Premium
Comparison of the Hazard Mapping System (HMS) fire product to ground‐based fire records in Georgia, USA
Author(s) -
Hu Xuefei,
Yu Chao,
Tian Di,
Ruminski Mark,
Robertson Kevin,
Waller Lance A.,
Liu Yang
Publication year - 2016
Publication title -
journal of geophysical research: atmospheres
Language(s) - English
Resource type - Journals
eISSN - 2169-8996
pISSN - 2169-897X
DOI - 10.1002/2015jd024448
Subject(s) - environmental science , fire detection , land cover , satellite , remote sensing , meteorology , physical geography , geography , land use , ecology , engineering , architectural engineering , biology , aerospace engineering
Biomass burning has a significant and adverse impact on air quality, climate change, and various ecosystems. The Hazard Mapping System (HMS) detects fires using data from multiple satellite sensors in order to maximize its fire detection rate. However, to date, the detection rate of the HMS fire product for small fires has not been well studied, especially using ground‐based fire records. This paper utilizes the 2011 fire information compiled from ground observations and burn authorizations in Georgia to assess the comprehensiveness of the HMS active fire product. The results show that detection rates of the hybrid HMS increase substantially by integrating multiple satellite instruments. The detection rate increases dramatically from 3% to 80% with an increase in fire size from less than 0.02 km 2 to larger than 2 km 2 , resulting in detection of approximately 12% of all recorded fires which represent approximately 57% of the total area burned. The spatial pattern of detection rates reveals that grid cells with high detection rates are generally located in areas where large fires occur frequently. The seasonal analysis shows that overall detection rates in winter and spring (12% and 13%, respectively) are higher than those in summer and fall (3% and 6%, respectively), mainly because of higher percentages of large fires (>0.19 km 2 ) that occurred in winter and spring. The land cover analysis shows that detection rates are 2–7 percentage points higher in land cover types that are prone to large fires such as forestland and shrub land.