Comparative analysis between impact data related to landslides and extreme rainfall events in Southern Brazil: a proposal to establish potential impact indicators

Disaster monitoring and early warning systems are typically associated with the detection of extreme events capable of causing significant social impacts, particularly in cases of rain-related disasters such as floods, flash floods, and landslides. However, this traditional approach—focused solely on assessing the likelihood of threats materializing—proves insufficient when monitoring areas with high heterogeneity in terms of exposure and population vulnerability. In such cases, less extreme but more frequent events can result in recurring impacts that, when analyzed historically, surpass those of extreme events. In Brazil, approximately 90% of landslide occurrences are associated with low magnitude impact. Low magnitude events cannot be neglected because even though they cause low-severity losses, their high-frequency and cumulative effect adds up to a large number of losses and affected people. Understanding the impacts of low magnitude events can aid in defining risk scenarios as part of the potential impact dimension within a risk matrix. Thus, this study uses a database developed by the Brazilian National Center for Monitoring and Early Warning of Natural Disasters (CEMADEN) to better understand these relationships. Furthermore, it proposes an approach to develop a potential impact indicator based on retrospective risk analyses, linking average impact levels over time to extreme rainfall frequency data. The study focuses on Santa Catarina state (Southern Brazil), analyzing impact data from 80 municipalities between 2016 and 2024. During this time period, the monitored municipalities in the state reported 568 landslide/related impact events, affecting over 8,000 individuals. The analyzed data indicate 548 events with low magnitude impacts, which can be classified as extensive risk events (high frequency, low severity), typically characterized by situations that had 1 to 2 small landslides. On the other hand, 18 events were identified with medium magnitude impacts, where 3 to 10 landslides were generally recorded. Only 2 large magnitude events (>10 landslides) were recorded in the analyzed period, which can be classified as intensive risk events (low frequency, high severity). The results reveal distinct municipal profiles, highlighting two key scenarios: i) areas where the combination of frequent heavy rainfall events and a high potential impact indicator result in very high climate risk and, ii) contrasting situations where significant impact occur despite of low frequency of heavy rainfall suggesting a bigger weight of social vulnerability and exposure of human systems. In addition to providing critical insights for enhancing CEMADEN's decision-making in disaster early warning issuance, the study offers valuable information for prioritizing risk reduction measures and climate adaptation actions.