Global warming is shifting the relationships between fire weather and realized fire-induced CO2 emissions in Europe

Abstract

Fire activity has significantly changed in Europe over the last decades (1980-2020s), with the emergence of summers attaining unprecedented fire prone weather conditions. Here we report a significant shift in the non-stationary relationship linking fire weather conditions and fire intensity measured in terms of CO₂ emissions released during biomass burning across a latitudinal gradient of European IPCC regions. The reported trends indicate that global warming is possibly inducing an incipient change on regional fire dynamics towards increased fire impacts in Europe, suggesting that emerging risks posed by exceptional fire-weather danger conditions may progressively exceed current wildfire suppression capabilities in the next decades and impact forest carbon sinks.

Figure 1

Fire weather dynamics in European IPCC regions. (a) A map summarising the distribution of the IPCC regions analysed. The map was produced using QGIS v3.16 (see “Methods”). (b) Observed trends in the Seasonal Severity Rating index (SSR) in Southern, Central and Northern Europe over 1980–2019. Colored shaded areas highlight years characterized by higher SSR values over the analyzed period. Ordinary least squares fits are indicated.

Figure 2

Observed changes in the non-stationary relationships between fire weather danger (SSR) and fire impacts (CO₂ emissions, MtC) in Southern Europe. (a) Changes in the explained variation (R²) observed in moving-window correlation analyses over the last two decades (2000–2020). Asterisks in panel a (*) indicate significant correlationships, observed only in the last decade. (b) Observed relationship between Seasonal Severity Rating index (SSR) and satellite-estimated fire-induced CO₂ emissions in the last decade. (c) Projected increase in the 90th percentile of SSR during the fire season predicted for the 2079–2098 time period under RCP 4.5. SSR increases are represented relative to the values estimated for the reference period (1986–2005, Fig. S4). (d) Projected increase in the 90th percentile of SSR under RCP 8.5. Maps were produced using Python Programming Language version 3.8 (see “Methods”).