depth, soils map resolution, and building inventory accuracy. Models run with each of the four ground shaking attenuation functions available in HAZUS for the central U.S. revealed differences in building loss estimates of up to ± 4.0% (±$1.75 million). Varying the hypocenter depth from 4 km to 10 km to 15 km revealed little or no change in the building loss estimates. The comparison of HAZUS building loss estimates using a generalized soils map from FEMA and a detailed soils site class map for southern Illinois revealed large difference in building-loss estimates of 46 to 140% ($4.6 to $24.0 million). Loss estimates using the detailed soils map were less conservative. Comparison of HAZUS-MH building stock with a county-level building inventory for Jackson County, IL revealed a difference in building inventory of 3,485 structures. These extra structures were likely the major contributor to the observed difference of 220 to 240% ($56 to $75 million) in building-loss estimates for the model run using the HAZUS-MH building stock vs. the Jackson County building inventory.

These results suggest that site-specific soil and building inventory data sets produce less conservative and likely more realistic loss estimates. The two parameters tested, attenuation relationship and hypocenter depth, appear to have at most a small impact on building loss estimates. HAZUS-MH is the official FEMA tool for applying geologic, seismic, engineering, and infrastructure data in order to quantify social impacts of earthquakes, but testing the sensitivity of this tool to model assumptions, parameter selections, and data resolution is a vital step in producing reasonable and trustworthy results.