The spatial and temporal distribution of industrial/biomass fire points during the decade 2012–2021 in China

Biomass and industrial fire points are crucial to forest fire prevention and industrial carbon emissions research. However, only a few investigations in the literature focus on fire point classification tasks using different spatial resolutions fire datasets in China. A comprehensive and accurate analysis of the spatial and temporal distribution characteristics and trends of fire points in China during the past decade is lacking. In this study, industrial heat source data, four fire point data, including VIIRS 750-m Nightfire (VNF), VIIRS 375-m Active Fire points (ACF), MODIS 1000-m fire points (MF) and Landsat-8 30 m fires data (LF), and land cover type data were integrated to more accurately classify fire points into biomass or industrial fire points. The result shows that our classification is more accurate than ACF’s results, and several conclusions were obtained from the spatial and temporal analysis of the total/biomass/industrial fire points across the four fire point datasets in China. (1) There was a high spatial and temporal correlation across all four datasets in China between 2012 and 2021. The annual numbers of fire points from the four fire point datasets all increased from 2012 to 2013, peaked in 2014, and have since declined. (2) The number of industrial fire points across the four datasets was static and persistent over the time series and had tight spatial aggregation, with little variation in their annual numbers and spatial distribution over time. In contrast, biomass fire points exhibited more significant changes in their spatial distribution, and the annual number declined after 2014. (3) The distribution of biomass fire points shifted northward over time, gradually moving from the Yangtze-Huai Plain and Yunnan Province in 2012 to northeastern China after 2018. These findings highlight the importance of considering temporal factors when analyzing fire point data, as well as the potential benefits of utilizing multiple datasets to achieve more accurate results.