Identification and Extraction of Faults Based on RS & GIS and Geomorphic Evidences of their Activity in the Qoshadagh Fault Zone, NW of Iran

Introduction
Similar to other parts of Iran, Azerbaijan Plateau is active tectonically due to the Eurasia-Arabia cluster convergence, and its hydrogeomorphological impacts are evident as a strike-slip fault, folding, horst, graben, changes in the topography, and drainage system (Abedini, 2016: 75). Therefore, investigating and understanding the function of faults can significantly contribute to analyzing the tectonic activities, the occurrence of earthquakes, their geomorphic hazards as well as the environmental planning and management, the adaptation of the land use, and constructions regarding the tectonic situation of the regions and reduce their hazards.
The current research aims to identify the geomorphic hazards caused by the tectonic activities, especially the behavior of faults, for the planning and management before the occurrence of hazards and their resulted crises. In this regard, the principal purpose of this research is to study and identify the faults of the Qoshadagh fault zone and their geomorphic impacts in the northwest of Iran and a part of the structural block of Aras Fault Zone (Berberian & Yeats, 1999).
Data and Methodology
Three methods of manual, automatic, and mixture were used to precisely extract the region's faults, affected by the doublet seismic activities of 12 August 2012 with the magnitude of 6.4 and 6.2 Richter that had only 6 kilometers and 11-minute distance from each other (Donner et al., 2015). Moreover, survey study and manual measurement of the replacements were used. Hence, Envi 5.1 software was used to apply the techniques and process of images. Also, PCA Geometrica was used to automatically extract the lineaments, and ArcMap 10.4.1 was applied to draw the output map. In the automatic method, Landsat 8 satellite image of OLI sensor with 33 rows and 168 paths was fused with its panchromatic band (15 meters). Then, the edge detection, thresholding, and extraction of fractures were conducted using algorithms. The map of lineaments was prepared using filtering, PCA, and RGB color model in the manual or visual method, and, was adapted using the automatic method.  Images of Sentinel-B2 and Spot were also used due to some properties to ensure. Finally, the obtained lineaments were checked by the field data.  
 Discussion and Conclusion
After analyzing the automatic method results, applying the lineaments manually along with the field controls, the obtained result is displayed in an overlaying map. Then, by investigating the obtained lineaments and removing the repetitive and incompatible lineaments automatically, and consultant with the tectonics experts and geomorphologists familiar with the region, the lineaments with 80% possibility of being a fault were drawn in a final map of the region’s faults. It is evident that the extraction of the definite faults of the region requires different RS sources, such as radar data, GPS, gravimetric satellites, etc. Rose diagram was used to understand the length and direction of the faults. As a result, the length of the faults was more in 100-280 degree and 145-325 degree, and no-fault was extracted along the 80 degrees, although this length can be seen in the field observations. The results of the field study, overlaying some of the geomorphological effects, including the obstruction and deviation of the waterways, formation of the pressure ridges, and fault pools, have confirmed and improved the software outputs.
Results
Using purposive remote sensing along with the field studies can play a significant role in identifying the structural ruptures that are usually hardly visible and emphasize the spatial-temporal changes after the earthquake and its role in reducing the ground hazards (Yang et al., 2017). In this regard, the faults of the Qoshadagh zone were extracted using automatic, manual, and semi-automatic methods with satellite images. The obtained results were compared with the field study and adapted to the effects and fractures caused by the earthquake and the new faults were also identified. The obstruction and deviation of the waterways, fault edges, tension, and pressure ridges, etc., are the most significant evidence that contributed to extracting the fault lineaments using the satellite images. In the current study, the faults of the region were displayed as the fault system instead of single fault lineaments that were in line with the previous studies with a different purpose (Copley et al., 2012; Ghods et al., 2015). Also, the privilege extension of faults in western and eastern directions was identified. Using automatic and manual methods is not sufficient. Therefore, the mixture method was used to obtain better and more precise results. It is evident that the extraction of the definite faults of the region cannot be obtained only by having optical images and other RS resources are required, such as radar data, GPS, and gravimetric satellites (with a high spatial separation). Furthermore, the field control and survey and investigating the geomorphologic forms must be considered an inseparable part of these studies.