HESFIRE: an explicit fire model for projections in the coupled Human–Earth System
Abstract. Vegetation fires are a major driver of ecosystem dynamics and greenhouse gas emissions. Potential changes in fire activity under future climate and land use scenarios thus have important consequences for human and natural systems. Anticipating these consequences relies first on a realistic model of fire activity (e.g. fire incidence and inter-annual variability) and second on a model accounting for fire impacts (e.g. mortality and emissions). Key opportunities remain to develop the capabilities of fire activity models, which include quantifying the influence of poorly understood fire drivers, modeling the occurrence of large, multi-day fires – which have major impacts – and evaluating the fire driving assumptions and parameterization with observation data. Here, we describe a fire model, HESFIRE, which integrates the influence of weather, vegetation characteristics, and human activities in a standalone framework, with a particular emphasis on keeping model assumptions consistent with fire ecology, such as allowing fires to spread over consecutive days. A subset of the model parameters was calibrated through an optimization procedure using observational data to enhance our understanding of regional drivers of fire activity and improve the performance of the model on a global scale. Modeled fire activity showed reasonable agreement with observations of burned area, fire seasonality and inter-annual variability in many regions, including for spatial and temporal domains not included in the optimization procedure. Significant discrepancies – most notably regarding fires in boreal regions, in xeric ecosystems, and fire size distribution – are investigated to propose model development strategies. We highlight the capabilities of HESFIRE and its optimization procedure to analyze the sensitivity of fire activity, and to provide fire projections in the coupled Human–Earth System at regional and global scale. These capabilities and their detailed evaluation also provide a solid foundation for integration within a vegetation model to represent fire impacts on vegetation dynamics and emissions.