Atmosphere response to pre-earthquake processes revealed by satellite and ground observations. Case study for few strong earthquakes in Xinjiang, China (2008-2014)

We are presenting the latest results of multi-sensor observations of short-term pre-earthquake phenomena preceding significant earthquakes. We study satellite thermal infrared radiation (OLR) anomalous signals in association with three major earthquakes, which occurred in Xinjiang province, China at different periods M7.3 of 02.12.2014; M6.2 of 08.12.2012; and M7.2 of 03.20.2008. We systematically apply multi-sensor satellite thermal data and ground temperature /humidity and estimates of the atmospheric chemical potential (ACP) parameter. Data analyses include NOAA NPOESS, the Chinese geostationary satellite FY-2D, and in-situ hourly NOAA data from the Hotan weather station. In all three cases, we detected atmospheric satellite OLR anomalies developed near the epicenter area and ACP increases (significant change for 2008 and 2014, weak for 2012) over the major Altyn Tagh fault lines within 10-20 days before the earthquake event. These findings demonstrate the occurrence and reoccurrence of transient variations of these parameters, implying their connection with the earthquake preparation process.


Introduction
The observational evidence from the last twenty years provided an extensive scientific collection of anomalous transient patterns preceding earthquakes. Some of them have shown [Tronin et al., 2002;Liu et al., 2004Liu et al., , 2010Boyarchuk, 2004, Pulinets andDavidenko, 2013;Tramutoli et al., 2005Tramutoli et al., , 2013Hattori et al., 2006] that studying atmospheric variability could also reveal active tectonic processes in the Earth's crust.
Despite the multiple singular reports for enhancement in the transient atmospheric fields around the time of significant earthquake events [Ouzounov et al., 2007;Nemec et al., 2009;Pergola et al., 2010;Lisi et al., 2010], there is still of lack of coordinated observations. Therefore it is very much needed to understand the atmospheric and environmental processes associated with active tectonic faulting and significant earthquakes. Two extended summaries of the multi-parameter approach for studying pre-earthquake processes were published recently Pulinets and Ouzounov, 2018]. These two volumes show the variety of parameters seismic, atmospheric, electromagnetic, and geochemical and the historical perspective of this research and are opening this topic to a broader geoscience community.
The latest major earthquake in Xinjiang province, China, had a magnitude of 7.2 and occurred in the Southern part of the province on 12 February 2014 ( Figure 1) followed by several aftershocks (see Table 1).
A similar type of earthquake occurred in the same region in 2008 and 2012. This study explores the relationship between atmospheric processes and the occurrence of the significant earthquakes in Xinjiang province -M7.2 of 02.12.2014; M6.2 of 08.12.2012; M7.2 of 03.20.2008 by analyzing data from ground and satellite observations. For the first time, we test the reappearance of this relationship for major earthquakes occurred in the same seismotectonic region by using multi-sensor physical observations. We have analyzed two different physical parameters characterizing the atmosphere's state during the time of earthquakes in Xinjiang province: 1. Earth outgoing radiation measured at the top of the atmosphere from different Dimitar Ouzounov et al. satellite platforms (polar and geosynchronous); and 2. Atmospheric temperature, relative humidity, and computed atmospheric chemical potential from ground observations. This approach provides a complex view of the scale and physics of changes in the atmospheric processes related to tectonic activity.
Our first reports on the pre-earthquake detection of satellite-based anomalies associated with major seismicity in Xinjiang province was in response to the M7.3 earthquake in 2014 [Sun et al., 2014;Ouzounov et al., 2015].
Meanwhile, OLR data associated with five earthquakes (M > 6.0) in the Xinjiang area of China between 2008 and 2014 were analyzed [Jing et al., 2016]. The spatial-temporal analysis results confirmed that three major events (the same event we analyze in this paper) had been identified with significant thermal anomalies.

Satellite observations
One of the main parameters we characterize the Earth's radiation environment is the outgoing long-wave-earth radiation (OLR, eight to12 μm). OLR occur at the top of the atmosphere integrating and are integrated emissions from the ground, the lower atmosphere, and clouds [Ohring and Gruber, 1982] and primarily been used to study Earth's radiative budget and climate [Gruber and Krueger, 1984;Mehta and Susskind, 1999]. The National Oceanic and Atmospheric Administration (NOAA) Climate Prediction Center provides daily and monthly OLR databases on POES polar orbit observations. The algorithm for analyzing the Advanced Very High-Resolution Radiometer (AVHRR) OLR data is not directly measured, but rather is calculated from the raw data using a different algorithm [Gruber and Krueger, 1984]. The NOAA estimate of total longwave radiation flux is obtained by applying a regression equation (infra-red) IR window observations.

3
Atmosphere response to pre-earthquake processes -is the radiance equivalent of brightest temperature BT (K) in the IR window over nadir, is the flux equivalent BT(K), a and b are the regression coefficients.
OLR is mainly sensitive to near-surface and cloud temperatures. A daily mean data footprint covering a significant area (90° N -90° S, 180° W -180° E) and with a spatial resolution of 2.5° was used to study the OLR variability in the zone of earthquake activity [Ouzounov et al., 2007Xiong at al., 2010]. An increase in radiation and a transient change in OLR were recorded at the top of the atmosphere over seismically active regions. This phenomenon was proposed to be related to thermodynamic processes in the earth's surface. In the first approximation, we can define the atmospheric anomaly in the Euler frame of reference by subtracting the mean value. The mean can be defined as the average for the same day of the year, local time, and location over more than ten years. The advantage of this approach is its simplicity and effectiveness with the availability of long-time satellite observations. Following this, the OLR anomalous variations were defined [Ouzounov et al., 2007] as an E_index. This index is similar to the definition of an anomalous thermal field proposed by Tramutoli et al. 2005.
The E_index was constructed as statically estimated variability in OLR values for specific locations and periods: Where: = 1, K days, * ( , , , , ) is the current OLR and * ( , , , , ) is the computed mean of the OLR field, defined for multiple years of observations over the same location and same local time, , is the standard deviation. We use the Anomaly Index, a modified version of _ (Eq 2), which represents the regional calibration of _ estimates we apply for Xinjiang province and define as: Where A and B are regional calibration coefficients, A -is a mask, mainly defined by the regional seismotectonic patterns and frequency of appearance of OLR anomalies for the historical events. The Xinjiang province values vary around 0.6-0.8; B -normalizes each of NOAA 15 and 18-time series of OLR data to the same time coverage (ten years). Range of B for the Xinjiang province region 1-3.5. The Anomaly Index data have been processed with a resolution of (2.5° × 2.5°), and the output maps have been processed with additional preprocessing to avoid aliasing of short wavelengths and with spatial filtering based on a "minimum curvature" algorithm .
In addition to NOAA OLR daily data, we analyzed and OLR from the FY-2D Chinese geostationary satellite.
FY-2D is very similar to GMS and GOES-3/7, and the primary payload is a Visible and Infrared Spin Scan Radiometer (VISSR), which obtains hourly full-disk images of the Earth, in five channels: VIS (0.55-0.90 μm), IR (10.3-11.3;11.5-All anomalies were building near the epicenter area, over the major Altyn Tagh fault lines within 10-15 days before the seismic event. Thus, it probably is connected with the gas release enhancement and additional flux emission and provides information about major fault activation in this area. A join plot of OLR NOAA and FY-2D over the epicentral areas for January 2 until February 15 is presented in FY-2D data respectively [Ouzounov et al., 2007]. The differencing days of appearance ( Figure 5) probably result from differences in the spatial/temporal resolution, local time of observations, and the different types of sensor payload on both NOAA15/18 and FY-2D platforms. More investigation is needed to perform a cross-platform correlation between the two OLR products.

5
Atmosphere response to pre-earthquake processes

Ground atmospheric measurements
To understand the day-by-day variability of the daytime and nighttime temperature during the time earthquake events, we analyzed the hourly temperature and relative humidity near the epicenter. We used in-situ hourly global surface air temperature for Hotan weather station provided by NOAA NCDC. (Location 37°N 80°E, about 270km north from the epicentral area). The presence of ions in the atmosphere creates a possibility for water vapor molecules to join these ions through the hydration process, which is different from condensation. In the process of condensation, the chemical potential is equal to the Latent heat, which is Q = 40.683 kJ/mol or U0 = 0.422 e V per one molecule. The evaporation/condensation process, which is the phase transition of the first order, always takes place during the chemical potential equality. However, newly formed ions have different chemical potential; what we should consider in the one-component approximation we introduce the correction to the chemical potential DU which considers this circumstance.
In this case, the real chemical potential can be expressed as: Where U0 is the chemical potential for pure water, and U(t) is the chemical potential where the ionization and hydration are considered. By cos2t we consider the daily changes of solar radiation. The increase of the water molecules chemical potential ΔU (which can be derived from (4)) indicates the strength of the nucleation process and can be used as an indicator of earthquake approaching according to Boyarchuk et al., [2010].
We analyzed hourly surface air temperature data obtained from Hotan ground station for two months around  Figure 6 reparents the natural atmospheric variability during the thermodynamic changes. Plots are presenting "raw" in-situ data without statistical validation. We choose this approach only to illustrate the first order of processes in the atmosphere's different states. Of course, statistical analysis is needed to show the significance of those processes and separate the real anomalies from the false alarms.

Discussion and conclusions
Our preliminary analysis of atmospheric parameters from satellite and ground observations during the three demonstrated the presence and re-occurrences of related variations of these parameters, implying their connection with the earthquake preparation process (see Table 2).

9
Atmosphere response to pre-earthquake processes which suggests physical links between the different atmospheric variations and tectonic activity . The primary process is the air ionization produced by the increased geogas emanation (CO2, CH4, radon) from the Earth's crust in the vicinity of active tectonic faults. The near-earth surface processes include air molecules ionization (by a-particles -a product of radon decay), which leads to the formation of complex molecular ions and the formation of large ion clusters up to aerosol size. This chain of events leads to changes in the air conductivity and the latent heat release (increasing of air temperature) due to water molecules attachment (condensation) to ions [Prasad et al., 2005;Pulinets et al., 2006]. Our findings provided evidence of the thermal buildup in increasing mean air temperature in the atmosphere. The initial increase in the ACP, T, and drops in RH have been followed (see Table 2  Our results show the re-occurrence of transient anomalies in the atmosphere in advance of the three major earthquakes in Xinjiang province, China -M7.3 in 2008, M6.3 in 2012, and M7.2 in 2014. These also suggest that systematic use of multi-parameters and multisensory satellite and ground type of observation provides additional physical validation of pre-earthquake anomalies triggered by the coupling processes between the lithosphere and atmosphere.