Global seasonality of small-scale livelihood fire

Landscape fires are increasingly represented in dynamic global vegetation models to understand impacts on carbon emissions and climate. Deliberate human fire use and management influence landscape fire characteristics, varying in space and time depending on social, economic, and ecological factors. For example, fire is used variously in rural livelihoods involving e.g., agriculture, hunting, gathering, and for other cultural practices, often depending on the time of year. Yet existing global fire models typically represent human fire use as a constant function of gridded datasets such as population density or gross domestic product.

Recently, initiatives have begun to draw together available data on global fire use from across multiple disciplines and disparate sources into coherent databases. We draw on information from one of these databases, the Livelihood Fire Database (LIFE), which includes case studies in 587 locations worldwide, to assess the availability of data on seasonality of anthropogenic fires associated with small-scale rural livelihoods. By defining seasonal cycles relative to the local variation of precipitation and evapotranspiration at each case study location, we look for patterns in the spatiotemporal nature of anthropogenic fires associated with different fire-use purposes - such as clearing vegetation for agriculture, maintaining pasture for livestock, or driving game when hunting - and consider the potential for this analysis to inform fire models.

For many fire types, especially those related to hunting, gathering, human wellbeing, and social signalling, there are limited quantitative data available, but it is possible to draw qualitative insights from case studies. Where quantitative data are available, we find some correspondence between fire seasonality and the intended fire-use purpose, suggesting that distinguishing between distinct fire-use purposes could improve the representation of human fire use in fire models, and consequently the seasonal cycle of fire emissions. Case studies demonstrate that environmental and social conditions drive variation in fire use for the same purpose, reiterating that a wide range of factors influence human behaviour and that assumptions of uniform drivers of anthropogenic fire may be misleading. Many of the fires now being revealed in global burned area data by new fine-scale remote sensing products are likely human-set; continued collection, collation, and analyses of data on human fire use globally is important to ensure appropriate anthropogenic representation in fire models.