Identifying local‐scale meteorological conditions favorable to large fires in Brazil

Abstract

This study aims to investigate local-scale meteorological conditions associated with large fires in Brazil during recent decades. We assess whether there are large fire types with preceding predictors. Our results show that large fires, defined with a threshold of a daily burned area >95th percentile of the historical record, mainly occur in August and September in Brazil, and Amazônia and Cerrado experience much higher numbers of large fires than the other biomes. There are two large fire types that have robust meteorological signatures: (1) a wind driven type, characterized by peak wind speed on the day of the fire, and anomalously high wind speed a few (∼3) days before and after the fire; and (2) a Hot-Drought driven type, characterized by anomalously high temperature, low relative humidity, and consistent drought conditions indicated by anomalously high fuel aridity starting as far back as 5 months prior to the fires. A third one is characterized by no anomalous meteorological conditions. The wind driven type most frequently occurs in southern and southeastern Amazônia, Pantanal, and western and northern-to-central Cerrado, with some occurrences over the western Caatinga region bordering Cerrado, southern Cerrado, and southern Mata Atlântica; whereas the Hot-Drought driven type most frequently occurs in southern and southeastern Amazônia, Pantanal and western and northern-to-central Cerrado, with some occurrences over the western Caatinga region bordering Cerrado, southern Cerrado, central-tosouthern Mata Atlântica, and a few occurrences over Northern Brazil where the Amazônia meets Roraima. Southern and southeastern Amazônia, Pantanal and western and northern-to-central Cerrado are the major large fire prone regions. Our results highlight that understanding the temporal and spatial variability of the meteorological conditions associated with large fires is essential for developing spatially explicit forecasting, and future projections of large fire hazards under climate change in Brazil, in particular the Hot-Drought driven type.

Amazon fires in the 21st century: The year of 2020 in evidence

“…In line with this finding, Li et al (2022) indicated that drought-driven large fires accounted for <30% of the total large fires registered in the Brazilian Amazon from 2001 to 2015.…”

“…In other words, the majority of anomalously high fire occurrences in the Amazon since 2003 may not have been influenced by droughts. In line with this finding, Li et al (2022) indicated that drought‐driven large fires accounted for <30% of the total large fires registered in the Brazilian Amazon from 2001 to 2015.…”

Drought–heatwave nexus in Brazil and related impacts on health and fires: A comprehensive review

“…195 Fires in Cerrado have been shown to be linked with meteorological conditions, in particular rainfall and temperature. 196,197 For instance, the Daily Severity Rating (DSR in short, an extension of the Canadian Forest Fire Weather Index System) explains 71% of the interannual variability of burned area in Cerrado. 198 Here, we show further evidence for the Cerrado's fire seasons (defined here as August-October) from 2001 to 2019 of a link between CDHW conditions with fire activity over four Cerrado ecoregions (Figure 5): Bico do Papagaio, Araguaia Tocantins, Bananal, and Alto Parnaíba.…”

“…This fire‐prone biome sees high fire activity every year as the largest contributor to Brazil's annual burned area and a major fire hotspot worldwide 195 . Fires in Cerrado have been shown to be linked with meteorological conditions, in particular rainfall and temperature 196,197 . For instance, the Daily Severity Rating (DSR in short, an extension of the Canadian Forest Fire Weather Index System) explains 71% of the interannual variability of burned area in Cerrado 198 .…”

The climate science for service partnership Brazil

“…Thus, developments in rainfall modelling and predictability can pull through to actionable advice. Li et al (2021) show how predictability of events such as fires may be improved using observed links to meteorological conditions which themselves can be predicted better than the impacts directly. Example services supported by CSSP Brazil include early warning of the annual maximum river height in Manaus (Chevuturi et al, 2021) and seasonal warnings of fire risk in Brazil's Protected Areas (Anderson et al, 2021)-an activity which has now produced multiple warnings and is transitioning to routine production.…”