Talk: Why do emissions estimates differ, and what can we learn from the differences?
Abstract: Active fire detections are often used in atmospheric applications because of the higher repeat rate and temporal coverage they can provide relative to methods based on detection of burned areas. Decades of research with these observations have shown them to relate quantitatively to fire activity and to emissions, but the quantitative response is seen to vary dramatically between sensors and among different ecosystems. This talk will describe some of the key components of this variation, and demonstrate the framework of a model of fire detection efficiency that permits quantitative normalization of fire detections from multiple sensors, and improved description of spatial and temporal patterns of fire activity from individual sensors. A key focus of this talk will be the interaction between the sampling pattern of the MODIS sensor and the diurnal variation in fire activity. This talk will lead into a discussion of scale issues associated with the use of satellite-based emissions inventories in atmospheric models.