Spatial Analysis of Flood Connectivity in Brazil: Implications for Risk Management in a Warming Climate

Abstract: Spatially compounding floods represent a class of compound events characterized by the concurrent occurrence of flooding in multiple catchment and water bodies. Despite the substantial impact of intense floods on 93% of Brazilian municipalities between 2013 and 2022, the scientific literature lacks comprehensive descriptions of methods for detecting, monitoring and forecasting spatially compounding floods. Effectively addressing this issue necessitates understanding the drivers underlying these events' spatial interconnectedness. As such, identifying regions susceptible to co-occurring floods is of paramount importance in formulating adaptation and mitigation strategies. Therefore, we conducted an in-depth analysis of spatially compounding river flood connectivity within catchments across Brazil. We investigate a 39-year (1980-2019) dataset of 511 catchments and extract flood events at individual sites using the Peak-Over-Threshold (POT) approach. We established the 15th percentile of annual maxima as a threshold and considered a minimum time lag of 10 days between consecutive events. Subsequently, we quantified the spatial connectedness of flood events through the number of co-occurrences between pairs of catchments. A flood event was considered to co-occur with another if the catchments experienced a POT event within a window of ±2 days from a given date of occurrence. Our results revealed significant variations in flood characteristics and connectivity patterns across the Brazilian territory, with an average of 2 flood events per year per catchment. Particularly, the Central-Northeast regions demonstrated heightened connectivity, indicating a higher likelihood of experiencing multiple and widespread flood occurrences. Conversely, the Northwestern Amazon and catchments in the Southern part of the country displayed lower connections to other regions, suggesting a prevalence of independent flood events in these locations. Our findings hold valuable implications for adaptation, mitigation, and resilience for extreme events, to guide policymakers in risk management and urban design in flood-prone areas.