Our analyses focused on the portion of the South Atlantic watershed that lies within Rio Grande do Sul state (RS), hereafter referred as SA-RS, covering approximately 141,700 km2 and 50% of the state area (Fig. 1). The SA-RS encompasses an ecotone area between the Atlantic Forest and Pampa, and it harbours several native vegetation types, including forests, grasslands, wetlands, and costal vegetation on sandy soils (restingas). This region is intensively used for agriculture, including croplands, forestry, and livestock (Rezende et al., 2018; Baeza et al., 2022).

Fig. 1.

Map of South America highlighting the study area, the portion of the South Atlantic watershed located within Rio Grande do Sul state (SA-RS), Brazil, with its remaining native vegetation, Protected Areas, and the flooded area in May 2024.

Using climate historical data available from the WorldClim database (Fick and Hijmans, 2017; https://www.worldclim.org), we calculated the historical mean annual precipitation and the historical May precipitation in the SA-RS for comparison to the observed precipitation in May 2024. The mean annual precipitation is 1,508 ± 179 mm (mean ± standard deviation) and the mean May precipitation is 115 ± 16 mm. In May 2024, this region experienced an extraordinary heavy rainfall. Porto Alegre recorded 536 mm, and other cities reached 845 mm (INMET, 2024).

The map of the flooded areas was generated using two datasets: one for the Guaíba hydrographic region (Possantti et al., 2024) and the other for the Patos hydrographic region (Silva et al., 2024). For the Guaíba hydrographic region, we used the area directly affected on the peak day, May 6, 2024 (GIS layer: analise_ada_inundacao_20240506). As similar data were not available for the Patos hydrographic region, we used a simulation of inundated areas considering a flood quota of 2.73 m (GIS layer: SIG_inundacao_cota273), which is slightly below the measurements made on the peak day at this region, May 16, 2024. This estimate of the flooded area indicated that over 9,800 km² in the SA-RS were directly affected.

We used land cover maps (MapBiomas, 2023, Collection 8) to assess which vegetation types, hereafter referred to as ecosystems, were most impacted by the flooding. We overlapped the distribution of these ecosystems with the flooded area map.

Brazilian Native Vegetation Protection Law (Brazil, 2012) demands that native vegetation must be kept on special areas (especially at water bodies margins and areas with slopes exceeding 45 °) – the Permanent Preservation Areas (PPAs). Following this law requirements, we created an estimated PPAs map for both waterbody margins and areas with a slope exceeding 45 ° using a hydrography map from Hasenack and Weber (2010) and the ANADEM Digital Terrain Model (Laipelt et al., 2024).

For rivers, the PPA ranges from 30 to 500 meters from river margins, depending on river width. Therefore, we measured widths of the main rivers in the SA-RS: Taquari, Sinos, Jacuí, Vacacaí, Vacacaí-Mirim, Pardo, Antas, Caí, Gravataí, Camaquã, and Mampituba. Measurements were taken every two kilometers, starting from the headwaters to the river mouth, continuously changing PPA size in accordance with river width. For all other rivers, a standard 30-meter buffer was applied for PPA estimation. For natural lakes, PPAs of 100 meters were assigned to lakes larger than 20 hectares, while 50 meters were assigned to smaller lakes. These PPA estimates adhered solely to the general rule of the Native Vegetation Protection Law, excluding the transitional provisions outlined in Chapter XIII, which offer mitigations primarily for small rural properties. This PPAs map was overlaid with the flooded area and with the ecosystem maps to identify which PPAs areas were directly affected by the flooding and whether they preserved their natural cover, in accordance with the legal requirements.

Apart from PPAs, Brazilian legislation establishes Protected Areas (PAs) through the National System of Protected Areas (Brazil, 2000). These PAs are legally designated, have defined boundaries, and are established to conserve nature and ensure the sustainable use of natural resources. Data on PAs was obtained at National Register of Conservation Units – CNUC.

We analysed which tetrapod species (amphibians, reptiles, birds, and mammals) were potentially affected by the flooding through spatial analysis. We obtained distribution range maps for amphibians and reptiles from the IUCN Red List database (IUCN, 2024), for birds from the BirdLife International and Handbook of the Birds of the World (2023), and for mammals from Marsh et al. (2022). Marine species (e.g., turtles and whales) and exotic species were excluded. For birds, only resident and breeding ranges were considered.

To identify species potentially affected by the flooding, we overlapped species distribution range maps with flood extent. A spatial intersection determined the area within each species' range affected by the flooding. The percentage of species distribution affected by the flooding was then computed as the ratio of the affected area to the total distribution area of each species. All spatial operations were performed using the sf package (Pebesma and Bivand, 2023).

We calculated the proportion of species that were potentially affected by the flooding in relation to all species of each group occurring in the RS and in the SA-RS to estimate the impact on the tetrapod diversity. We also calculated this proportion in relation to the number of threatened species in the RS (Rio Grande do Sul, 2014) that were potentially affected by the flooding. For this analysis, we considered only the categories Vulnerable (VU), Endangered (EN), and Critically Endangered (CR). The categories Near Threatened (NT) and Data Deficient (DD) were considered as non-threatened. Since the red list of threatened species in the RS was published ten years ago, we also considered the IUCN Red List, classifying a species as threatened based on the higher threat level from either list.

To assess how the percentage of affected area of species distribution varied among taxonomic groups and between threatened and non-threatened species, we selected only those species that had more than 0.1% of their distributions affected by the flooding and we run a one-way analysis of variance (ANOVA), followed by a Tukey test for comparison between groups. Furthermore, we used a generalized linear model (GLM) to assess if the percentage of species distribution affected by the flooding was related to species conservation status, i.e., varied between threatened and non-threatened species. For this analysis, we ran four GLMs, one for each taxonomic group and one GLM including all species. For amphibians and reptiles, we assembled all species, since there are few species considered threatened in each group, which we called herpetofauna. Since the data is right skewed, we used GLM with an inverse gaussian family. Additionally, we also compared species richness between flooded and unflooded areas for each group using one-way ANOVA (see results in Supplementary Material). All analyses were run in the R environment, version 4.1.3 (R Core Team, 2022).

Our results revealed that at least 1,200 km2 of forests and 1,020 km2 of grasslands were flooded. Proportionally, the most affected ecosystems were wetlands, locally called banhados, aquatic ecosystems and costal vegetation on sandy soils. Native vegetation in the SA-RS covers approximately 50% of its original extent, with only 2.8% located within a PA. Total affected area of native ecosystems, ca. 4,300 km2, is larger than all the PAs in the SA-RS, which cover about 4,000 km2. This already limited network of PAs was also highly impacted, with 25.7% of strict PAs and 16.9% of sustainable use PAs flooded. This represents more than 825 km2 of PAs directly impacted by the flooding. The extent of this impact highlights the urgent need to restore ecosystems in areas that had already lost most of its native vegetation (Strassburg et al., 2020) and to expand the PA network in underrepresented areas (Lima et al., 2020; Malecha et al., 2023), incorporating climatic risk scenarios into conservation planning (Nguyen et al., 2022).

PPAs are mandatory to maintain their natural vegetation and aim to preserve water resources, landscapes, geological stability, biodiversity, and soil quality, thereby ensuring the well-being of human populations (Brazil, 2012). However, we showed that 40% of the PPA related to water body margins had already lost their original cover within the SA-RS. This number rises to 67% when considering the PPA directly affected by the flooding. PPA related to areas with slopes exceeding 45 ° are in an even worse situation. Despite 93% of these PPAs on the SA-RS preserving their natural cover, those directly affected by the flooding had already lost about 80% of their natural vegetation, increasing landslide risks. Maintaining and restoring native vegetation is a nature-based solution with high efficiency in diminishing flooding risks after extreme rainfall events (Manes et al., 2024). While native vegetation in all the landscape is important for mitigating the effects of rainfall, the importance of PPAs is even higher, because they designed to protect areas of higher vulnerability. Therefore, it is imperative to prioritize the conservation and restoration of PPAs to safeguard ecosystems and human communities. This preservation should be reinforced into national and regional legislations. However, recent legislative changes in the RS indicate otherwise. The recently reviewed Environmental Code of Rio Grande do Sul (Rio Grande do Sul, 2020) significantly weakened natural vegetation protection (Ferraz et al., 2024). At the national level, despite the general rule mandating the preservation of PPAs, the transitional provisions of the same law ease these requirements in many situations (Brazil, 2012), weakening the protection of these vulnerable areas. Further than legislative requirements, carbon market could provide a financial support to small landowners for maintaining natural vegetation on PPAs (Evans, 2018; Dutra et al., 2024) and should be implemented on flooding vulnerable areas.

A high number of species were potentially affected by this major flooding, representing 85% of the tetrapod species in the SA-RS. However, as IUCN range maps might underestimate the complete species' geographical range (Pineda and Lobo, 2012), the number of affected species might be even larger. Amphibians and reptiles had the highest percentage of their distributions affected. Furthermore, amphibians and reptiles are considered the most vulnerable to flooding due to their specific habitat requirements and low mobility (Zhang et al., 2021). Specially, the endangered lizard Liolaemus arambarensis, who lives in coastal sand dunes, was the most affected species. Mammals and birds, in general, are considered more resilient due to their higher mobility. However, they still may face significant challenges related to reproduction areas and resource depletion. Additionally, higher mobility is not true for all birds and mammals. For instance, affected populations of tuco-tucos (Ctenomys), which have burrowing lifestyle, with three species among the top-ranked affected species, may have suffered high mortality rates, exacerbating their conservation status in the future. Noteworthy, other burrowing animals were potentially affected, such as caecilians and cavies. These results are concerning because extreme events have already caused tetrapod’s extinctions, especially for small-ranged species (Lips, 2018). Almost all top-ranked affected species are small-ranged, aggravating the flooding impact on their conservation status. Worrisomely, threatened species were disproportionately affected by the flooding, especially amphibians, reptiles, and mammals. Eighty-four threatened species were potentially affected, and the percentage of their distribution affected was twice as large as that of non-threatened species. These findings emphasize the urgent need for targeted conservation efforts to protect species small-ranged, with low mobility and that are already threatened from the increasing frequency and intensity of floodings.

Unfortunately, all climate models indicate that many regions of the world will face extreme rainfall and, therefore, major flooding events in the decades to come (IPCC, 2021). The SA-RS experienced its largest flooding ever recorded, and the impacts on biodiversity are massive, in addition to socioeconomic impacts. To prevent effects of future extreme rainfall events, it is imperative to promote native vegetation restauration, especially in the PPAs and PAs. To mitigate the effects of this flooding, many initiatives are necessary. Both traditional and nature-based solutions can be applied to mitigate flooding caused by extreme events. Under traditional solutions, sustainable technologies (e.g., permeable pavements) and gray infrastructures (e.g., ditches and culverts) have positive impacts both ecologically and socially. However, nature-based solutions, such as urban parks, rain gardens, and green infrastructure, most effectively attenuate flooding hazards (Prado et al., 2024). More importantly, maintaining and restoring native vegetation is key to diminish flooding risks and landslides (Manes et al., 2024).

In addition to emergency funding to rebuild infrastructures adapted to extreme events, it is necessary to promote funding for ecosystems and habitat restoration, especially for (i) amphibians and reptiles, (ii) tetrapod species with reduced mobility, (iii) small-ranged species, and (iv) threatened species. Additionally, special grants should be promoted to further research on: (i) the effects of this flooding on aquatic ecosystems, (ii) the stability of the affected ecosystems post-disaster with resilience and resistance measures, and (iii) fieldwork to evaluate the effects of this flooding on biodiversity in situ, especially for the top-ranked affected species.

We emphasize that our estimations are based on pre-disaster species distribution data, and we did not account for potential range contractions, mortality estimates, or habitat loss resulting directly from the flooding. However, using range maps is a first approximation to obtain the magnitude of this extreme event. Therefore, future studies should include post-disaster surveys and monitoring programs to evaluate population trends, particularly for threatened species. Standardized populational monitoring, using techniques such as mark-recapture, camera traps, acoustic monitoring, and eDNA, can provide critical data on species persistence, recolonization potential, and habitat use after extreme events. Long-term ecological studies will also be essential to understand delayed impacts and ecosystem recovery trajectories. Here we focused only on tetrapod species, but fishes, plants and invertebrates were likely to be impacted by the flooding as well. Generating such data is vital for building more accurate impact models and informing conservation and restoration priorities under increasing climate variability.