Detection, Evaluation, and Analysis of Global Fire Activity Using MODIS Data

Abstract

Global biomass burning plays a significant role in regional and global climate change, and spaceborne sensors offer the only practical way to monitor fire activity at these scales. This dissertation primarily concerns the development, evaluation, and use of the NASA Terra and Aqua MODIS instruments for fire monitoring. MODIS is the first satellite sensor designed specifically for global monitoring of fires. An improved operational fire detection algorithm was developed for the MODIS instrument. This algorithm offers a sensitivity to small, cool fires and minimizes false alarm rates. To support the accuracy assessment of the MODIS global fire product, an operational fire detection algorithm was developed and evaluated for the ASTER instrument, which provides higher resolution observations coincident with the Terra MODIS. The unique data set of multi-year MODIS day and night fire observations was used to analyze the global distribution of biomass burning using five different temporal metrics which included, for the first time, mean fire radiative power, a measure of fire intensity. The metrics show the planetary extent, seasonality, and interannual variability of fire. Recognizing differences in fire activity between morning and afternoon overpasses, the impact of the diurnal cycle of fire activity was addressed using seven years of fire data from the VIRS sensor on-board the TRMM satellite. A strong diurnal cycle was found in all regions, with the time of peak burning varying between approximately 13:00 and 18:30 local time. Given interest in area burned among atmospheric chemical transport and carbon cycle modelers, a data set was developed utilizing the MODIS global fire and vegetation cover products to estimate monthly burned area at 1-degree spatial resolution. The methods, products and results presented in this thesis provide the global change research and fire management communities with products for global fire monitoring and are currently being used in the development of the next generation of operational satellite fire monitoring systems.