Contact Authors
Davide Faranda (IPSL-CNRS, FR) - davide.faranda@lsce.ipsl.fr - EN/FR/IT
Suzana Camargo (Columbia University, USA) – suzana.camargo@columbia.edu – PT/EN
Haosu Tang (University of Sheffield, UK) – haosu.tang@sheffield.ac.uk – ZH/EN
Neven Fučkar (University of Oxford, UK) - neven.fuckar@ouce.ox.ac.uk - EN/HR
Tommaso Alberti (INGV, IT) – tommaso.alberti@ingv.it – IT/EN
Citation
Faranda, D., Camargo, S. J., Tang, H., Fučkar, N. S., & Alberti, T. (2026). Heavy precipitation in February 2026 Brazil Floods likely influenced by both human-driven climate change and natural variability. Zenodo. https://doi.org/10.5281/zenodo.18787782
Press Summary
Depressions similar to those producing the February 2026 Brazil floods are up to +6 mm/day (15–20%) wetter than they were in the past.
This event was driven by very rare meteorological conditions.
We mostly ascribe the heavy precipitation during the February 2026 Brazil floods to human-driven climate change, while natural climate variability likely played a modest role.
On 23–24 February 2026, exceptionally heavy rains battered southeastern Brazil, especially in the state of Minas Gerais, and its cities of Juiz de Fora and Ubá, triggering catastrophic flooding, multiple landslides, and widespread disruption across urban and rural areas. Torrential downpours caused rivers such as the Paraibuna to burst their banks, with some areas recording rainfall far above the normal February totals, turning streets into raging torrents and saturating slopes above communities. The deluge was associated with persistent convective rainfall fed by moist tropical air over the mountainous terrain, producing intense precipitation that overwhelmed natural and built drainage systems. The city of Juiz de Fora experienced its wettest February on record so far, more than 240% above climatology (170.3 mm), and recording 229.9 mm for 22-24 February (INMET), prompting local authorities to declare a state of calamity as municipal services struggled to cope. Authorities warned of continued rain and heightened risk of further floods and mudslides as rescue and assessment operations proceeded.
The human toll was severe: at least 53 people were confirmed dead across affected areas, with dozens more unaccounted for and presumed missing after sudden floods and hillside collapses engulfed homes and neighbourhoods, particularly overnight when residents were asleep. In Juiz de Fora, dozens of deaths were attributed directly to landslides that buried houses in hillside districts, while the nearby city of Ubá also reported multiple fatalities as floodwaters surged through low-lying areas. Emergency services reported that more than 3,000 residents were forced from their homes. Around 440 people evacuated to shelters as roads were blocked by debris and bridges, and other infrastructure were damaged or washed away. The Brazilian President announced deployment of national disaster response teams,healthcare workers and mobile care units to support local efforts. Schools and public services were suspended as the crisis unfolded.
Officials emphasised the ongoing hazard posed by saturated soils and swollen watercourses as meteorological forecasts indicated continued rainfall in the region. Fire departments, civil defence units, and volunteer brigades worked through mud and debris in search and rescue missions, while authorities urged residents to avoid flood-prone zones and follow safety directives as recovery and relief operations continued.
The meteorological conditions during 23–24 February 2026 were characterised by persistent negative surface pressure anomalies over southeastern Brazil, with values locally below −2 to −4 hPa relative to climatology across Minas Gerais, indicating a sustained low-pressure system favouring large-scale ascent. This surface pressure pattern was combined with westerly circulation at altitude close to the tropopause, which supports cloudy weather with persistent rain throughout the state. Near-surface temperature anomalies were predominantly negative, reaching around −2 °C in parts of the affected region, consistent with widespread cloud cover and intense rainfall limiting daytime warming. Daily precipitation totals were exceptionally high, locally exceeding 50 mm/day, with the most intense rainfall concentrated over central and southern Minas Gerais, including the Belo Horizonte and Sete Lagoas areas. Near-surface winds were moderate, generally below 30 km/h, suggesting that the severity of the event was primarily driven by prolonged and intense rainfall rather than extreme wind conditions. The combination of low pressure, sustained moisture convergence and persistent precipitation created highly favourable conditions for flooding and landslides across the region.
Climate and Data Background for the Analysis
The impact of climate change on changes in precipitation and flooding in the state of Minas Gerais in Southeast Brazil is characterized by a mix of increasing extreme precipitation events and changes in regional precipitation patterns, with implications for drainage systems, agriculture, and water resources. Changes in regional precipitation have been observed in different parts of Brazil over the past century. For example, significant increases in extreme precipitation have been noted in southeastern Brazil, while non-significant decreases have been found in central Brazil (IPCC AR6 WGI FR - Page 1116). In southeast Brazil, there has been a region of highly significant decrease in rainfall in both wet and dry seasons during the period 1979-2011 (IPCC AR6 WGI FR - Page 2012). A warmer climate is projected to intensify both very wet and dry weather events, increasing the risk of flooding or drought events in the region (IPCC AR6 WGI SPM - Page 19). In particular in the South-Eastern South America, an increase in both mean and extreme precipitation has been observed since 1960 with high confidence; the intensity and frequency of extreme precipitation and pluvial flooding is projected to increase with medium confidence from a 2oC of global warming and above (IPCC AR6 WGI TS4.3.2.4 - Page 140). Changes in precipitation and extreme temperatures are impacting agricultural production in the region, while also increasing the risk of landslides. Although this increase in mean precipitation can positively impact agriculture in some areas, the extremely long dry spells that comes with it negatively affects economies in southeastern Brazil (IPCC AR6 WGII FR - Page 1703).
Our analysis approach rests on looking for weather situations similar to those of the event of interest having been observed in the past. For this event we have medium-high confidence in the robustness of our approach given the available climate data, as the event is similar to other past events in the available data record.
ClimaMeter Analysis
We analyze here (see Methodology for more details) how events similar to the meteorological conditions leading to the Brazil floods have changed in the present (1988–2025) compared to what they would have looked like if the event had occurred in the past (1950–1987) in the region [50W, 35W, 15S, 30S]. Surface-pressure changes show a modest but coherent signal, with present-day conditions associated with slightly higher surface pressure than in the past over Minas Gerais, suggesting a weaker low-pressure anomaly for comparable events today. Temperature changes indicate a clear and robust warming signal, with increases of about +0.8 to +1.5 °C across the affected region, consistent with a warmer background climate state during similar meteorological configurations. Precipitation changes are spatially heterogeneous but indicate locally wetter conditions over central and eastern Minas Gerais, with increases locally exceeding +3 to +5 mm/day, aligning with the areas most impacted by flooding during the event. Wind speed changes are generally weak and slightly negative, indicating that comparable events today tend to be associated with similar or slightly reduced near-surface wind speeds, reinforcing the conclusion that rainfall, rather than wind, is the dominant hazard driver.
Similar past events show a seasonal redistribution, with a higher fraction of analogues occurring in February and March in the present compared to the past, suggesting a shift toward later-summer occurrence of such heavy-rainfall configurations. Changes in urban areas are consistent with the regional signal: Juiz de Fora and Sete Lagoas exhibit warmer and wetter conditions compared to past analogues, while Patos de Minas shows warming but weaker or even slightly reduced precipitation signals. These changes imply increased exposure of urban areas to flood risk during similar meteorological situations today, driven primarily by enhanced rainfall intensity occurring on top of a systematically warmer climate background.
Finally, we find that sources of natural climate variability, notably the Pacific Decadal Oscillation, may have influenced the event. This suggests that the changes we see in the event compared to the past may be partly due to human-driven climate change, with a contribution from natural variability.
Conclusion
Based on the above, we conclude that precipitation associated with depressions similar to those causing the February 2026 Brazil floods has increased by up to 6 mm/day, corresponding to roughly 15%, in the present compared to the past. We interpret these floods as an event driven by very rare meteorological conditions, whose thermodynamic characteristics can be largely ascribed to human-driven climate change.
NB1: The following output is specifically intended for researchers and contain details that are fully understandable only by reading the methodology described in Faranda, D., Bourdin, S., Ginesta, M., Krouma, M., Noyelle, R., Pons, F., Yiou, P., and Messori, G.: A climate-change attribution retrospective of some impactful weather extremes of 2021, Weather Clim. Dynam., 3, 1311–1340, https://doi.org/10.5194/wcd-3-1311-2022, 2022.
NB2: Colorscales may vary from the ClimaMeter figure presented above.
The figure shows the average of surface pressure anomaly (msl) (a), average 2-meter temperature anomalies (t2m) (e), cumulated total precipitation (tp) (i), and average wind speed (wspd) in the period of the event. Average of the surface pressure analogs found in the counterfactual (b) and factual periods (c), along with corresponding 2-meter temperatures (f, g), cumulated precipitation (j, k), and wind speed (n, o). Changes between present and past analogues are presented for surface pressure ∆slp (d), 2-meter temperatures ∆t2m (h), total precipitation ∆tp (i), and wind speed ∆wspd (p): color-filled areas indicate significant anomalies with respect to the bootstrap procedure. Violin plots for past (blue) and present (orange) periods for Quality Q analogs (q), Predictability Index D (r), Persistence Index Θ (s), and distribution of analogs in each month (t). Violin plots for past (blue) and present (orange) periods for ENSO (u), AMO (v), and PDO (w). Number of the analogues occurring in each subperiod (blue) and linear trend (black). Values for the peak day of the extreme event are marked by a blue dot. Horizontal bars in panels (q,r,s,u,v,w) correspond to the mean (black) and median (red) of the distributions. (x) Number of analogues found in sub periods when analogues are searched in the whole reanalysis period.