Assessing Postdisaster Recovery to Inform Prediction of Economic Loss in Future Floods

This is the authors' abstract. We don't add key points for this paper.

Economic loss assessments in floods traditionally treat each disaster event as isolated and independent, overlooking how the legacy effects of postflood recovery are associated with changes in a system’s sensitivity to sequential climate shocks. By analyzing 3,754 flood events (2012–2024) and 13 years of daily nighttime light data, this study builds a temporal recursive model to quantify how recovery states from prior floods inform the prediction of economic loss in future flood events in US coastal states, with losses measured using FEMA-reported building and content damages. Our results show that within a critical 4–9 months between two flood events, antecedent recovery is associated with a 34.8% lower level of subsequent losses ( − 0.52 ** ), whereas insufficiently consolidated recovery investments during shorter intervals ( < 4 months) are associated with higher subsequent losses. A structural equation model was created to examine nonlinear temporal pathways associated with future loss shifts. Governance factors, including timely aid approval (0.09**), cumulative mitigation investment (0.08*), and renter rate (0.17***), show significant positive associations with postdisaster recovery (T1 recovery) in the SEM. However, their direct SEM paths to subsequent flood loss (T2 loss) are not statistically significant, indicating postdisaster recovery operates as a statistical mediator within the modeled pathway. These findings highlight the necessity of taking recovery states of locations from prior disasters into loss prediction in future events, supporting a shift from static loss assessment toward dynamic, adaptive loss management contingent on disaster sequencing and recovery status to enhance societal resilience to recurrent flooding.