Occurrence frequencies and regional variations in Visible Infrared Imaging Radiometer Suite (VIIRS) global active fires.

Abstract

Active fires are considered to be the key contributor to, and critical consequence of, climate change. Quantifying the occurrence frequency and regional variations in global active fires is significant for assessing carbon cycling, atmospheric chemistry, and postfire ecological effects. Multiscale variations in fire occurrence frequencies have still never been fully investigated despite free access to global active fire products. We analyzed the occurrence frequencies of Visible Infrared Imaging Radiometer Suite (VIIRS) active fires at national, pan-regional (tropics and extratropics) to global scales and at hourly, monthly, and annual scales during 2012-2017. The results revealed that the accumulated occurrence frequencies of VIIRS global active fires were up to 12,193×104 , yet exhibiting slight fluctuations annually and with respect to the 2014-2016 El Niño event, especially during 2015. About 35.52% of VIIRS active fires occurred from July to September, particularly in August (13.06%), and typically between 10:00 and 13:00 Greenwich Mean Time (GMT; 42.96%) and especially at 11:00 GMT (17.65%). The total counts conform to a bimodal pattern with peaks in 5°-11°N (18.01%) and 5°-18°S (32.46%), respectively, alongside a unimodal distribution in terms of longitudes between 15°E and 30°E (32.34%). Tropical annual average of active fire (1,496.81×104 ) accounted for 75.83%. Nearly 30% were counted in Brazil, the Democratic Republic of the Congo, Indonesia, and Mainland Southeast Asia (MSEA). Fires typically occurred between June (or August) and October (or November) with far below-average rainfall in these countries, while those in MSEA primarily occurred between February and April during the dry season. They were primarily observed between 00:00 and 02:00 GMT, between 12:00 and 14:00 within each Zone Time. We believed that VIIRS global active fires products are useful for developing fire detection algorithms, discriminating occurrence types and ignition causes via correlation analyses with physical geographic elements, and assessment of their potential impacts.