A Study On The Relationship Between Thermal Information Of Satellite Remote Sensing Data And Earthquake Activities

Thermal anomalies prior to earthquakes are common phenomena. Satelliteobservation is an innovative technique that has been used to analyze the thermalanomalies prior to earthquakes. With contrast to surface observations, remote sensingobservations have the advantages in getting more information, covering area widelyand updating database easily. Furthermore, using of computer in the satellite dataprocessing will improve geo-information mining and geo-data base building.During the last decades, scientists have made some progresses in using remotesensing for earthquakes prediction based on the observations of thermal anomaliesprior to earthquakes, but there are still many problems need to be studied. Some ofthem are listed below.Identification methods of the anomalies need to be improved or even rebuilt. Thesatellite infrared brightness temperature is highly dependent on environmentalconditions, such as terrain, topography, season and some other meteorologicalparameters. The artificial visual interpretation method is intuitionistic, but hard to beoperated, tested and applied for it depends on experiences badly. The method of imagedifference analysis by subtracting an image with the background image can provideinformation of anomaly distribution, but the background image is difficult to bedetermined.Infrared radiation, latent heat exchange, and sensible heat exchange are all heatexchange processes and the first and the second one are the most important. Thethermal anomalies prior to earthquakes can lead to not only infrared radiationanomalies but also surface latent heat flux (SLHF) anomalies. At present, muchattention is paid to the study on satellite infrared anomalies. Both anomalies need tobe studied in the future.The study on the surface latent heat flux (SLHF) anomalies is just at thebeginning. The preliminary results show that SLHF anomalies are only associatedwith coastal earthquakes, and highly dependent on factors such as the region whereearthquakes happened, the season when they took place and the proximity from theocean. The meteorological parameters can induce SLHF anomalies too, so how toidentify anomalies caused by earthquakes is a challenge problem.The distribution range of thermal anomalies is known incompletely for itscomplexity. Most of the researchers consider that the anomaly is only presented in anepicentral region, ignoring many anomalies in other regions, and losing much usefulinformation.This thesis focuses on three scientific problems: 1) extracting the information ofthermal infrared anomalies prior to earthquakes, 2) developing the method to identifyanomalous behaviors of Surface Latent Heat Flux (SLHF) prior to earthquakes andanalyses of spatial distribution of the SLHF anomalies, 3) studying the relationshipbetween distribution of earth surface temperature anomalies and the regionalseismicity in one year or more and exploring the possible physical meaning behind thecorrelation. A software system has been developed, including the spatial analysisfunction of GIS, to process the satellite images, extract the thermal information andstudy the relationship between thermal anomalies and active faults. Based on theunderstanding of the physical mechanism of thermal anomalies, some representativeearthquake cases are studied to aim at the scientific problems mentioned above.1. Satellite infrared brightness temperature anomalies prior to earthquakesImages of satellite infrared brightness temperature anomalies prior toearthquakes are very complicated. The brightness temperature is highly dependent onenvironmental conditions, such as terrain, topography, season and othermeteorological parameters. High temperature regions are not always the regions ofthermal anomalies, whereas the low temperature regions sometimes are anomalousregions. The intensities of thermal anomalies are different in varied tectonic. Theincrease of infrared brightness temperature is higher inside active faults than outsidethem since the active faults are the shortcut channels of underground heat flux tocome out the earth surface. The Yaoan Ms6.5 earthquake on Jan.15, 2000, KunlunMs8.1 earthquake on Nov.14, 2001 and Mani Ms7.9 earthquake on Nov.8, 1997 aretaken as cases to study the spatial and temporal relationship between the satelliteinfrared anomalies and active faults and draw information about infrared anomalieswith explicit tectonic meaning.Two methods are used to study the earthquake cases. One is contrasting ofaverage temperature image in anomalous periods with that in normal periods. It isfound that there was an obvious belt-like thermal anomaly along the Red River fault20 days before the Yaoan Ms6.5 earthquake on Jan.15, 2000. The anomalous belt wasfading away after the event.Another method is the inside-outside temperature relation analysis method(IOTRAM). The inside region is defined as the buffer region created by GIS in 15 kmradius to the segment of the Red River fault. The outside region is the buffer region in15～30 km radius to the fault segment. IOTRAM calculates the difference of theaverage brightness temperature of the inside region and outside region in each night.IOTRAM was used in the middle segment of the Red River fault. The result showsthat the inside temperature is about 0～1.5℃ higher than the outside temperature inthe normal periods. But the brightness temperature inside the Red River fault is 2℃higher than that outside the fault in most anomalous periods. It is easy to identify theanomalous processes of appearance, acceleration before the earthquake anddisappearing after the mainshock from the curve of the average brightnesstemperature differences inside and outside the Red River fault belt.The Kunlun Ms8.1 earthquake took place on Nov.14, 2001 at the KunlunMountain. Image difference analysis indicates that the monthly average temperaturedecreases in most the northern Tibetan plateau just before the earthquake. Whereasthere are some temperature increase areas, especially near the epicenter a belt-liketemperature increase zone extends along the eastern Kunlun fault. It is considered asthe earthquake precursor. Daily temperature difference is studied by using IOTRAMin the Hoh-Sai-Hu segment of the eastern Kunlun fault in 2001. The result shows thatthe inside temperature was about 2℃ lower than the outside temperature in thenormal periods. But half and one month before the earthquake the brightnesstemperature inside region was 1℃ higher than that outside region from Oct., 2001(one month and half before the earthquake). It returned to normal after the event.The Mani Ms7.9 earthquake occurred on Nov.8, 1997 in northernQinghai-Tibetan plateau. Curves of the average brightness temperature in the eastsection of the Margaichaca fault belt and the difference inside and outside the regionof the fault belt in 1996, 1997 and 1998 were contrasted. The result shows thattemperature difference in Oct. and Nov. in 1997 was 2-3℃ higher than that in 1996and 1998, though the inside temperature in the same term in 1997 was the lowest inthree years because of a heavy snow. It returned to normal after earthquake.The cases above show that IOTRAM works well to extrat the precursorinformation from satellite infrared images in some regions or fault zones. Thetemperature difference between the inside region and outside region fluctuates in asmall range in normal periods, whereas the temperature inside the fault belt willincrease much more than that outside the fault belt because the fault is activated. Thismethod can remove not only the background differences produced by static factorssuch as terrain and landform, but also dynamic meteorological noises. Meanwhile itcan calculate the anomalous degree visible in thermal infrared images quantitatively,and draw some invisible information hidden in images.2. Analysis of surface latent heat flux (SLHF) anomaliesThere may be not only thermal radiation anomalies but also the surface latentheat flux anomalies prior to strong earthquakes. SLHF is the heat energy exchange byphase translation due to solidification, evaporation or melting. In this thesis, SLHFmeans heat released from earth surface when water evaporates.The surface latent heat flux (SLHF) data have been used to analyze theprecursors before the Mw9.0 earthquake that occurred on 26 December 2004 in thewestern coast of Sumatra, Indonesia. Several phenomena are found: (1) The SLHFshowed anomalous behavior before the strong earthquake near the epicenter. Themaximum anomaly appeared on Dec.7, 2004 when the SLHF exceeded 400 W/m2suddenly which was the maximum in 2004. The SLHF in the following ten days wasalso highest in the same date in 25 years from 1980 to 2004. It is a miracle. (2) Themaximum anomalous increase of SLHF presented in the middle part of the subductionzone between the India Plate and Burma Plate, where the rapture propagated from theepicenter to the north and the aftershocks concentrated in that region. (3) Theanomalies were scattered and disorderly at the beginning, and then rapidly increasedand concentrated, finally disappeared after the mainshock. The anomalies developedfrom the east tensional boundary of the Burma plate to the western compressiveboundary where the subduction zone exists.The anomalies of SLHF prior to the earthquake may be attributed to the increaseof stress, causing eruption of underground heat fluid into the ocean water, stimulatingthe energy exchange between the surface and atmosphere, and resulting in anomaliesof SLHF.3. Analysis of diagnostic surface temperature anomaliesDiagnostic surface temperature is assimilated mainly from the upper airobservations, and partly from surface observations in margin. The amplitude andspatial distribution of diagnostic surface temperature anomalies before some strongearthquakes took place in western China are studied. These cases are the Ashitu Ms6.9earthquake in 1996, Kunlun Ms8.1 earthquake in 2001, Jiashi Ms6.8 earthquake in2003 and Ms7.9 earthquake at the Russia-Mongolia-China border in 2003. It is foundthat the temperature in the anomalous region is 5℃ higher than the averagetemperature in 25 years from 1980 to 2004, and the patterns of anomaly distributionare correlated to regional seismicity in the following time. The anomaly distributioncan be delimited into two patterns.When the anomalies prior to a strong earthquake are distributed in several activetectonic belts (or regions) in a large region, it will follow several major earthquakesnear the anomalous belt in one year or more. For example, before the Ashitu Ms6.9earthquake in 19 March 1996,the temperature anomaly occurred not only in theepicenter region but also along the Kalakunlun fault belt where the Kalakunlun Ms6.9earthquake took place in 19 Nov 1996, in the Mani active fault where the Mani Ms7.9earthquake took place in 8 Nov 1997 and in the Yaluzangbujiang fault where severalearthquakes with magnitude above 6.0 happened form March 1997 to 1998. So did thetemperature anomalies prior to the Jiashi Ms6.8 earthquake in 2003 distributed notonly near the epicenter but also in other several regions where earthquakes fellowed inone year or more, which include Russia-Mongolia-China border Ms7.9 earthquake in2003, Delingha Ms6.8 earthquake in 2004 and a Ms6.5 earthquake in 2004 in thesouth of the Mani fault.When temperature anomalies centralize to a special active tectonic segment, asingle strong earthquake usually happen there. For examples, before the KunlunMs8.1 earthquake in 2001 the short term anomaly only occurred near the epicenterregion, as the result the seismicity in the China mainland was very weak withoutearthquakes above Ms6.5 happened in more than one year after the big event. Thesame situation happened for the Russia-Mongolia-China border Ms7.9 earthquake in2003.The above analysis indicates that thermal anomaly patterns are closely related tothe distribution of tectonic stress concentration regions (belts) though it depends onthe first focal body. When several strong earthquakes are going to take place in atectonic region, the earthquake sources may be all in a state with high stress. Once oneof them is about to rupture and anomalies arise, other earthquake sources may respondto that with anomaly emergences at the same time. But if there is only one earthquakesource in strain accumulation, the thermal anomaly may only correlate to the physicalprogress of that earthquake preparation.Conclusions:(1) A new method used in extrating the information of thermal infraredanomalies prior to earthquakes was developed. IOTRAM has some advantages inanomaly identification. This method can remove not only the background differencesbrought by static factors such as terrain and landform, but also dynamicmeteorological noises. Meanwhile it can not only calculate the anomalous degreevisible in thermal infrared images quantitatively, but also draw some invisibleinformation hidden in images.(2) The relationship between the surface latent heat flux (SLHF) and theoccurrence of earthquakes is studied for the first time in China. The SLHF anomalywas found prior to the Mw9.0 earthquake that occurred on 26 December 2004 in thewestern coast of Sumatra, Indonesia. The reliability of the precursor was discussed inthe thesis.(3) It is found that there are two distribution patterns of the diagnostic surfacetemperature anomalies prior to strong earthquakes occurred in the Qinghai-Tibetanplateau and Xingjiang region in western China. A Single piece of anomaly distributioncorrelated to a single strong earthquake and several anomalous regions (belts) respondto several strong earthquakes one year or more later.(4) A satellite observation database is built. A software system is developed toanalyze several kinds of satellite observation images and extract thermal information.Furthermore, several problems about the study and application of satelliteinfrared brightness temperature anomalies, the surface latent heat flux (SLHF)anomalies and diagnostic surface temperature anomalies are discussed.