Multispectral remote sensing satellite imagery has become a cost-effective and accessible tool for the preliminary identification of hydrothermal alteration zones and mineral exploration. In this study, the capability of the thermal infrared (TIR) bands of the ASTER sensor was evaluated for mapping siliceous and carbonate outcrops associated with low-sulfidation epithermal gold mineralization in the Kahyaz prospect, northwest of Nain, within the Urumieh–Dokhtar Magmatic Arc (UDMA), Iran. To achieve this objective, band ratios, Principal Component Analysis (PCA), False Color Composites (FCC), and Decorrelation Stretch (DS) were applied to enhance siliceous and carbonate units at the regional scale. Subsequently, quartz and calcite distribution maps were generated using the Spectral Angle Mapper (SAM) and Matched Filtering (MF) classification algorithms and validated through field investigations, petrographic observations, and X-ray diffraction (XRD) analyses. The results indicated that the second principal component (PC2) showed the highest correspondence with the 14/12 band ratio and effectively enhanced the spectral contrast between siliceous and carbonate units. In addition, the SAM and MF algorithms successfully delineated the spatial distribution of quartz and calcite with high classification accuracy. The overall accuracy ranged from 89% to 97.5% for the MF method and from 95% to 98% for the SAM method, while the kappa coefficient ranged from 0.6 to 0.7. The results demonstrate that, despite the spatial resolution limitations of the ASTER TIR bands, they provide a reliable means for mapping siliceous and carbonate outcrops associated with low-sulfidation gold mineralization when combined with appropriate image processing techniques and field validation. Therefore, ASTER TIR data can serve as an effective tool for reconnaissance-stage mineral exploration.
 
Introduction
The economic deposits such as gold and copper usually are related to the hydrothermal altered rocks (John et al., 2010). Recognition of the different types of alteration, as an evident signature of probable mineralization, is important during the ore mineral exploration (Dilles and Einaudi, 1992). Detection the hydrothermal alteration zones by using the remote sensing data effectively and widely has used for the ore deposit prospecting, including the epithermal gold deposits (Bolouki et al., 2020). Remote sensing data are useable for the recognition and detection of the gold-silver epithermal deposits exposures, because the host rocks of this type of deposits have the mineral assemblages, which show the diagnostic absorption features at visible-near infrared (VNIR), short wave infrared (SWIR), and thermal infrared (TIR) regions of the electromagnetic spectrum (John et al., 2018).  
Kahyaz prospect is located in the central part of the Urumieh-Dokhtar Magmatic Arc (UDMA). It is a low sulfidation epithermal gold deposit, which characterized with colloform and crustiform quartz veins, quartz-calcite-adularia alteration minerals, low content of sulfides, high content of silver (Electrum).
 
Materials and Methods
In this study used the thermal bands (TIR) of an ASTER L1T scene
AST_L1T_00309192002072139_20150425042239_103813), which it covered the Kahyaz prospect and several kilometers of its around. This image scene, which is free cloud and vegetation, acquired on September 19, 2002. Data processing did by Envi version 5.6 software. During the data pre-processing, at first the IARR atmospheric correction was done, and then resampled to 15 m resolution. Afterward, the methods such as band ration, false color composite, decorrelation stretch, matched filtering, and spectral angle mapper used for image processing. 
 
Results
14/12 band ratio was used to map the quartz-rich area. In the resulted image, quartz bearing area are mainly white in color.  The carbonates-rich area was detected by 13/14 band ratio, and showed with white in color. These minerals are predominantly found in the Quaternary alluvial terraces, and also in the sandstone and conglomerate associated to tuffaceous volcanic rocks, andesitic lava and diabase.  FCC method was used to quartz and carbonate detection. ASTER bands of 14, 12, and 10 set in R (Red), G (Green), B (Blue) channels, respectively. The rocks containing the quartz are mainly pale-pink and dark in color, while carbonate-bearing rocks exhibit white color.  Based on MF and SAM methods, quartz coincides with the alluvial unit, dacitic-rhyodacitic rocks, andesi-basalt and andesitic lava. Carbonate overlaps with the quartz in most place and predominantly associated with the Quaternary sediment unit, and also andesitic rocks, and subvolcanic dacitic-rhyodacitic rocks.
 
Discussion
Using RGB band ratios of 12/14, 14/13, and 10/12, siliceous zones were identified in pink, orange, and yellow, while carbonate areas appeared in green color. The SAM and MF algorithms effectively revealed the distribution of lithological units and outcrops associated with quartz and carbonate. For example, the accuracy of the SAM method for siliceous and carbonate samples was 99.10% and 98.98%, respectively. The calcite mineral map derived from the MF algorithm is mostly consistent with Quaternary sedimentary units, whereas the map produced by the SAM algorithm is mainly associated with andesitic volcanic, pyroclastic, and dacitic–rhyodacitic subvolcanic units. The quartz mineral map obtained from the MF algorithm generally corresponds to dacitic–rhyodacitic, andesitic, and Quaternary sedimentary deposits, while the quartz mineral map derived from the SAM algorithm is related to dacitic–rhyodacitic subvolcanic units, andesitic units, tuff, sandstone, and conglomerate. Based on the validation carried out, there is good agreement among the remote sensing results, petrographic studies, and XRD analysis. The main mineralized vein in the Kahyaz prospect, composed of quartz–calcite, overlaps with the silica and carbonate signatures identified in the TIR band images. The results obtained from the SAM and MF methods are consistent with the XRD analysis performed on most pure calcite and quartz samples, with kappa coefficient values ranging from 0.6 to 0.7. These results, in turn, demonstrate the reliability of the remote sensing-derived maps in this study.