Publikationsdatum:
2020-02-12
Beschreibung:
Within a few years after the German reunification, most open pit lignite mines in the Central German Lignite Mining District have been shut down leaving more than 200km2 of devastated land in Saxony (Germany) even now. These areas are in demand of short term conceptions for future landuse and suitable recultivation methods. This thesis investigates, how thermal infrared (TIR) remote sensing in particular can provide valuable information, which is required to assist the landuse planning in identifying and classifying critical areas. In order to establish the mining of the lignitic seams the overburden Quatemary and Tertiary sediments had been removed and dumped in conveyor bridge dumps. The Tertiary sediments contain the mineral pyrite, which reacts to sulfuric acid under wet oxidizing conditions. The acidification of the dumps is not only strongly affecting the chemical characteristics of groundwater and residual lakes but also the stability of slopes leaving the risk of slides and subsidences. For a risk assessment of the aff~cted areas the geochemical and mineralogical composition as well as compositional changes of the dumps have to be explored. This is giving strong motivation for the application of remote sensing methods to answer the short term demands of landuse planners. Hyperspectral scanner data were acquired over two open pits (Espenhain, Zwenkau). The used scanner is DLR's Digital Airbome Imaging Spectrometer (DAIS 7915) which records data in 79 bands from the visible to thermal infrared. To allow a quantitative approach by means of empirical methods, the airbome data had to undergo a number of correction procedures and accompanying ground truth measurements had to be carried out. For a quantitative analysis of the target area, the data were calibrated and atmospherically corrected. Atmospheric modeling is based on the software packages SENSAT and ATCOR using the MODTRAN code. Interna! distortions of the imagery were corrected from flight attitude information and the data were finally geocoded to Gauß-Krüger-coordinates. A total of 77 field sarnples were taken to develop quantification algorithms based on spectral characteristics. Chemical and mineralogical analyses (GFZ Potsdam) as well as thermal infrared laboratory spectrometry (2.5-14μm) were performed. The samples were found tobe mainly composed of quartz (50-90 weigth percent), alkaline feldspars (5-20 wt.% ), kaolinite (0-15 wt.%), illite (2-20 wt.%) and pyrite (0-4 wt.%). In the TIR, many rock-forming minerals (e.g. silicates) show characteristic spectral features. To investigate the influence of different grain sizes on the spectral behavior, the particulate material was ground to grain sizes 〈63μm. Both, the fine homogenized and the original dump sediments were measured. The mineralogical quantitative sarnple composition was correlated with the spectral features measured in the laboratory. Different approaches were investigated to quantify the rnineralogical composition including absorption band depths and spectral angle mapping (SAM). The analysis in the laboratory allowed a determination of kaolinite and quartz contents of the overburden dumps. By transferring the regressions obtained from laboratory spectroscopy to atmospherically corrected and calibrated DAIS 7915 data, quartz contents were successfully mapped. Due to the broad TIR bands of DAIS, the spectral features of kaolinite were not resolved in the airbome data. A temperature map derived from night-time DAIS thermal infrared data were evaluated with respect to exothermic pyrite oxidation, too. The fresh Tertiary dumps in Zwenkau show thermal anomalies which seem to be related to exothermic oxidation processes. As a final outlook, the potential of the ASTER sensor, which is planned to be launched in early 1999, was evaluated. The TIR bands of ASTER are likely tobe a powerful tool for the mapping of quartz contents of particulate material. However, the low geometric resolution in the thermal infrared (90m pixel size) causes limitations for a structural analysis. Therefore, the Multi-Sensor-Multi-Resolution Technique (MMT) was used to unmix the low resolution bands using the higher resolution reflective ASTER bands (15-30m pixel size). The results show a significant improvement of the images. The MMT algorithm allows a further spectral investigation of the unmixed TIR data. The results presented in this work show that remote sensing data can provide useful information for a quantitative assessment and monitoring of environmental impacts. The combined use of the wavelength region from the VIS to TIR produces characteristic features that allow a differentiation of the mineralogical components, which could not have been achieved with either data set separately.
Schlagwort(e):
550 - Earth sciences
Sprache:
Deutsch
Materialart:
info:eu-repo/semantics/doctoralThesis
Format:
application/pdf
Permalink
