Coal fires in Northwest China. Detection, monitoring, and prediction using remote sensing data

Coal fires in China occur within a region that stretches over 5,000 km in the east- western part and 750 km. in the north-southern part. These fires cause an economic and environmental threat by making a significant contribution to the global CO2 budget. The studies made in this thesis can be divided into two parts. Part one is based on field work and laboratory analysis that includes the dating of the paleo coal fires; part two concerns remote sensing applications for the active coal fires. In Chapter 2, the evolution of the paleo coal fires in Toutunhe and Xinjiang areas are studied. Several age groups of burnt rock have been recognized and their relationships with the river terraces will be discussed. The causes of the paleo coal fires are addressed, and the areas of coal fires with different ages have been dated. In Chapter 3, the physical basis of thermal infrared remote sensing for the detection and measurement of coal fires are addressed with an emphasis on the spatial, spectral, and radiometric resolution. In Chapter 4, a method to reduce the effect of solar heating, the main factor of confusion when investigating the thermal anomalies of coal fires, is discussed with the help of a DEM. In Chapter 5, as the coal fires normally occupy only part of one pixel of the Landsat TM thermal channel data, the capability of sub pixel coal fire detection is addressed. In Chapter 6, the airborne data from different wavelengths acquired at different times are studied to analyze the spatial thermal characteristics of the coal fires. Spreading direction and different types of coal fires are studied. Chapter 7 presents, based on multi-sensor data fusion techniques, a hierarchical methodology for detection and monitoring of the coal fires. 120 refs