Forest Fuel Parameterization and Large-Scale Fire Simulation Modeling Using High- Performance Computing in the Himalayas

Abstract. Large-scale forest fire modeling necessitates high-resolution spatial data and complex fuel parameterization, resulting in substantial computational demands. While significant progress has been made in countries such as the USA, Australia, and parts of Europe, particularly in region-specific fuel parameter development. Comparable studies for Indian forests remain limited. This research addresses that gap through extensive field investigations in the Sikkim Himalayas, leading to the generation of localized fuel parameters and their integration into high-resolution (1 m) fire simulations. We employed the Fire Dynamics Simulator (FDS), a Computational Fluid Dynamics (CFD)-based tool, to model forest fire behavior, executing simulations on a High-Performance Computing (HPC) platform. Pre-processing and formatting of input data were facilitated using open-source GIS software, ensuring compatibility with FDS requirements. Validation was performed by comparing simulated outputs with burnt area extents derived from post-fire satellite imagery. By integrating FDS, HPC and field investigations specific to Sikkim forests, this approach offers new insights into forest fire simulation tailored to the Sikkim context. Leveraging parallel computing enables faster simulation outputs, providing valuable support to decision-makers for more effective fire management and mitigation strategies.