Investigation acid mine drainage minerals using spectral characteristics and satellite images processing of Landsat- 8, a case study: Darrehzar mine, Kerman Province, Iran

Introduction:
Oxidation of iron sulfide is a common phenomenon in mining areas and sometimes generates acid drainage. Acid drainage is produced under certain conditions of pH and Eh when metallic sulphides (mostly iron) are exposed to oxygen and water. Copiapite (Fe2+Fe3+4(SO4)6(OH)2•20(H2O)), jarosite ((SO4)2KFe3(OH)6), schwertmannite (Fe3+16O16(OH)12(SO4)2), goethite (FeO(OH)), ferrihydrite (Fe5HO8.H2O), lepidocrosite (FeO(OH)), and hematite (Fe2O3) minerals are formed by oxidation of pyrite with increasing pH (from very acidic to neutral) and are frequently observed in mining areas with acid drainage pollution.
Landsat 8 can be used for environmental studies and detection of secondary iron minerals associated with acid drainage, as it has characteristics such as a relatively good temporal (16 days) and spectral resolutions in the visible and infrared (VNIR) ranges. This research focuses on the study secondary iron minerals associated with acid mine drainage using investigation of spectral characteristics of AMD minerals and image processing of Landsat 8 data in the Darrehzar mine.
Material and method:
Spectra from USGS spectral library and Zabcic (2008) were chosen for identification of spectral characteristics of secondary iron minerals associated with acid drainage such as copiapite, jarosite, schwertmannite, goethite, ferrihydrite, lepidocrosite, and hematite. Furthermore, the resampled spectra of these minerals based on bands’ centers of Landsat 8 were also investigated. Lepidocrosite, ferrihydrite, goethite, and schwertmannite have the same spectrum on the resampled spectrum of USGS spectal library. These minerals are created in acidic to near neutral conditions. Simillarly, jarosite and copiapite which are associated with very acidic conditions, have the same features in the visible range.
The Fast Line-of-sight Atmospheric Analysis of Spectral Hypercubes (FLAASH) correction applied on the images to compensate for the atmospheric effects. The NDVI masking was used to remove vegetation before performing principle component analysis technique. Then selected principal component analysis (PCAs) technique was used for image processing. The ferric iron absorption (Fe3 +) of band 1 at copiapite, jarosite, schwertmannite, goethite, ferrihydrite, lepidocrosite, and hematite (0.433-0.453 micrometers), absorption of band 7 (2.100-2.300 micrometers) for copiapite and jarosite due to SO4 and OH vibrations and reflectance values in band 4 (0.630-0.680 micrometers) for copiapite and jarosite and band 6 (1.560-1.660 micrometers) of Landsat 8 for jarosite, schwertmannite, goethite, ferrihydrite, lepidocrosite, and hematite considered to perform this technique.
According to the eigenvector loading and spectral characteristics of Schwertmannite, goethite, ferrihydrite, lepidocrosite, and hematite minerals, PC2 is appropriate for discrimination of these minerals as bright pixels due to the high positive loading in band 6 and high negative loading in band 1. Jarosite will be presented by dark pixels in PC3 because of high positive loading in band 7 and high negative loading in band 6 which are absorptive and reflective bands respectivelly. This PC was multiplied by -1 to convert pixels related to Jarosite as bright pixels. Copiapite and jarosite will be presented by bright pixels values in PC4 based on high positive loading in band 4 and high negative loading in band 1 which are reflective and absorptive bands respectivelly.
Eventually, field survey was conducted for evaluation and verification of results of image processing. The rocks and water samples were taken out of the areas identified by image processing. Five water samples were taken from different parts of the mining area that consist of influx water to the mine (Da31), Piezometer well in the mine (Da32), water pit located in the eastern part of mine (Da23), accumulated water in the western part of mine (Da20), and discharging water from the mine (Da24).
Water samples were taken by Polyethylene bottles and recorded their temperatures in situ and transferred to Graduate University of Advanced Technology laboratory for EC and pH measurements. Their pH and EC were measured by Metrohm 827 lab pH meter and Metrohm 712 Conductometer EC meter. Rock samples were taken from southern, western, and eastern dumps and in the central part of the mine based on image processing results and transferred to Graduate University of Advanced Technology laboratory for spectroscopy measurments. Spectroscopy of rock samples was conducted by ASD FieldSpec®3. The measured spectra were compared with the spectra in USGS spectral library.
Result:
Schwertmannite, goethite, ferrihydrite, lepidocrosite, and hematite that have been detected in PC2, were located around the mine and pixels having high values (marked by red colour) have been presented in the east and northwest mine (Fig 1C). Secondary iron minerals such as hematite and goethite were observed on the eastern and western dumps during the field survey. Spectroscopy results were presented goethite and hematite on the western dump and goethite on the eastern dump.
Jarosite which were detected as bright pixels in -PC3, cover the interior of mine and tailing dumps. Furthermore the bright pixels in –PC3 correspond to argillic and phyllic zones in Darrehzar mine. Pixels having high values (marked by red color) have been observed within and eastern parts of mine (Fig 1B). Copiapite and jarosite have been situated along the main stream in the middle part of the mine and southern dump (Fig 1A).
EC and pH measurements of water samples determined that the water samples taken from eastern part of mine (Da23) and accumulated water in the western part of mine (Da20) have high EC and low pH while influx and discharging water of the mine have low EC and neutral pH. Water sampling areas having high EC and low pH (Da23 and Da20) conform to copiapite and jarosite in PC4. Despite precence of water with low pH and high EC inside of the mine, discharging water of the mine has low EC and neutral pH. It shows that during the sampling time acidic water inside of mine has no influence on the discharging water because acidic water is preserved in a reservior inside the mine. However, acidic water may be cause pollution during high rainfall period and it should be consider at the management planning of the mine.
Conclusion:
Investigations spectral features of secondary iron minerals associated with acid mine drainage revealed that strong absorption of the minerals associated with very acidic environments such as copiapite (0.430µm) and jarosite (0.436µm) is located at lower wavelengths than strong absorption of the minerals associated with acidic to neutral environments such as schwertmannite (0.489 µm), goethite (0.480 µm), ferrihydrite (0.489 µm), lepidocrosite (0.480 µm) and hematite (0.480 µm). In addition, jarosite (associated with very acidic conditions) show strong absorption feature in short wave infrared (SWIR) region (2.26µm) due to SO4 and OH vibrations while the other minerals such as schwertmannite, goethite, ferrihydrite, lepidocrosite and hematite do not have any absorption features in this region. The results of spectral processing revealed that despite there are some similarity between schwertmannite, goethite, ferrihydrite, and lepidocrosite; it is possible to discriminate AMD minerals using spectroscopic studies due to large numbers of spectral channel in spectrometers.
The result of Landsat- 8 image processing showed that the OLI sensor in this satellite could identify secondary iron minerals associated with acid drainage and determine environments having different acidic conditions. However it could not separate hematite, goethite, lepidocrosite, ferrihydrite, and schwertmannite which have relatively the same spectra from each other and also copiapite and jarosite. Based on PCA results lepidocrosite, ferrihydrite, hematite, goethite, and schwertmannite that are associated with acidic to neutral conditions, were detected around mine while copiapite and jarosite that are generated in acidic condition (low pH), have been detected inside mine. Results of laboratory and field analysis conformed image processing results so that spectroscopy analysis on samples of eastern and western dumps revealed hematite and goethite which correspond to discriminated minerals like hematite, goethite, lepidocrosite, ferrihydrite and schwertmannite in PC2.
Acknowledgment:
The authors are sincerely grateful to the geologists and staff of the Sarcheshmeh copper mine especially Mr. Khosrowjerdi, Dr. Sahrayi, and Mr. Salajegheh for providing the facilities and kindly helping us during our field work.