Using remote sensing for exposure and seismic vulnerability evaluation: is it reliable?

Abstract

Conducting field surveys for exposure and seismic vulnerability evaluation is the most costly, resource-intensive task when assessing earthquake risk. During the past decade, risk analysts have been trying to alleviate this using remote sensing for building characterization. However, the use of vulnerability databases created with remote sensing had not been sufficiently validated thus far. In this paper, we have created an exposure and seismic vulnerability database in Port Prince (Haiti) using freely accessible aerial ortho-imagery and LiDAR points. We have validated this database against two reference datasets from different, independent studies. Then, we have computed an earthquake damage scenario to test whether remotely sensed data are actually valid for seismic risk evaluation. We have seen how our vulnerability database yields an accurate damage distribution with a low Mean Absolute Percentage Error of 3.78% when compared to the damage obtained with the reference vulnerability dataset. Further, we have conducted a thorough comparison of the cost that entails creating a vulnerability database using remote sensing with a traditional field survey. Twelve international experts have collaborated in the cost estimation of a typical in-field building inspection. As a result, we have found that using remote sensing techniques allows for saving up to 75% of the cost and 85% of the time. These outcomes seem to prove both, the technical and economic feasibility of remote sensing for seismic vulnerability assessment. Thus, we have proposed a 5-step approach for evaluating building vulnerability that combines both, the analysis of remotely sensed data and a reduced, targeted field survey to optimize time and cost. The final goal is to help cities reach the Sustainable Development Goal nr. 11.B to increase their resilience against disasters.