Evolving Dependence Structures Between Compound Flood Drivers Under Future Climate Scenarios: A case study over Greater Boston

The assessment of compound flood risk often relies on the assumption that the dependence structure between flood drivers (e.g., rainfall intensity, coastal water levels, and streamflow) remains stationary under changing climatic conditions. Yet, traditional approaches that inherently assume stationary dependencies, or rely solely on historical relationships, may misrepresent flood risk and fail to identify hotspots of emerging infrastructure vulnerabilities. This study aims to (a) characterize the dependence structure between compound flood drivers using a parsimonious parametric framework, and (b) explore potential changes in this structure under future climate scenarios, by leveraging outputs from the Coupled Model Intercomparison Project Phase 6 (CMIP6) regional climate projections. An ensemble of synthetic and historical storms is employed to simulate flood impacts across the Greater Boston region, forming the basis for statistically modeling the conditional dependence of the main flood drivers. Changes in the marginal distributions of these drivers, informed by CMIP6 simulations under various Representative Concentration Pathways (RCPs), are also integrated into the dependence framework to evaluate future trajectories of compound flood risk. The findings focus on determining whether shifts in the dependence structure offer a more nuanced understanding of evolving flood risk profiles, as well as identifying areas where traditional stationary assumptions may result in systematic errors. Ultimately, the study advances understanding of the dynamic interplay between flood drivers under future climate scenarios, and supports the development of adaptation strategies for regions vulnerable to compound flooding.