.png){kind=icon}
.png){kind=icon}
Return
Improving flood damage estimation by integrating property elevation data
.png){kind=icon}
PRE
AI
DOI：10.1016/j.ijdrr.2025.105251.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
The research objective is to use property elevation data to enhance flood damage models. To accomplish this, the research assesses two super-neighborhoods: Meyerland and Edgebrook in Harris County, Texas. In this study, property elevation refers to standard characteristics of a household, such as lowest floor elevation (LFE) and the height difference between the street and lowest floor (HDSL). The research uses these characteristics as predictors of flood damage to inform decision makers on a property's susceptibility to flooding. These characteristics were gathered from Streetview images (SVIs) using a computer vision model. For flood damage ground truth data, the research uses Property Damage Extent (PDE) data which is computed based on a combination of National Flood Insurance Program (NFIP) and Individual Assistance (IA) claims. Moreover, Hurricane Harvey flood depth estimates are obtained from HEC-RAS 2D model and incorporated into the damage models. The results of the one-way ANOVA test show that the two neighborhoods are statistically different in terms of LFE, HDSL, PDE, and flood depth; therefore, future analysis in this study will be done independent of the neighborhoods. The Pearson correlation analysis shows statistically significant results between property elevation and PDE, thus are suitable predictors of damage. Finally, log-log regression models are used to analyze the effect of both flood depth and property elevation on PDE. The results show that models with property elevation have higher R2 values, statistically significant coefficients, and can complement traditional flood models by providing more insight on damage. This study addresses the data gap in flood damage models, by demonstrating how computer vision obtained elevations can be empirical predictors of damage. Additionally, the introduction of HDSL into damage assessment offers new insights that can complement the findings from traditional parameters such as LFE and flood depth.
Keywords:
Lowest floor elevation
Height difference between Street and lowest
floor
Rapid damage assessment
Flood risk
Journal
IF:
4.5
Papers:
6.0K
Citations:
2.1W