Coal fire burning around the world is an environmental catastrophe characterized by the emission of noxious gases, particulate matter, and condensation by-products. In this study, coal fire temperature is retrieved based on Landsat 5 TM images and Generalized Single-Channel Algorithm (GSCA), in Wuda coalfield, Inner Mongolia, China. Then coal fire zones are extracted by Jenks′ natural breaks and threshold methods based on temperature images. Changes of coal fire zones are analyzed from 1989 to 2008. The results are summarized as follows: 1) The coal fire temperature retrieval method based on Landsat 5 TM and the GSCA model is effective and feasible, because the temperature error is relatively small (from –2.9℃ to +2.6℃) between the measured temperature and the retrieved temperature. 2) The accuracy is relatively high to extract coal fire zones through the Jenks′ natural breaks and threshold methods, because 83.56% of surveyed area is located in the coal fire zones extracted in 2005. 3) The coal fire area increased 9.81 × 10 5 m 2 from 1989 to 2005, and the annual growth is about 6.1 × 10 4 m 2 , with an annual increasing rate of 2.48%. The area of coal fire decreased by 8.1 × 10 5 m 2 from 2005 to 2008.
展开▼
