Enhancement of Fire Early Warning System in Vietnam Using Spatial Data and Assimilation

Accurate and timely information on vegetation fires is crucial for mitigation and rehabilitation measures. With the advent of spatial technologies, fire risk can be mapped at varied spatial scales integrating multiple datasets. In Vietnam, forest protection department (FPD) leads the forest and fire management activities. FPD routinely generates fire early warning maps at a district level that depict fire risk varying from level I to level V with increasing severity. The FPD fire risk maps are based on an algorithm that only uses ground-based meteorological inputs. In this study, we improve the fire risk assessment through assimilating meteorological as well as satellite data and map the fire risk at 0.1 × 0.1° grid cells. We use MODIS active fires to test the relative accuracy of FPD-generated fire risk map and our approach. Results suggest a significant enhancement in fire risk using our approach. Our results outperformed the FPD results in terms of both spatial details and fire risk information, i.e., we found a much higher fire density at level IV and level V at 0.1 × 0.1° grid scale than the FPD district-level maps. Our results highlight the potential of data assimilation for an improved fire early warning in Vietnam.