Fill and Floods: Analysis of the Impact of Parcel Fill on Residential Flood Damages

Abstract

As property losses from flooding continue to rise in coastal communities, the need to examine the changing dynamics of these damages in relation to specialized mitigation methods at multiple scales becomes increasingly necessary. An example of such mitigation method at the parcel level is ‘Fill’, which allows development to occur in floodplains by raising parcels of land above base flood elevation using dirt/sand. This flood avoidance technique is often associated with low-density development in flood-prone suburban communities. However, with repetitive flood damages recorded in recent times, it is uncertain whether structures on fill still incur flood damages, or whether these fill parcels have potential adverse impacts on unprotected adjacent parcels. This research addresses the lack of comprehensive knowledge on the impact of fill by posing the following questions: 1) is filling a parcel effective in mitigating flood damages? 2) Does filling a parcel adversely impact flood damages of adjacent unprotected parcels? These questions are answered using a two-step analysis of propensity score matching and a spatial autoregressive model. First, propensity score estimation identified parcels that have the probability of receiving the ‘treatment’ of fill using machine learning methods with increased predictive accuracy compared to other traditional parametric measures. These ‘treatment’ parcels are then matched with fill parcels using appropriate treatment effects and matching calipers, creating a pooled sample of both fill and non-fill parcels in the study area. A post-match analysis of 6,059 filled and non-fill parcels shows a 7% difference in flood damages between fill and non-fill parcels between years 2000 to 2014 in the Clear Creek watershed. A second order analysis of flood damage clusters using a bivariate Ripley’s K point pattern analysis indicates significant clusters of flood damages relative to fill parcel locations. These results highlight the importance of examining parcel-level flood mitigation methods that have cross-jurisdictional economic and planning implications, and the cumulative effect on flood damages at both the community and regional watershed scale. This research also provides insight into the need for synergistic flood risk reduction and incentives to compensate for the use of fill in floodplain development and planning.