Extreme rainfall and landslides as a response to human-induced climate change: a case study at Baixada Santista, Brazil, 2020

Abstract In March 2020, an extreme rainfall in Baixada Santista, Brazil, led to a series of landslides affecting more than 2,800 people and resulting losses exceeding USD 43 million. This attribution study compared extreme rainfall in two distinct runs from the UK Met Office Hadley Centre HadGEM3-GA6 model: one with all forcings and the other with natural forcings only, considering antecedent conditions and soil moisture (extreme 60-day rainfall, Rx60day) and heavy rainfall (extreme 3-day rainfall, Rx3day) which may trigger landslides. The long-term set-up became 74% more likely, while the short-term trigger was 46% more likely. The anthropogenic contribution to changes in rainfall accounted for 20-42% of the total losses and damages. The greatest economic losses occurred in Guarujá (42%), followed by São Vicente (30%) and Santos (28%). Landslides were responsible for 47% of homes damaged, 85% of homes destroyed, all reported injuries, and 51% of deaths associated with heavy rainfall. Changes in land cover and increased urbanization showed a pronounced increase in urbanized area in Guarujá (107%), São Vicente (61.7%) and Santos (36.9%) and a reduction in farming area. Population growth estimates also indicate higher exposure to extreme rainfall events. The proportion of irregular and/or precarious housing indicates vulnerability, with Guarujá being the municipality with the highest number (34.8%) of dwellings of this type. Our estimates suggest that extreme precipitation events are having shorter return periods due to climate change and increased urbanization and population growth is exposing more people to these events. These findings are especially important for decision-makers in the context of disaster risk reduction and mitigation and adaptation to climate change.