A bottom-up savanna fire fuel consumption inventory and its application to savanna burning in Kafue National Park, Zambia

Background: Tropical savannas are the most extensively and frequently burned biome worldwide. To establish accurate emissions inventories for burning in tropical savannas, detailed biomass information is required. Available pan-tropical or global biomass maps currently focus on standing vegetation and largely ignore surface layers, a key component of fuel consumption in the tropics. Aims: In this paper, we propose a methodology for building a high-resolution regional bottom-up fuel inventory, and examine the effectiveness thereof in a local scale case study in Kafue National Park, Zambia. Methods: We scaled up fuel measurements using drone-mounted cameras and Sentinel-2 imagery. We examined inter-annual fire variability's effects on emissions. Key results: The fuel model performs well for surface level fuel, with an error margin of ±27%. Accuracy is reduced when mapping more stochastic fuel layers such as coarse woody debris, or fuel layers with a structural component. Conclusions: Current pyrogenic emissions models underestimate emissions from Kafue National Park. Implications: Timing of burning is an important factor for total burned area as well as for emissions.