Thermal Analysis in the field

The figure shows a photograph taken with a thermal camera. The underground coal fire that can only be indirectly observed by cracks and fissures in the bed rock can clearly be recognized in the thermal image. Several hot cracks and vents were closely monitored and are thus comparable on a multi-temporal basis. Thermal cameras are a good tool to determine if extinguished coal fires stay extinct, and how the thermal intensity of a fire develops over time.

Thermal image over ordinary photograph. Note: The geometry of anomalies is related to the crack pattern

All 17 coal fires in the Wuda syncline were mapped during field campaigns in 2003, 2003 and 2004. These field mapping campaigns act as ground truth for thermal satellite data analysis.

Close up of photograph and thermal image of a crack above a coal fire.

In the course of the project methods were developed to retrieve coal fire temperatures and coal fire related energy releases from thermal anomaly clusters. Hopefully in the future these methods can estimate the amount of coal burnt with in a subsurface fire.

Thermal Analysis Using Spaceborne Data

Thermal infrared satellite data are used for detection and quantitative analysis of the surface and near-surface coal fires. Such fires usually show enhanced brightness temperatures in the thermal part of the electromagnetic spectrum. Depending on the observing system there are usually one, two or more spectral bands available to detect coal fires.

Within this project we use imagery of the following satellite sensors: ETM, BIRD, ASTER and MODIS. The operational MODIS sensor has several thermal infrared channels at a pixel spacing of 1km, the DLR experimental satellite BIRD has one band in the mid-infrared and one band in the thermal infrared channel and has a pixel spacing of 185m. ASTER spans the 8-12 micron region with five contiguous bands at 90m pixel spacing. ETM has only one thermal band between 10.4 and 12.5 microns at a spatial sampling interval of 60m.

Quantitative analysis of coal fires in the Wuda coal field.

There are different methods to thermally assess the coal fires. In order to detect the thermal anomalies a simple theshold at a certain brightness temperature can be used to delineate the fires, however this often lead to over or underestimation of the fire area and it can not be applied over large areas. Therefore new statistical methods are being developed within this project, which are suitable for large area thermal anomaly detection.

If two thermal channels, picking up the coal fire signals are available, a quantitative assessment of the effective energy released from a coal fire can be estimated (Dozier, 1981). However, with ETM having only one thermal channel, this method is not applicable, other techniques are being developed.

Results show that BIRD, ASTER and ETM imagery have strong potential for coal fire detection and analysis. While BIRD data allow the detection and analysis of hot surface coal fires, ASTER and ETM imagery can be used to also map large area subsurface coal fires. For all these sensors it can be said that night time data is more applicable for coal fire mapping due to smaller influence of the uneven solar heating effects.