Changes in subsidence and uplift and the nighttime land surface temperature anomaly related the distance to the earthquake epicenter and the faults using Sentinel and MODIS imageries


 When an earthquake occurs, the faults of the region usually heat the rocks and soil of the region due to their movements. The purpose of this study was to analyze the uplift, subsidence, cloud cover and changes in nighttime land surface temperature (nLST) anomalies around faults and the earthquake epicenter in Kermanshah, Iran (date and time of earthquake 12 November, 2017 at 18:18 Coordinated Universal Time (UTC) and at 21:48 Iranian time(. Heat changes were investigated by considering the effect of other cooling factors such as vegetation (EVI), land altitude and soil moisture, rainfall and water areas. Using the MODIS sensor product, the amount of cloud cover and cooling factors were obtained. Using sentinel 1A the amount of earth uplift and subsidence were calculated. The results showed that using statistical analysis, a significant difference was observed in the nighttime land surface temperature around the faults and around the uplift and subsidence on the night of the accident, before and after night of earthquake. However, there was no significant difference between nighttime temperature and changes in the rate of spatial variation of cooling factors. It was found that the earthquake caused an increase in temperature at the fault and earthquake epicenter location. It also causes changes in height such as uplift and subsidence. Cloud cover situation showed before the earthquake, the cloud density was high and after the earthquake, the cloud density decreased. Crises managers can consider these results for monitoring metropolices for more readiness before earthquake accordance.


Introduction
The current progress in remote sensing sciences reveals different processes related with earthquakes i.e. earth's deformation, surface temperature anomaly, atmospheric gases, aerosol exhalation, ionospheric total electron content and electromagnetic disturbances in the ionosphere (Eleftheriou et al., 2016;Asim et al., 2017aAsim et al., , 2017bJilani et al., 2017;Barkat et al., 2017;Awais et al., 2017). Barkat et al., (2018)  anomalies with seismic and earthquake activity. These processes carry precursory information related with seismic and earthquakes and can serve as potential indicators within the context of earthquake predicting. Most of the precursory signals contain signi cant data and information for earthquake forecasting along with their limitations, but the satellite thermal infra-red (TIR) signal has gained more attention and support from the scienti c and expert community across the world (Saradjian and Akhoondzadeh, 2011;Akhoondzadeh et al., 2018). This can be attributed to its ability of providing valuable precursory information prior to near or distant earthquakes (Xie et al., 2013;Eleftheriou et al., 2016;Bhardwaj et al., 2017).
Researches that are pointed to as 'thermal anomalies' have been reported to precede earthquakes worldwide since the 1980s see, for example, (Tronin, et al. 2002;Tramutoli et al., 2015;Khalili et al. 2020).
Several references describe sudden increases in brightness temperatures (BT) recorded by satellite sensors (Lisi et al., 2010;Ouzounov et al., 2006;Tramutoli et al., 2001); in land surface temperatures derived using either satellite observations (SREEJITH et al., 2016;Lisi et al., 2015; or numerical simulations (Alvan et al., 2014;Qin et al 2012); in air temperatures recorded with groundbased meteorological stations (Pulinets and Dunajecka, 2007;Panda et al., 2007); in satellite-based outgoing longwave radiation (OLR) (Lu et al., 2016;Ouzounov et al., 2007); in surface latent heat ux (SLHF) (Cervone et al., 2006;Dey and Singh., 2003); and in soil temperatures measured in the eld on-site (Rezapour et al., 2010;Liu et al., 1999).The physical link between observed anomalies and earthquakes has not been established so far, however, and the main reason can be traced back to methodological shortcomings in existing literature and background (Jordan et al., 2011). De nitions of what is an earthquake related thermal anomaly vary among researchers and among different methodologies.
Observed anomalies also seem to vary for the same earthquake: they may appear a few hours ) to a few years (Yao and Qiang, 2010) prior to an earthquake, and they may reappear shortly after an earthquake (Tronin, 2000;Zhang et al. 2021). The spatial extent of reported anomalies is not clear, because the spatial resolution of the input data andimages can be as limited as point observations from a meteorological station (e.g., (Jie and Guangmeng, 2013)) or as coarse as gridcells of * (e.g., (Singh et al., 2010)). Study areas are often limited, in location and in time, around the earthquake (e.g., approximately one month and only in the pixel covering the earthquake epicenter in Ouzounov and Freund, 2004)). Such settings do not allow for detailed examination of the spatiotemporal coincidence of earthquakes with distinguished anomalies and do not permit to detect false positives. Statistical evaluation of the results is often missing and relevant discussion can be found rarely (Eneva et al., 2008). In addition, anomalies are often found to relate to atmospheric in uences and artefacts due to data processing (Blackett et al., 2011). Pavlidou et al., (2016) considered that an earthquake is large when it has magnitude larger than Mw 5.5 and that an earthquake is shallow when it has focal depth <35 km. They studied 20 earthquake cases in 10 study areas around the world, with different local environmental and climatic conditions. They applied a methodology which suppresses large-scale patterns in the satellite signal time series and isolates only spatially localized uctuations (Pavlidou et al., 2016). This methodology allows to limit the spatial extent of detected anomalies and the time of their occurrence. Thermal anomalies might appear for a variety of reasons other than earthquakes, including spatiotemporal variations of surface spectral emissivity (Tramutoli et al., 2005;Khalili et al. 2020) and local atmospheric conditions, like atmospheric inversions (QU and SHAN, 2006). (Pavlidou et al. (2016) tested the hypothesis that more anomalies would be detected at closer distances to the earthquake, shortly prior or during the earthquake, and only in years with earthquake occurrence. This hypothesis is supported by some published research such as (Xiong and Shen, 2017), concluding that nature anomalies increase with increasing earthquake magnitude; nature anomalies are found predominantly near the epicenter, one day before and on the day of the earthquake; and anomalies are more easily seen during shallow earthquakes than the deep ones. They statistically evaluated their ndings, taking into account the spatial and temporal occurrence of detected anomalies and earthquakes. Also, Thermal infrared (TIR) remote sensing has recently emerged as a promising technique for distinguishing seismic precursors. Anomalous TIR emissions have been detected by satellite sensors before the occurrence of major earthquakes (Piroddi et al., 2014;Wei et al. 2020).
Meanwhile, (Bellaoui et al., 2017) studied the 21 May 2003 Boumerdes earthquake and detected a TIR anomaly that had persisted for 1 week during the prior month (Bellaoui et al., 2017). Furthermore to analyzing the TIR anomalies for a single earthquake, (Tramutoli et al., 2013;Zhang et al. 2021) studied the causes of TIR anomalies a test over an area affected by variable gas emissions -to showe the correlation between TIR anomalies and seismicity and found that general gas dispersion models and spatial features lend support to the hypothesis of a robust correlation between greenhouse gas emissions and TIR anomalies related to seismic activity (Tramutoli et al., 2013;Wei et al. 2020).
Gaps seen in the background include unrelated uplift or subsidence to earthquakes, or nighttime temperature from satellite imagery and its association with proximity to faults or epicenters. In this study, lling these gaps was considered. Regarding the uniqueness of this research, it can be mentioned that using RAdio Detection And Ranging (radar) data from Sentinel satellite as images with new technology, the places that were affected by the earthquake were raised or subsided. Then, using data related to night temperature, satellite images, cloud cover, vegetation cover, humidity and water areas and other in uencing factors, their impact around faults as well as areas with subsidence and uplift were investigated. The hypotheses were that there was a relationship between a signi cant increases in temperature in the days before, after and the day of the earthquake near the faults and the epicenter of the earthquake. There have also been landslides and subsidence near the epicenter. Another assumption is the presence of dense clouds in the days before the earthquake and their disappearance after the earthquake. The aim was to study the trend of changes in night temperature, uplift and subsidence of the areas around the fault and the epicenter of the earthquake in the occurrence of the earthquake by eliminating the factors that have a disturbing effect on temperature.

Study area
The epicenter of the earthquake was Sarpol-e-Zahab city, which is adjacent to Thalas Baba Jani city from the north to Qasr Shirin city from the west, to Gilangharb from the south, and from Iraq and Qasr Shirin city from the west. The date and time of the earthquake, November 12, 2017, at 9:18 pm local time, was 7.3 magnitude in Kermanshah province, the border between Iran and Iraq. The average altitude of the city is 550 meters above sea level in a sudden change in altitude compared to the east of the city in a distance of less than 10 km to more than 1000 meters, which creates a very attractive and pleasant nature with the climate of the surrounding lands. In addition, this has created surface water currents and nourished the aquifers of the fertile plains of the city. Sar Pol-e Zahab city has a climate with mild winters and hot summers, with maximum rainfall in winter and a small amount of spring, and dry autumn and summer without rain. Based on the division has three different climates with mild winters and hot and long summers in the central part and cold winters and dull summers in the north and mild winters and hot summers in the northwest with an average rainfall of 500 mm per year and a temperature of 3.4 °C is the coldest month and 44.8 °C is the hottest month of the year. In Figures 1, 2 and 3, you can see the anamolic diagrams of temperature, temperature and precipitation, and the diagram of temperature and precipitation distribution for Sarpol-e-Zahab region. (https://weatherspark.com/y/104353/Average-Weather-in-Kermanshah-Iran-Year-Round) The nLST of case study (with minimum -1.81°C, maximum 17.75°C and mean 10.36 °C with standard deviation 2.93 °C) were extracted for all of study points (number of points 5014).
The examiner of hypothesis was that in each buffer of surrounding surface earthquake epicenter, faults, subsidence area and uplift area, the range nLST was not the same. This hypothesis has been done with non-parametric statistical test of Kruskal-Wallis.
In addition the range of subsidence and uplift was calculated and was processed with RADAR images from sentinel-1A in C-band that has caught in 2017 November 7 th and 19 th . (Table 1). 3.9 m × 2.6 m × 2.5 m 5m Cband The earthquake epicenter was speci ed and surrounding buffers was delineated in several circles to group area for nLST changing trend. The points with the equal distance (systematic) were made in buffers. Intersection of points with nLST was done in GIS environment for the area that earthquake took place. These points that were included data of temperature were intersected with different polygon buffers. Because of there were more two buffers (as groups) in each examination, the non-parametric statistical test of Kruskal-Wallis was chosen in analysis and process of hypothesizes. The relationship of night temperature and distance of surface earthquake epicenter and distance of faults were studied. The nLST map was included the two nights before and after of earthquake event. The distribution and scatter plot were examined and the gradient of the t line in each process was compared. In control of that changing temperature in surrounding of earthquake epicenter or faults should be not originated by others criteria such as changing elevation, changing vegetation or water bodies (river, dam lakes…), those points was intersected with elevation, soil moisture and vegetation density maps. With zonal and intersect order in GIS, the mean and values of elevation, soil moisture and vegetation density in each buffer was identi ed. Then these changes were compared for statistical assessments. If the changes be same we subsequently can inference that earthquake could change nLST. If with far from earthquake epicenter the range of elevation, soil moisture and vegetation density change signi cantly, it expect that we face with the trend of changing temperature that can disarrange our hypothesis about changing temperature consequently changing nLST cannot be from the earthquake event.

Results And Discussion
It was found that the difference in buffers is signi cant in terms of temperature. So, the temperature is higher near the earthquake epicenter or faults. The Kruskal-Wallis comparison results rejected the null hypothesis that the temperature was equal in buffers, and the difference in temperature was proved in buffers by rejecting null hypothesis. On the other hand, according to the obtained graphs, it turned out that the temperature (°C) was higher in the buffers closer to the earthquake epicenter, faults, uplift ( cm), and subsidence ( cm) due to earthquake (Fig. 2). The scatter plots revealed that the distance to the earthquake epicenter and the land surface temperature of the night, showed the reverse relationship (slight negative line slope). Before and after the earthquake occurrence, the distribution of temperature and distance to the earthquake epicenter did not show a strong negative relationship, but just at the night of the earthquake occurrence, this relationship was shown to be strong negative. In the distance to the faults, the same results were shown. to the cloud were also taken twelve days before and one week after the earthquake (Fig. 3). These images show that before the earthquake, the cloud density in the region was high and after the earthquake, the cloud density decreased.
In this study, thermal anomaly images in ve months before and three months after the earthquake were prepared using MODIS sensor products (Fig. 4).

Statistical study of temperature changes around the epicenter
The null hypothesis is that the temperatures are equal in all the buffers around the epicenter, the rejection of which means that the buffer temperatures were statistically different. This indicates clear and signi cant thermal changes around the surface epicenter of the earthquake. Graphs and data also show that the temperature near the epicenter was higher than around (Fig. 5).
Then the average altitude buffers were obtained from the zonal function and the altitude changes were compared with the temperature changes in the region to control the ineffectiveness of altitude on the warmer epicenter. Colder areas farther from the epicenter were not due to higher altitudes. Also, features such as river or vegetation that form areas with lower temperatures were controlled in the same way and it was found that the low temperature was not due to the density of these natural features.
Statistical study of temperature changes with distance from faults, uplift and subsidenceStatistical comparisons around faults in strips at distances of 1000, 3000, and 5000 m showed a signi cant difference (Table 3). As it moved away from the fault line, the nighttime temperature in the bands decreased. Similarly, signi cant temperature differences were observed around the uplifted area caused by the earthquake. There was a decrease in temperature as we moved away from the bulge area as well as the subsidence area caused by the earthquake (Fig. 6).

Ordinary least squares (OLS)
Ordinary least squares or OLS Regression In statistics, ordinary least squares (OLS) is a type of linear least squares method for estimating the unknown parameters in a linear regression model.
Calculations were also performed with unclassi ed or unbuffered data from the earthquake epicenter and night temperature. Kolmogorov-Smirnov test or K-S test, which is for normal data, which is not signi cant thus indicates the data is not normal. Therefore, non-parametric statistics were used. In these calculations, the number of data was 770 and the slope of the regression line was -0.27, which indicates the negative relationship between these two factors. That is, the greater the distance from the epicenter of the earthquake, the lower the temperature. This con rms the research hypothesis (Table 4, Table 5, and Table 6). With unclassi ed or unbuffered data, the distance from the fault and the night temperature were also evaluated for normality. The data were not normal and the number of data was 2138 and the slope of the regression line was -0.13. The regression was signi cant at the 0.99 con dence level. This negative relationship also shows the inverse relationship between these two factors. That is, the greater the distance from the fault, the lower the temperature. This con rms the research hypothesis (Table 7, Table  8, and Table 9). In assessment of height, vegetation, and water impacts on nLST, it was found that changes in altitude in buffer were not signi cant due to distance from the earthquake epicenter and faults. Therefore, altitude changes did not effect on the nLST. Additionally, with the distance from the earthquake epicenter and faults, the amount of vegetation density did not signi cantly decrease or increase and this factor also did not effect on nLST changes. Furthermore, Lakes and rivers have the equal distribution around the earthquake epicenter. In addition, distribution of soil moisture in area was same in statistical assessment. Thus, the nLST changes did not due to water distributions or soil moisture differences. These are consistent with the researches (Piroddi et al., 2014;Wei et al. 2020).
Independent-Samples Kruskal-Wallis Test was done for cooler factors (Table 7, and Fig. 11). These are statistical analysis consistent with this research (Eneva et al., 2008) and advances mentioned in (Zhao et al. 2021). The result of this test showed no factors had signi cant change around the faults and the epicenter, thus changes of the nLST are proved can occur just from the earthquake. In this study, the reverse relationship between nLST, distance from earthquake epicenter and the distance from the fault were proved. In many studies, the nLST was not used, which is the strength and unique aspect of our research (Tramutoli et al., 2005;Tramutoli et al., 2013;Wei et al. 2020;Khalili et al. 2020). The nLST makes it possible to check and assess the real effect of the heat generated from these sources while in the daytime LST the effect of the heat of the sun makes some errors.

Conclusion
In this research was studied and used a combination of RADAR remote sensing techniques, thermal remote sensing, GIS integration, and statistical analysis. These explorations, propose to use the night data in an online bed, to improve the alert of the earthquake events. These explorations could test the complex factors and behaviors and their relationships.
Limitations in this study are high uncertainty in the study of cloud status. There were few sources on this.
During an earthquake, the interference of heat from different sources of heat or cold in some cases, low temperature changes is one of the limitations of this study. The use of points in spatial statistical analysis of earthquake-related factors was a strong point in this study. The advantages of this method are the use of a set of data, including subsidence and uplift using radar images, as well as the use of nLST anomaly using satellite thermal images, as well as considering the factors of changes in clouds position in earthquake. The implications and applications of this study are in identifying areas similar to the study area. This can prepare governments for retro tting buildings. In some cases, by studying the temperature changes according to the position of the faults and in some cases according to the condition of the clouds to some extent to warn of earthquakes to be effective. It is also possible to perform the necessary crisis management operations by zoning cities and provinces, according to the factors of faults, heat, clouds, and land subsidence in high-risk areas.  Earth surface temperature anomaly (°C) in 2017, the star shows earthquake situation on the map and graphs (https://www.ncdc.noaa.gov/ Access date 14 Nov 2019) Figure 3 Flow chart of study stages Cloud images of the quake-hit area taken from the MODIS sensor a few days before and after the earthquake Figure 7 Page 20/21 Monthly average of thermal anomalies several months before and after the earthquake in Iran around the epicenter Figure 8 Buffer 2000 to 12000 m around the picenter of the earthquake, Buffer 2000m had max nighttime temperature and buffer 12000m has least. Buffers' temperatures are signi cantly different. This indicates the effect of the earthquake on the temperature of the region. These temperatures were from one week before to one week after the earthquake Figure 9 Statistical comparison of Kruskal-Wallis test of independent samples of nLST from one week before to one week after the earthquake is shown in three buffers of 1000, 3000 and 5000 meters and the Page 21/21 difference between the three buffer groups was signi cant Figure 10 Scatter plot of nLSTand distance to faults (row 1), to earthquake epicenter (row 2), expantion of points and moree steepness of trend line in night of accidenc is obviouse Figure 11 Statistical comparison for cooler factors around uplift center, these factors are same in buffers and have no effect on nLST changes