Vulnerabilidade natural e socioambiental relacionando eventos de precipitação na região metropolitana do Recife-PE.

Abstract

Mass movements are natural disasters often triggered by precipitation, causing significant socioeconomic impact in urban areas, especially when combined with natural susceptibility and local population vulnerability. This study aims to analyze the influence of natural and socioeconomic factors on the occurrence of these in the Metropolitan Region of Recife (Pernambuco, Brazil), associated with rainfall events, between 2016 and 2022. The region was selected due to its recognized vulnerability during rainy periods. The methodology followed a multidisciplinary approach structured in five stages: (1) compilation and spatialization of the mass movement inventory based on CEMADEN data; (2) determination of critical rainfall thresholds using APAC data, considering the intensity–duration (I–D) and accumulated precipitation–duration (E–D) relationships at 24-, 48-, 72-, 96-, and 120-hour intervals; (3) assessment of natural susceptibility based on topographic, geomorphological, environmental, and anthropogenic factors, with a comparative application of machine learning models (Random Forest, KNN, XGBoost, Naïve Bayes); (4) development of a social vulnerability map based on IBGE socioeconomic indicators; and (5) integration of the susceptibility and vulnerability maps to define Socionatural risk. A total of 221 mass movements were identified, and critical precipitation thresholds were established. The E–D relationship showed a higher correlation with the occurrences (R² > 97.5%) than the I–D relationship (R² < 90.5%). Among the conditioning factors, land use and altitude proved to be the most relevant. The Random Forest model achieved the best performance (AUC = 0.935), followed by XGBoost (0.875), Naïve Bayes (0.828), and KNN (0.799). The construction of the Socionatural risk map integrated physical and socioeconomic factors. Five risk classes were identified, with the “high” and “very high” classes (16.5% of the area) concentrating 85% of the recorded occurrences. Neighborhoods such as Ibura (Recife) and Cavaleiros (Jaboatão dos Guararapes) exhibited critical risk due to high population density, inadequate infrastructure, and elevated natural susceptibility. The results indicate that the critical rainfall period corresponds to accumulations of up to three days, particularly for Z2, with a critical accumulated rainfall of 15.4 mm, highlighting the need for public policies should prioritize urban areas with overlapping social and natural vulnerabilities, thereby providing a technical basis for proactive measures to mitigation the impacts of mass movements in the Metropolitan Region of Recife.