Assessment of pollutant load in Tharsis mine (Huelva, Spain): A study of evaporitic salts using UAS-Based Hyperspectral

One of the major environmental problems caused by mining operations is water pollution. In sulfide mining, pyrite is exposed to atmospheric conditions causing oxidize and release acidity and metals. The water transports these contaminants, producing a leachate known as Acid Mine Drainage (AMD). The Iberian Pyritic Belt (IBP), which belongs the province of Huelva, is very rich in massive sulfide deposits that are mainly composed of pyrite. Mining activity has left numerous abandoned mines with enormous amounts of waste rich in sulfides, including Tharsis mines. Numerous acid leachates emerge from the waste from the Tharsis mines, which have not been active since 2001, and drain to rivers of the area. In the dry season, pollutants precipitate in the form of soluble evaporitic salts that are redissolved with the first important rains of autumn. This redissolution of salts generates a pH decrease and the release of high amounts of contaminants such as Fe, Al, Zn, Mn, As, SO₄, Cu, Ni, etc. Remote sensors provide a cost-effective, consistent and accurate approach to monitoring mining pollution. Multispectral and hyperspectral sensors have been widely used due to the distinctive spectral absorption characteristics of minerals. The emerging use of unmanned aerial systems (UAS), such as multicopters coupled with hyperspectral sensors, has become a tool for collecting data at a higher spatial resolution than most aircraft and satellites, resulting in greater accuracy.

To obtain the aerial images, a DJI Matrice 600 Pro octacopter UAS was used with a Headwall Hyperspectral (HS) Coaligned VNIR-SWIR sensor with visible and near infrared range (VNIR; 400-1000 nm) and shortwave infrared range (SWIR; 900-2500 nm) and equipped with a LiDAR to quickly obtain an DSM for georeferencing of the hypercubes. The data obtained have been preprocessed to obtain an orthomosaic with the VNIR and SWIR spectra. The data are processed with ENVI v 5.3.6 and QGIS v 3.26.3 software. Wavelengths from 2001 nm to 2450 nm are selected to identify the minerals that are present in the study area, resulting in an orthomosaic with 76 exploitable bands. MNF and PPI techniques are applied to this orthomosaic to obtain the spectral signatures of the study area. These spectral signatures are compared to the USGS mineral library to identify the minerals present. In this way, 8 minerals are identified: hematite, goethite, jarosite, epsomite, copiapite, illite, clinochlore and chlorite + muscovite. Afterwards, the pixels in the study area are classified and the location of the minerals is obtained. Jarosite, epsomite and copiapite are the most abundant minerals and mainly present in the surroundings of waste leachates. The first rains after summer, evaporitic salts and Fe oxides, hydroxides and hydrosulfates are redissolve and cause a considerable polluting load increase. The surface area occupied by these minerals is 3,861 m², which represents a high load of pollutants that reach the rivers in the area.