Advances in Development of Dasymetric Exposure Area for FEMA's Hazus Risk Assessment Tool

FEMA's Hazus Program supports natural hazard loss modeling software (Hazus) and provides national base datasets for users. These national datasets ensure that all users have equitable access to high quality modeling capabilities. Hazus methodology uses a dasymetric Census block-based approach for defining flood hazard building exposure areas. The 2020 decennial Census, TIGER Census block boundaries updates allowed for updating Hazus dasymetric Census block geometries to reflect exposures more accurately. Previous dasymetric area definitions were generated by vectorizing cells with specific classification codes from the National Land Cover Dataset raster product to conduct spatial intersections with Census blocks. The development of Hazus 6.0 dasymetric geometries leveraged new datasets and improvements in spatial and database technologies to generate dasymetric output using a building-focused approach and business logic workflows employed on a Census block level.

This presentation will provide details on the improved data development methods used to produce dasymetric geometries and how an Agile approach allowed for rapid testing and refinement of the processing model. The Hazus 6.0 release has less total developed area in the dasymetric Census block boundaries dataset than previous Hazus versions, yet more accurately reflects areas of exposure to potential losses. The development process combined the use of terabytes of data leveraging building footprint datasets from Bing, Oak Ridge National Laboratory, and FEMA's Natural Hazards Risk Assessment Program; building information from the National Structures Inventory; National Land Cover Dataset; and water areas from TIGER areal hydrography datasets. A detailed logic tree analysis was used to produce the dasymetric output. Raster and vector data were utilized in their native format and business logic evaluated at the Census block level enabled the incorporation of site-specific logic variations and allowed for the simultaneous evaluation of a combination of factors. The dasymetric processing methods included serverless database technologies, open-source GIS, and custom-built raster-vector intersection tools to generate accurate output rapidly and cost efficiently.