Wildfires disproportionately affected jaguars in the Pantanal

The jaguar (Panthera onca) is globally listed as Near Threatened, with several sub-populations critically endangered. Although relative stability has been assumed for the Amazon and Pantanal biomes, mounting pressures in the Pantanal—including habitat loss, prey poaching, retaliatory killing due to livestock depredation, pollution, agricultural expansion, and infrastructure—are escalating conservation risks. Fire has historically been considered a lesser threat to jaguars range-wide, but the unprecedented severity of the 2020 Pantanal fires raises concern about unaccounted risks. An unusual number of fires began in the 2019 wet season and continued into the 2020 dry season, driven by human ignition and severe drought, burning through typically inundated areas. The purpose of this study is to quantify the reach and intensity of fires affecting jaguar conservation in the Pantanal over the last 16 years and to determine the disproportionate impacts on jaguar numbers, home ranges (HRs), and protected areas (PAs) overlapping HRs.

Background literature identifies multiple threats to jaguar persistence in and around the Pantanal: habitat loss and fragmentation, poaching of prey, human-wildlife conflict tied to livestock depredation, pollution from mining and pesticides, agricultural intensification, roads and dams. Prior to 2020, fires were generally considered to affect a small proportion of jaguars range-wide; however, recent analyses highlight that extreme drought, human-caused ignition, and accumulation of flammable biomass can produce severe wildfires in the Pantanal. Ecological studies indicate that fires can destroy riparian vegetation, keystone tree species, and patchy forests, and likely impact less mobile fauna (e.g., anteaters, armadillos, sloths, reptiles). Productivity (GPP) declines after fires have been documented in the Pantanal, with implications for jaguar space use because higher productivity areas are associated with higher jaguar densities and smaller HRs. The literature also underscores the critical role of PAs for biodiversity and jaguar conservation, while warning that PA extent in the Pantanal and upstream headwaters is limited and was reduced historically, potentially heightening fire impacts and species displacement.

Study area: The Pantanal, the world’s largest wetland, spans Brazil (about 78% of area), Paraguay, and Bolivia, embedded within the Upper Paraguay River Basin (UPRB). It is a mosaic of floodable and non-floodable habitats. Climate is seasonally wet tropical with strong spatial variation in precipitation and temperature; hydrology and fire jointly influence vegetation and wildlife dynamics. Data sources and spatial framework: The Pantanal extent was defined by merging the legal Brazilian Pantanal biome boundary with the Pantanal ecoregion within the UPRB, and analyses were stratified by country. Protected areas (PAs) polygons were compiled from national and global databases; some private PAs may be missing. Fire data: Near-real-time MODIS active fire detections (MOD14/MYD14; FIRMS/LANCE) were obtained via Google Earth Engine (GEE) for 2005–2020. Fire occurrence was defined for pixels with brightness temperature ≥325 K; intensity was quantified from fire brightness temperature. Fire metrics were computed across the Pantanal, within jaguar HRs, and within PAs overlapping HRs. Hydrological and climatic covariates (precipitation, river depth) and land-cover/land-use information (MapBiomas) complemented analyses. Jaguars: Two complementary datasets were used. (i) Published, range-wide spatial predictions of jaguar density were used to estimate the number of individuals present in fire-affected areas as a proxy for potential displacement/injury/mortality. (ii) HRs for 48 resident jaguars (tracked 2005–2016 across Brazilian, Paraguayan, and Bolivian Pantanal) were estimated using GPS telemetry and the autocorrelated kernel density estimator (AKDE). Individuals were included if residency was confirmed via movement semi-variograms and related diagnostics (POD > 0.95, asymptotic semivariograms). HR probability surfaces (AKDE-derived rasters) were aligned to the fire dataset resolution. Analytical approach: Annual overlays of fire occurrence/intensity with jaguar density rasters produced proxies of the number of jaguars affected per year and by country. Overlays of active fire pixels with HR rasters quantified the proportion and area of HRs affected, and similarly for PAs overlapping HRs. Temporal analyses summarized 2005–2020 trends and contrasted 2020 against the 2005–2019 distribution (median, mean). Complementary statistical assessments (e.g., ANCOVA, regression/GAMs) examined the dependence of fires within HRs on year and region. Assumptions: HRs from tracked resident individuals were treated as stable priority areas (2nd-order selection) likely to be occupied by resident jaguars across years if conditions remained similar. Due to sample-size and data availability constraints, annual estimates of numbers of jaguars, HRs, and PAs within HRs were assumed constant in some summaries.

- Fire activity and drought: Fire occurrences and intensity increased with drought conditions from 2019 to 2020. In 2020, mean fire brightness was the highest of the 2005–2020 period (352.3 K), 16 K above the previous 15-year median. - Burned extent: In 2020, fires affected 31% of the Pantanal (87% of burned area in Brazil), equivalent to 33% of the Brazilian biome’s legal boundary area. PAs were heavily impacted, with 62% of Brazilian Pantanal PAs affected. - Jaguars affected (population proxy): 2020 fires overlapped areas of high jaguar density, yielding an estimated 746 individuals affected out of 1,686 estimated in the Pantanal (about 45%). This was 3.3 times the median annual value over the prior 15 years. - Home ranges (HRs): 2020 had the highest proportion and extent of HRs burned in 16 years; 38 of 48 HRs (79%) were affected. The median burn extent within HRs was 78%, totaling about 2,718 km² across HRs. Mean fire intensity within HRs was five times the 2005–2019 median. In the northern Pantanal, a median of 97% of HR area burned (mean 87%), with approximately 2,098 km² of HRs affected. - Protected areas within HRs: In 2020, 78% of the total area of PAs overlapping HRs burned. Fires occurred in 54% (n = 26) of HRs with PAs, with a median of 94% of each PA area burned. The total PA area burned within HRs was 1,504 km² across the Pantanal (Brazil: 970 km², 72%). In the northern Brazilian Pantanal, mean PA burned within HRs was 91% (median 100%). - Multi-year context: Besides 2020, years with high proxies of jaguars affected included 2019, 2007, and 2005. Within HRs, burned areas in 2019 and 2005 reached 1,196 km² and 870 km² (3.6x and 2.8x the 2005–2019 median of 329 km²), while 2020 reached 8.3x the median. Fire intensity within HRs in 2019 and 2005 was ~2x and 2.4x the historical median (46 k), while 2020 exceeded 5x.

The analyses demonstrate that the 2020 Pantanal wildfires disproportionately impacted jaguars compared to the nominal fraction of the biome burned. Nearly half of the estimated jaguar population occupied fire-affected areas, and most resident HRs, including those overlapping PAs, experienced extensive burning. These findings directly address the study’s question by quantifying the spatial co-occurrence of fire with jaguar density, HRs, and PAs over 16 years, highlighting the exceptional severity of 2020 relative to the historical baseline. The ecological significance is substantial: fires destroyed core habitats, riparian vegetation, and keystone plant species, injured jaguars (with documented rescues and mortalities), and likely caused displacement, increased territorial conflicts, reduced fecundity, and diminished survival. Given that jaguar density and HR size are linked to productivity, documented post-fire reductions in GPP in the Pantanal may reduce habitat quality and carrying capacity. PAs, though critical refugia with higher biodiversity and jaguar densities, were also severely affected; their limited extent in the Pantanal and surrounding headwaters compounds vulnerability and may force jaguars into suboptimal, human-dominated habitats, increasing conflict and mortality risks. The results reinforce that regional land-use changes (agriculture, ranching, dams/roads) and climate change intensify droughts and fuel accumulation, thereby elevating fire risk. Effective conservation will require integrated fire management and policy actions across the UPRB and Pantanal, including safeguarding headwaters, expanding and effectively managing PAs, regulating fire use, and pre-positioning firefighting brigades ahead of dry seasons.

Wildfires in 2020 burned nearly a third of the Pantanal but disproportionately impacted jaguars: about 45% of the estimated population, 79% of tracked resident HRs, and extensive portions of PAs overlapping HRs. The study provides a 16-year perspective that situates 2020 as an extreme event, with fire extent and intensity far above historical medians. These findings emphasize that fire is now a major threat to jaguar conservation in the Pantanal and, by extension, to the region’s ecological integrity. Immediate priorities include: reducing anthropogenic drivers that amplify drought and fire, protecting springs and headwaters, expanding and adequately resourcing PAs, regulating fire use, and deploying permanent/seasonal fire brigades and rapid-response protocols. Future research should track long-term demographic and movement responses of jaguars to fire, monitor habitat recovery and GPP trajectories, refine real-time land-use and fire-risk monitoring, and evaluate the effectiveness of mitigation actions (e.g., PA expansion, restoration in the UPRB, conflict reduction strategies) for maintaining viable jaguar populations.

Key limitations include: (1) HR-based inferences are derived primarily from jaguars tracked in Brazil (45 of 48 individuals), with limited representation from Paraguay and Bolivia; (2) HRs of resident individuals were assumed to be stable priority areas over time, which may not hold under changing conditions; (3) annual estimates of the number of jaguars, HRs, and PAs within HRs were treated as constant in some summaries due to sample-size and data constraints; (4) the proxy for jaguars affected is based on overlaying fire pixels with density predictions and HR probability surfaces, not direct mortality or displacement counts; (5) MODIS active fire detections and the ≥325 K threshold may miss low-intensity or understory fires and have spatial resolution limits; (6) PA datasets may be incomplete, particularly for private protected areas; (7) potential uncertainties in land-cover, precipitation, and hydrological datasets used for contextual analyses.