Final Report Summary - SEMEP (Search for electro-magnetic earthquake precursors combining satellite and ground-based facilities) SEMEP is an FP7 funded project under the scheme EU-Russia Cooperation in GMES (SICA). Its aim is the study of electromagnetic phenomena such as the propagation of radio signals, perturbations of the plasma in the ionosphere, and ionospheric plasma waves in relation to increases in seismic activity that occur as a precursor to a large earthquake. The identification of anomalies that may be associated with increased seismic activity provides the possibility of using future observations as one method to forecast the possibility of an impending large earthquake. The amalgamation of several, independent methods of observation will lead to a more robust and reliable forecast enabling mitigation plans to be put into effect to reduce damage to infrastructure and saving lives in the process.As well as indicating the types of electromagnetic anomalies that may result from increased seismic activity this study also indicates the type of measurements that are required in future to improve our forecasts based on electromagnetic phenomena. The more information we can gain, the higher the accuracy of any forecast of an impending earthquake and the better we can be prepared for it.It is well documented that the change in the Earth's crust in the period leading up to a large earthquake can set up instabilities within the lower atmosphere that are capable of launching waves into the upper atmosphere and ionosphere where they dissipate their energy, creating localised perturbations in the plasma.The monitoring terrestrial radio transmitter signals can provide a wealth of information regarding the state of the ionosphere. By studying the reception of these signals by groundbased receivers, SEMEP has demonstrated that when the propagation path lies close to a region of enhanced seismic activity the properties of the received signal (amplitude and phase) change in a systematic way. This change typically began a few days before a major earthquake, lasting after the main event.Perturbations of the ionosphere may result in the generation of periodic oscillations of the density and temperature of the plasma. Based on satellite data, SEMEP has shown that such effects are observed in the vicinity of earthquake epicentres.Ionospheric perturbations also modify the characteristics of transmitter signals as they propagate through the inner magnetosphere and observed by satellites in low Earth orbit. In particular, SEMEP has shown that the relative amplitudes of signals from dual frequency transmitters can exhibit large changes due to changes in their propagation characteristics resulting from small changes in the composition of the ionosphere, induced by localised seismological effects.Perturbations of the ionospheric plasma lead to the generation of instabilities and an increase in plasma wave noise. Results from SEMEP indicate increases in the ULF noise level in the vicinity of an impending earthquake epicentre during the earthquake precursory period.The characterisation of anomalies in the electromagnetic environment reported during the SEMEP project provides insight into a set of observations that may be used in future to aid the generation of accurate forecasts for impending major earthquakes. This will not only help to safe guard infrastructure but also to save lives by providing a warning period and enabling a faster emergency response.As well as indicating the types of electromagnetic anomalies that may result from increased seismic activity this study also indicates the type of measurements that are required in future to improve our forecasts based on electromagnetic phenomena. The more information we can gain, the higher the accuracy of any forecast of an impending earthquake and the better we can be prepared for it.It is well documented that the change in the Earth's crust in the period leading up to a large earthquake can set up instabilities within the lower atmosphere that are capable of launching waves into the upper atmosphere and ionosphere where they dissipate their energy, creating localised perturbations in the plasma.The monitoring terrestrial radio transmitter signals can provide a wealth of information regarding the state of the ionosphere. By studying the reception of these signals by ground-based receivers SEMEP has demonstrated that when the propagation path lies close to a region of enhanced seismic activity the properties of the received signal (amplitude and phase) change in a systematic way. This change typically began a few days before a major earthquake, lasting after the main event.Perturbations of the ionosphere may result in the generation of periodic oscillations of the density and temperature of the plasma. Based on satellite data, SEMEP has shown that such effects are observed in the vicinity of earthquake epicentres.Ionospheric perturbations also modify the characteristics of transmitter signals as they propagate through the inner magnetosphere and observed by satellites in low Earth orbit. In particular, SEMEP has shown that the relative amplitudes of signals from dual frequency transmitters can exhibit large changes due to changes in their propagation characteristics resulting from small changes in the composition of the ionosphere, induced by localised seismological effects.Perturbations of the ionospheric plasma lead to the generation of instabilities and an increase in plasma wave noise. Results from SEMEP indicate increases in the ULF noise level in the vicinity of an impending earthquake epicentre during the earthquake precursory period.The characterisation of anomalies in the electromagnetic environment reported during the SEMEP project provides insight into a set of observations that may be used in future to aid the generation of accurate forecasts for impending major earthquakes. This will not only help to safe guard infrastructure but also to save lives by providing a warning period and enabling a faster emergency response.Project context and objectives:The main concept of the project was the investigation of electromagnetic phenomena related to large earthquakes in the global lithosphere-atmosphere-ionosphere coupled system using simultaneous satellite and ground-based observations. These seismo electromagnetic effects, observed as electromagnetic emissions in a large frequency range, perturbations in the ionospheric particle populations, anomalies in the records of very low frequency (VLF) transmitter signals, and night airglow observations, may be caused by natural geophysical activity such as earthquakes and volcanic eruptions. Such phenomena are of great interest because they are observed several days or a few hours before the main shock occurs and can be considered as short term precursors. The investigations were based on the analysis of the data obtained from the micro-satellites - French Demeter, Russian Cosmos-900, together with publically available data from the Cluster (ESA) and THEMIS (NASA) missions. The main objective of the Demeter satellite was to search for and characterise ionospheric perturbations that can be associated with seismic activity. Up to now, Demeter has been the only satellite that has the capability of surveying the Earth's electromagnetic environment in the ionosphere on a global scale. The high-quality database created during the mission (which contains measurements from October 2004 to December 2010) has enabled us to carry out the statistical analysis that is necessary to ascertain the link between the recorded perturbations and seismic activity. In addition to the DEMETER data, the records from the plasma density analyser onboard Cosmos-900 have been used to provide information about plasma environment. The occurrence of plasma waves has also be investigated based on data obtained from the Cluster and THEMIS missions that provide measurements of the electric and magnetic field fluctuations in the ULF frequency range.The search for earthquake related phenomena using satellite observations has an evident advantage. Uniform and global-size observations of the possible ionospheric effects from many earthquakes can be carried out in a short time together with an estimation of the size of the seismo-active region. However, ground based measurements are also necessary when investigating their relationship with seismic and volcanic activity since this is the only way to connect the physical processes in the crust with the signals recorded by the satellite.In addition to satellite observations, datasets from a recently created specialised network of VLF / LF receiver stations distributed in Russia (Petropavlovsk-Kamchatsky, Yuzhno-Sakhalinsk and Moscow), Europe and Japan were analysed. The network, which operates in conjunction with powerful transmitters deployed in Europe, United States, Asia, and Australia, enables the search for electromagnetic earthquake precursors in highly seismo-active regions such as the Far East (Sakhalin Island, Kurile Islands, and Japan) and Southern Europe. Sakhalin Island and the Kurile Islands lie inside the Pacific seismic zone, which is characterised by extremely strong seismic activity. The hypocenters of earthquakes that occur in the region of the Kurile Islands or in the Ohotskoe Sea are mainly localised in a seismofocal zone nearly 70 km wide that runs under the continent at depths of up to 650 km. This zone crops up near the deep Kurile trench, about 60 - 70 km east of Kuriles. Seismic activity reaches its maximal value in this region in which more than half of the earthquakes occur at depths of 30 - 50 km. Usually one earthquake with ? = 8 and around 10 other events with ? = 7 are recorded each decade. The hypocenters of strong earthquakes in the vicinity of Sakhalin Island are associated with the West-Sakhalin, Central Sakhalin, and East-Sakhalin faults or with their feathering faults. The seismic activity of the Sakhalin crust is more moderate, typically with one earthquake with ? = 6 and around 10 events with ? = 5 recorded on Sakhalin territory each decade. The most destructive earthquakes that have occurred in this region during the last 18 years are as follows: Shikotan (1994, Mw = 8.3 h = 40 km), Iturup (1995, Mw = 7.9 h = 50 km), Neftegorsk (1995, Ms = 7.2 h = 18 km); Uglegorsk (2000, Ms = 7.2 h = 18 km) and Nevelsk (2007, Mw = 6.2 h = 10 km).In contrast, the seismic activity of Southern Europe is much weaker. Earthquakes with ? = 6.5 occur very seldom. However, due to their shallow depth (hypocenters about 10 km), their impact on society is much greater. Europe has a rather high population density and so these earthquakes may result in extensive damage to a nations infrastructure as well as large numbers of fatalities. A joint satellite and ground-based analysis in two different seismic zones increases the chance of finding events related to different geophysical conditions. The overall goal of the project was the development of methodology for short-term earthquake forecasting. In order to reach this goal the following four major scientific objectives were identified:1. Investigation of electromagnetic ionospheric perturbations observed above seismo-active regions and their correlation with ground VLF / ULF observations and there use as a diagnostic aid to advance our understanding of the coupling between the solid earth and the ionosphere prior to the occurrence of earthquakes. 2. The development of physical models to explain the relationship between electromagnetic turbulence identified in satellite data and increases in seismic activity.3. Identification of common, reliable signatures for the precursor phases of earthquakes from temporal and spatial variations of the signal parameters obtained from multi-station VLF / ULF measurements in Europe and the Far East. This objective has improved our knowledge of the physics of earthquake precursors and enables the development of practical methods for the short-term forecasting of strong earthquakes.4. The investigation of the convective mechanism for the generation of gravity waves in the Earth's atmosphere and the coupling of large scale atmospheric structures with ionospheric perturbations.The realisation these objectives was based on the following activities:- development of short-term methods for determining the time, location and magnitude of impending earthquakes using multi-station ground observations and satellite observations;- analysis of simultaneous satellite and ground VLF signals of ionospheric perturbations, including both local events such as those of seismic or volcanic origin as well as global processes related to magnetic storms, solar flares etc.;- consideration of non-linear wave-wave interactions from correlations of VLF signal spectrum broadening, ionospheric turbulence and electron density perturbations in the upper ionosphere and their possible association with seismic disturbances, volcanic eruptions and magnetic storms;- investigations of plasma waves and plasma parameters;- estimation of the spatio-temporal scales of earthquake related precursory phenomena and study of the dependence of seismo-electromagnetic phenomena on geological and geophysical properties of the crust; - development of theoretical models of energy penetration from the earthquake epicentre, through the atmosphere and ionosphere to the satellite altitudes.The first objective relied on milestones 1 - 3 that involved the development of data processing methods for VLF / LF signals received on board a satellite. These methodologies were used for the search of long-term seismic effects. The development of physical models was based on milestones 4 - 7 which resulted from the investigations of plasma waves and plasma parameters. The third objective was based on milestones 3 and 8. Theoretical investigation of lithosphere-atmosphere-ionosphere coupling depended on milestone 9.The participants of the SEMEP project have already accumulated substantial experience in study of earthquake related phenomena in different using ground-based and satellite observations. A new method investigating the global diagnostics of seismicity using both Demeter and ground-based VLF signals had been suggested and tested over Japan. However, there is a necessity in comparative analysis (involving the integration of different advanced theoretical and experimental studies) for clarifying the physical mechanisms describing the relation between seismic activity and related ionospheric perturbations and developing theories for the transfer of energy from the earthquake epicentre, through the atmosphere and ionosphere and up to satellite altitudes. The implementation of the SEMEP project has filled several significant gaps in our understanding of the physics of the formation of ionospheric disturbances caused by seismic activity. These results have enabled the further development of methods used for short-term forecasting and early detection of large-scale natural disasters such as earthquakes and volcanic eruptions.Project results:WP 1: VLF/LF signals from ground-based transmitters received on the Demeter satelliteOne of the first activities was to be able to determine when the satellite Demeter is passing in the vicinity of an earthquake epicenter. The data from satellite Demeter above the region of analysis were selected. The Demeter data are available from the web server of the mission (see http://Demeter.cnrs-orleans. fr/ online) that is located in Orléans. This web server contains tools to select the data by dates, times, earthquakes, and specific regions. The ephemeris of the satellite is written to specific files that can be read with software (written in IDL) provided by the centre. These data have also been merged with the data files for each experiment. A seismic event catalogue, created by merging a list of all earthquakes with a magnitude M>5 with the satellite ephemeris, has been produced. This provides the possibility to search for specific earthquakes and retrieve a list of the times when Demeter approached the epicenter location during the time period covering 30 days before and 30 days after the occurrence of the earthquakes. Software to read this seismic catalogue is provided in IDL.The Demeter data were processed by the method based on difference between model signal and current signal. Selected data from the electric field receiver (ICE) for night orbits were used for modelling. Using measurements of the VLF / LF signal to noise ratio an average reference surface was constructed for the region of interest. To build the reference surface we used the method of local polynomial interpolation, which provides a polynomial expression for the reference level as a function of longitude and latitude. Data for three months were used in order to remove any effects due to seasonal variations. The modelling consists of the following procedure:1. computing a polynomial expression for the surface as a function of longitude and latitude;2. construction of the regular latitude and longitude grid 0.32°;3. computing of a net point model.Using the reference surface, at any time and for any longitude and latitude in an active region, it is possible to define the anomalous variations of the VLF signal as the difference between the measured amplitude (t, longitude, latitude) and the reference value (longitude, latitude).Analysis for three periods of seismic activity in November 2006, June - July 2008 and January 2009 was made. 1. November 2006Data from the VLF/LF station in Petropavlovsk-Kamchatsky and the ICE receiver on board the Demeter satellite covering the period 1 October 2006 until the end of January 2007 have been used for the analysis. During this period, a very strong earthquake with M = 8.3 took place near Simushir Island in the Central Kuril Region (Russia) on 15 November 2006. Following this, the series of strong aftershocks (M = 5-6.5) was observed, lasting for several months. The earthquake epicenter lay within the sensitivity zones of the wave paths JJY-PTK, JJI-PTK and NWC-PTK. The Demeter data analysed were measured during night-time passages of the satellite above the active seismic region. The zone of analysis has a width of 25°, resulting in at least one orbital passage by Demeter per day. There is an evident decrease in the amplitude of VLF / LF signals both in the ground and in the satellite data in association with increased levels of seismic activity. The amplitude anomalies are always negative. This signature can result from the effects of either magnetic storms or seismic activity and is due to signal losses as the signal propagates through irregularities in the ionosphere. Phase anomalies can be both positive and negative. It depends on the length of the path. In the present case, the anomalies in the phase of the JJY signal are positive. 2. June - July 2008Data from NWC transmitter received by the Demeter satellite and data from JJI transmitter received in Petropavlovsk-Kamchatsky were used for the analysis. The period of analysis was from the 1 June to the end of August. Two strong earthquakes occurred during this period in the Honshu region of Japan. The first earthquake with magnitude 6.9 took place on 13 June 2008 and the second earthquake with M = 7.0 happened on 20 July 2008. The epicenter of the first earthquake lay within sensitivity zone of the propagation path JJI-PTK, the epicenter of the second earthquake was outside sensitivity zones of any path for ground observations. As in the previous case, differences between the observed averaged night time signals and synthetic, model generated quiet time signals are plotted. It is clearly seen that there is a decrease in the observed signal during the period 1 - 2 days before the earthquakes.3. January 2009During the period from the beginning of December 2008 until the end of February 2009 several earthquakes were recorded in Kuril-Kamchatka region. The strongest earthquake was recorded on January 15 and had a magnitude 7.4. The epicenters of all earthquakes lay outside the sensitivity zone for the propagation path from the transmitter JJY to the receiver PTK. As a result, the signal along this path was undisturbed during the period of analysis. Some weak earthquakes did occur within the propagation path JJI-PTK. However, the signal is unaffected. Data from the DEMETER satellite were analysed to investigate changes in the signal from the transmitter NWC (19.8 kHz). During the periods in which the earthquakes were observed there is no significant change in Dst, which indicates that the magnetosphere was relatively quiet. As can be seen from the top panel, three large decreases in the amplitudes of the electric field are observed. Each decrease occurs 1 - 2 days before the onset of an earthquake.A comparison between ground and satellite observations showed similar results. Such simultaneous analysis provides a cross validation of the observations and hence a higher reliability in the results. These propagation anomalies may result from the interaction of atmospheric gravity waves (AGW) with the ionospheric plasma. In a seismically active region AGW may be generated by the release of water and gasses from the crust. This release resulting in the generation of AGW that subsequently propagates into the ionosphere. The energy that these waves carry with them perturbs the ionospheric plasma. These changes in the ionosphere then modify the propagation characteristics of the signals from terrestrial transmitters as they pass over the seismically active zone.WP 2: Theoretical and experimental studies of plasma waves in the vicinity of earthquake epicentresWP 2 investigated the plasma wave environment of the inner magnetosphere to determine if the observations of plasma waves occurred due to increase in seismic activity. The French Demeter, and ESA satellites Cluster collected the data analysed. DEMETER had low Earth orbiting satellite, with a circular orbit of around 700 km and is ideal to investigate plasma wave activity in the upper ionosphere. Cluster has an elliptical orbit whose apogee is within the inner magnetosphere at an altitude of around 1200 km. Observations from Cluster were used to investigate wave activity in the inner magnetosphere. The Cluster data were supplemented with observations from the NASA THEMIS mission. During the second phase of its operation, the THEMIS spacecraft had an elliptical orbit whose perigee was in the inner magnetosphere.ULF Wave activityDemeter data were analysed to whether or not there were plasma waves observed at the time of the Sichuan earthquake (China) on 12 May2008. For this investigation, the individual spectra were averaged using 30second time bins and then compared against the average wave power for the whole orbit. This increase in wave activity may be the result of either terrestrial source such as seismic activity or possibly due to enhanced magnetospheric activity. To examine this latter possibility the Dst geomagnetic index was analysed. Dst is a measure of the surface magnetic field of the Earth in the equatorial region. It is strongly affected by changes in the terrestrial ring current that result from increased geomagnetic activity. In the period leading up the Sichuan earthquake, the value of Dst remains fairly constant and low throughout this period, which implies that there were no periods of strong geomagnetic activity in the period leading up to the Sichuan earthquake. For this period, levels of low frequency wave activity in the inner magnetosphere were investigated using data from the Cluster and THEMIS spacecraft. Both of these constellations had elliptical orbits with perigees of the order of a few thousand kilometres. The results show that there does not appear to be a strong increase in the level of wave noise within the inner magnetosphere during this period.WP 2 also reported on the characteristics of another possible source of waves that may propagate from the surface of the Earth into the ionosphere where they dissipate their energy, causing perturbations in the plasma: the acoustic wave. On the Earth's surface, acoustic waves may be generated by either natural events such as earthquakes, volcanic activity, tsunami, etc., or man-made events such as nuclear explosions. There have been several reports in the literature regarding observations of ionospheric plasma waves associated with nuclear explosions. The intense acoustic pulse arrives in the lower ionosphere inducing localised currents that propagate along the local magnetic field lines and are then closed by field aligned currents in the conjugate hemisphere. These currents can lead to instabilities and the subsequent generation of waves. The acoustic waves may also generate micro-instabilities such as small-scale density gradients that result in the generation of various wave modes. Overall, a whole raft of instabilities covering a range of different scale sises may be produced as a result of the interaction of the acoustic wave propagating into the ionosphere.Finally, by combining number of the results generated by the whole of the SEMEP project together, WP2 presented a scenario to link events observed on the surface of the Earth during the preparatory phase of a large earthquake with observations made in the ionosphere, and the propagation of IGW.WP 3: VLF transmitter broadening over seismically active regionsDuring the project, the focus of this WP has changed in respect to that given in the description of work. The original investigation envisaged a study of the broadening of signals from terrestrial transmitters that were observed by Demeter. However, early on in this study it was noticed that for dual frequency transmitters the relative amplitudes of the signal would vary significantly depending upon the propagation characteristics of the signals. This change is related to perturbations of the ionosphere density that have been observed to occur in relation to increased seismic activity. Thus the focus of task 3.1 was changed to investigate these experimental observations. Task 3.2 was subsequently modified to model the mechanism for these propagation differences. This change has resulted in a new method for the possible short term forecast for the occurrence of large earthquakes.Using the measurements from the Demeter satellite and the earthquake database from the US geological survey server, we have performed statistical analysis of the LHR frequency above seismic regions to show a statistically significant correlation between variations of the LHR frequency and gathering earthquakes. Based on this observation, we suggest monitoring of man-made VLF signals with frequencies f close to the maximum of the lower hybrid resonance (LHR) frequency (fLHR)max in the hemisphere opposite to that of the VLF transmitters.The propagation of signals close to (fLHR)max may be qualitatively different for f > (fLHR)max and f < (fLHR)max, and since (fLHR)max is very sensitive to plasma distribution at the 'base level' of the ionosphere. The observation of unexpected transmitter signal variations can indicate unexpected variations of the LHR frequency and, thus, of plasma distribution above the receiver which may be the result of localised seismic activity. A sharp increase in both the local LHR frequency (blue) and maximum (green) is clearly seen a few days before the earthquake. This is accompanied by a corresponding decrease in the spectral intensity at a frequency of 11.9kHz the lowest of the working frequencies of the groundbased VLF Alpha transmitters. It should be noted that changes in the LHR frequency profile may result from processes in the Earth's atmosphere, ionosphere, and the magnetosphere, in addition to increases in seismic activity. As a result, unexpected variations in the transmitter signal amplitude should only be considered as one possible indicator of precursor activity associated with a large earthquake and should be interpreted in conjunction with other features of precursory activity.A possible mechanism has been suggested of variation in the ion distribution and the LHR frequency profiles above seismically active regions. Such variations were studied previously (and reported within deliverable 3.1) on the basis of Demeter measurements. It has been shown that the variations mentioned above can take place due to particle transversal drift caused by large-scale electric field arising above the region of gathering earthquake. Registrations of such electric fields were previously reported in the literature. We have undertaken a detailed analysis of plasma redistribution caused by the particle drift under reasonable assumptions about the localisation and properties of the seismic-related electric field and initial particle distribution. The present consideration amplifies the previously performed analysis and earlier suggested method of monitoring of seismic activity.WP 4: Space plasma parameters from COSMOS-900 and Demeter dataWP 4 involves a study of the plasma parameters measured by the spacecraft Cosmos-900 and Demeter. These satellites were in low Earth orbit, Demeter at an altitude of around 700 km and Cosmos-900 from around 500 km down to 300 km as the mission progressed. Thus these satellites are ideally placed to investigate the occurrence of perturbations of the plasma in the lower ionosphere and to see if there is any connection with regions showing enhanced levels of seismic activity. The first two tasks involved the extraction and validation of the data sets to be used. Data from the Demeter IAP and ISL instruments was obtained from the Demeter data centre together with essential support datasets such as the background magnetic field, files describing the particular science mode of operation, and a catalogue of occurrences of earthquakes during the period in which Demeter was operational. Due to the high quality of the data from the Demeter data centre, no further processing was required and the data were written into files to be analysed later. The data sets from Cosmos-900, in contrast, are not in such good condition and a complete reprocessing of the raw telemetry was required. After being originally collected in 1977, the raw datasets were copied to newer media in the early 80's and then again in the early 90's. This latter dataset were reprocessed as part of task 4.2. The task of reprocessing also involved the cleaning of the data, i.e. the identification and removal of erroneous points that can occur due to byte error problems caused by either onboard storage, the downlink, or as part of the copying process to preserve the data. Once cleaned, the particle measurements were combined with the satellite ephemeris data to provide the location of the satellite and also various geomagnetic indices that can be used to characterise the level of solar and geomagnetic activity. Once the data sets were ready to be analysed software were created to search, extract, and display periods of data from either the Cosmos-900 or Demeter data banks created during this work package. Additional applications were created to list the satellite passes that occurred within a few days of a specified earthquake epicenter and analyse the waveform data using a wavelet transform to investigate any periodicities that occur within the data set. Both electrons and ions exhibit a periodicity at around 0.002 Hz that occurred in the vicinity of a seismically active region. An analysis of a number of regions has shown that the plasma parameters measured by the low Earth orbiting satellites Demeter and Cosmos-900 do appear to be perturbed by enhancements in seismic activity. These effects are most pronounced in measurements of the plasma density. In most cases, periodicities in the frequency range 0.002 - 0.003 Hz are observed. The spatial extent of these density oscillations is quite large, typically spreading ± 25 - 30° in longitude and ± 40° in latitude. These variations are likely to be caused by dynamic wavelike processes in the neutral atmosphere. In addition to these wavelike variations, sharp increases or decreases in the density are also observed close to the earthquake epicentre. These piston-like features are most probably caused by the appearance of transverse electric fields resulting from increases in seismic activity. These electric fields will perturb the local plasma, causing it to either drift upwards for an eastward electric field or downwards in the case of westward directed fields and so changing the density. WP 5: Ground-based observations in VLF / ULF networksThe initial phases of this WP, dealt with the development of data sets, software applications, and models that were used to analyse VLF / ULF signals and highlight possible anomalies of seismic origin. It is important to eliminate changes in the geospace environment that may give rise to anomalous signal propagation. To aid this identification a database containing parameters characterising the geospace environment was created. The VLF / ULF data used in this study were obtained from networks of VLF stations in the Far East and Europe. Investigations based on case studies, spectral analysis, statistical and correlation analysis and investigations of tsunami effects on the VLF signal.1. Case studyCase studies have been performed to analyse perturbations in the propagation characteristics of VLF / ULF signals. These investigations included the analysis of signals in connection with the Simushir earthquake (Kuril Islands) on 15 November 2006 (M = 8.3) the earthquake in the Honshu region of Japan on 13 June 2008 (M = 6.9) and the Tohoku earthquake in Japan on 11 March 2011 (M = 9). For the first two earthquakes, the analysis was performed using simultaneous observations from ground-based stations and the satellite Demeter. These results were described in WP1. Here we present results of analysis for the Tohoku earthquake. The signal from the JJI (22.2 kHz) transmitter received in Yuzhno-Sakhalinsk and Petropavlovsk-Kamchatsky was analysed in connection with very strong (M = 9) earthquake in Japan on 11 March 2011. The zone of seismic activity laid in the central part of the wave path JJI-Kamchatka. The interval of analysis was from 1 February to 22 May. Following the earthquake there was strong aftershock activity. Another drop in the signal and increase of dispersion were observed before strong aftershock on 7 April (M = 7.1). 2. Spectral analysisWe examined changes in the spectral composition of the LF sub-ionospheric signal from the NRK transmitter (37.5 kHz) in Iceland that was received in Bari (Italy) around the time of the earthquake that occurred in L'Aquila on 6 April 2009. The strongest anomalies in the signal were observed in the NRK-Bari propagation path during the period 5 - 6 days before the earthquake as well as during the series of aftershocks. These anomalies are similar to the anomalies during strong magnetic activity. Although the characteristics of the waveform of the NRK signal anomalies appear identical during the L'Aquila earthquake and for periods of strong geomagnetic activity their resulting spectra are different. We have applied spectral analysis to signals recorded on quiet days and compared them to similar results for days on which geomagnetic and seismo-induced anomalies were observed. For the analysis we have used the nighttime amplitudes of the NRK signal filtered in the frequency band 0.28 - 15 mHz which corresponds to wave periods in the range T from 1 to 60 minutes. Only the spectra of the signal propagating along the NRK-Bari path exhibits wave activity with periods of the order 10 - 20 minutes. These periods are absent from the spectra of the undisturbed signals, those used as control paths and from spectra of magnetic-disturbed days. Similar spectral changes were also observed in the LF (40 kHz) signal from the JJY transmitter in Japan measured in Petropavlovsk-Kamchatsky for strong earthquakes (M = 6.1 - 8.3) which occurred in the sensitivity zone of this wave path in November 2004, November 2005 and November 2006.3. Statistical analysisFor the purpose of determining the sensitivity threshold of the LF signal to the magnitude of an earthquake and unearthing probable periods of observation of anomalies caused by seismic activity, a statistical analysis has been made using seven years of observations of the wave path JJY - Petropavlovsk- Kamchatsky. We found that the sensitivity of the LF signal to seismic processes becomes apparent mainly for M = 5.5. The most probable times of anomalies are 6 - 7 and 3 days before an earthquake and 6 - 7 days after it. Statistically significant analysis (made for 80 events) shows that anomalies of LF signal are observed in 25 - 30 % cases for earthquakes with ? = 5.5 - 6.5. 4. Correlation analysisA method of estimating the sensitivity of VLF / LF signals to seismic processes using a neural network approach based on a three-layer perceptron has been developed. This type of network employs the back propagation technique and is referred to as a supervised network. The development of such a network involves two main stages to solve the problem, namely the training of the network and recognition (the prediction itself). In a supervised learning scheme, the network determines the relation between the known input output pairs, called the training set. In order to train the neural network, a training database was created that included both a catalogue of seismic events between 2005 and 2007 and the corresponding amplitudes and phases of the LF signal transmissions between the Japanese transmitter JJY and the receiving station in Petropavlovsk-Kamchatsky. Seismic events were excluded from the database for periods when the geomagnetic activity index Dst and the flux of relativistic electrons and protons exceeded predefined thresholds. After many experiments training and testing the neural network the optimal properties for creating the training database were determined. As a result, the training samples included features calculated from the amplitudes and phases of the signals measured for 5 days before the 40 seismic events of magnitude M = 5.5 and which occurred at a depth of not more than 150 km. To test the capability of the network to predict the occurrence of a seismic event from LF data a set of twelve time intervals in 2003, 2005, 2006, and 2007 were chosen. Each period consisted of between 6 and 8 days of data that included the day of seismic events of magnitude M = 5.5. 5. Investigations of tsunami effects in VLF signalThe network of VLF receivers sited in the Far East has, for the first time, been used to observe the response of the lower ionosphere to tsunamis resulting from the Simushir (15 November 2006, Kuril region) and the Tohoku (11 March 2011, Japan region) earthquakes. From results of the Tsunami Travel Time software the tsunami propagates approximately along the Hawaii - Yuzhno-Sakhalinsk path. It is clearly seen that the signals received at both stations (PTK and YSH) are very similar except for those propagating along the NPM-PTK and NPM-YSH paths that show large differences in comparison to the other transmitters. For this particular pair of propagation paths, the signal recorded in Petropavlovsk-Kamchatsky travels along an undisturbed path whereas that measured at Yuzhno-Sakhalinsk clearly shows an anomalous decrease in amplitude of about 10 db together with an increase in phase of up to 50 degrees. The wavelet spectrograms of the data reveal the frequency of the maximum spectral amplitude in the range of periods of 8 - 30 min that corresponds to the internal gravity wave periods. These periods are in compliance with the periods observed in data recorded by the DART sensor buoys.A qualitative interpretation of the observed effects is suggested in terms of the interaction of internal (atmosphere) gravity waves with lower ionosphere.WP 6: Theoretical investigation of the coupling between the Earth's lithosphere, atmosphere, and ionosphereWP 6 investigated the aspects of the coupling between the Earth's crust, atmosphere, and ionosphere. The coupling mechanism involves the propagation of internal gravity waves (IGW) created as a result of the changes occurring within the crust due to the increase in seismic activity. IGWs are associated with density and velocity perturbations of the atmosphere and consist of fluctuations of particles reacting to interaction of buoyancy forces and the force of gravity. IGW provide an effective mechanism for the transport of energy and momentum from the lower layers of the neutral atmosphere to the ionosphere. The correlation between effects observed in the lower atmosphere and perturbations in the ionosphere has been observed using ground-based electromagnetic sounding techniques and satellite observations. Therefore, the electromagnetic sensing of the ionosphere is closely related to the propagation of IGW. During the earthquake precursory period, two notable changes in the lithosphere are observed, namely the emanation of gas and the generation of low frequency oscillations of the Earth's surface. These modifications to the composition of the atmosphere lead to infrared radiation anomalies and the generation of a broad spectrum of IGW. Dissipation of IGW usually results in the generation of large scale nonlinear structures (convective cells) due to wave-wave and wave-mean flow interactions. This WP investigates the propagation and effects of IGW.Generation of large-scale zonal structuresThe generation of large scale zonal structures by IGW has been investigated. The generation mechanism is based on the parametric excitation of convective cells by finite amplitude IGWs. Energy is transferred from small scale IGW into large scale zonal structures by means of an inverse turbulent cascade mechanism, creating large structures from turbulent chaos. It was found that the maximum growth rate has a convective mode whose vertically directed wave vector corresponds to the periodic characteristics of zonal structures as the altitude increases. Numerical estimations based on typical values for the ionosphere reveal that the spatial scales of the zonal structures is of the order of 1 km, a value consistent with observations from the Demeter satellite.Tidal ionospheric structures above earthquake epicentresAs was mentioned in the previous section, changes in the crustal layers during the period leading up to a major earthquake lead to the modification of the lower atmosphere and the generation of IGWs. The IGW disturbances grow in amplitude by several orders of magnitude as they attain ionospheric altitudes. Their nonlinear interaction with zonal winds, discussed above, limits the growth of IGW perturbations. Nonlinear IGWs have previously been shown to exist in the form of vortex structures whose horizontal velocity should be of the order of or greater than the speed of sound. This motion should initiate a shock wave, a situation that is inconsistent with the initial assumptions. These models also omitted effects due to vertical zonal wind gradients where as in the real atmosphere the zonal winds are inhomogeneously distributed with height. Hence, this task sought to extend these models by including effects due to the finite scale of the vertical temperature gradient and the vertical shear of the zonal winds.One of the important parameters that govern vertical transport in the atmosphere is the vertical temperature gradient. The stability of the atmosphere may be determined from the height averaged Brunt-Vaisala frequency ?g. For ?g>0 the atmosphere is stably stratified whilst strong convective transport arises when the temperature decreases with height in which case ?g<0. This latter condition corresponds to the absolute instability of IGWs and the wave frequency is purely imaginary. The basic model used in this study assumed a stably stratified atmosphere with a characteristic finite scale temperature gradient. In this case, the atmosphere is always convectively unstable. In such an atmosphere, IGWs transport energy in the direction opposite to that of the temperature gradient i.e. if the atmospheric temperature decreases with height, IGW transport energy upwards and vice versa. Thus the waves provide a coupling mechanism between the lower and upper atmosphere.As mentioned at the beginning of this section, nonlinear IGWs in a turbulent atmosphere with zonal winds can generate vortex structures whose propagation velocity corresponds to the sound speed or greater, a result inconsistent with the initial assumptions of the model. Within this task these models were extended to include the effects of a shear in the zonal wind velocity. We hypothesize that, prior to an earthquake the dynamics of the atmosphere develop according to the following scheme. Large-scale thermal inhomogeneities appear in regions of increased seismic activity a few days before strong earthquakes, forming the source of IGWs. A new type of the IGWs instability in stably stratified atmosphere with finite temperature gradient was investigated. It was shown that an atmosphere with a finite scale for the temperature gradient is always convectively unstable. Generated in such an atmosphere, the IGWs carry the energy in the direction opposite to that of the temperature gradient. It is shown that the characteristic scale of the energy density of IGWs is equal to the characteristic spatial scale of the vertical temperature. Investigating the vortex structures in the Earth's atmosphere, taking into account the effects of vertical shear of zonal wind, we obtained a more realistic estimate for the velocity of vortex structures. We have shown that in a coordinate system moving with the zonal wind the velocity of the vortices can be substantially smaller than the velocity of sound. Owing to collective vortex motions in the atmosphere and related motion in the ionosphere, the density of plasma and atomic oxygen increase. This agrees with Demeter observations.Finally, numerical ray tracing calculations for the propagation of IGWs from their generation region assumed to be in the lower atmosphere boundary layer into the ionosphere through an atmosphere consisting of inhomogeneous zonal flows, and finally to the creation of ionospheric perturbations. These simulations enable the wave trajectory to be studied, highlighting the appearance of critical layers, and horizontal / vertical reflection layers in the ray path. In order to perform realistic calculations, experimental measurements of the Brunt-Vaisala frequency together with measurements of the vertical zonal winds were analysed and their profiles approximated using analytic functions. The results of these calculations have demonstrated the following:1. The propagation of large scale IGW, i.e. those with a horizontal wave number of the order of tens of kilometers into the ionosphere is strongly dependent upon the profiles of the vertical winds and Brunt-Vaisala frequency. Under some conditions, horizontal displacements of up to 6000 km were observed. In addition, the propagation time may also be very large, of the order of 3 - 5 days.2. The main obstacle that limits the passage of IGW is the occurrence of layers within the atmosphere with vertical wind speeds that possess a critical velocity and the vertical wave number may increase many fold in the absence of dissipation processes. The critical velocity decreases with horizontal wave number growth. This leads to a strong slowing of the IGW and wave absorption takes place. 3. For high-speed zonal winds, the passage of IGW becomes impossible due to the presence of these layers, which cause the vertical wave number to vanish.4. The occurrence of horizontal reflection layers greatly effects the horizontal displacement of IGW as they propagate from their source region into the ionosphere. The wave may even cross its original source region several times as it propagates.5. The propagation time may vary from hours to days depending upon the initial profiles of the atmosphere.The results of the propagation of IGW from their source region into the ionosphere are highly dependent upon the atmospheric profile of the winds and Brunt-Vaisala frequency, which in turn depend strongly on the location of the source region and also the season.Potential impact:Potential societal impact The studies performed within the SEMEP project have investigated the occurrence of anomalous electromagnetic effects in relation to increases in the level of seismic activity. Previous results have demonstrated that seismic effects such as gas emission, thermal anomalies, or changes in the electrical conductivity of the crust are able to set up instabilities in the lower atmosphere that result in the generation of IGWs. These IGWs may propagate into the upper atmosphere and ionosphere where they dissipate, creating anomalies within the ionospheric plasma. Such anomalies affect the propagation of the electromagnetic signals from terrestrial transmitters as observed by both ground-based receiving stations and low Earth orbiting satellites. They cause variations in the composition of the ionospheric plasma, creating density anomalies and other effects that drive instabilities resulting in anomalous levels of plasma waves whose propagation characteristics within the inner magnetosphere can be modified. One key property of all of these effects is that they are often observed for a few days or more before the occurrence of a major earthquake, i.e. during the precursory period. Thus, the observation of such effects could potentially provide advanced warning of the possible occurrence of a large seismic event. However, care must be taken not to erroneous predictions so these observations should be corroborated by other seismic related observations such as increases in thermal emissions, increases in the concentration of gasses in the lower atmosphere, changes in the TEC values of the ionosphere, etc. Thus they form part of a global tool to forecast the potential likelihood of the occurrence of a large earthquake. Whilst nothing can be done to prevent any impending earthquake, the ability to produce accurate forecasts of their location and severity has huge societal implications. It means that mitigation scenario may be put into action in the effected region to reduce the impact that such large events have on society such as reducing the number of deaths, minimising the potential danger due to the damage and destruction of infrastructure such as power plants, gas, electric and water supplies, and the initiation of aid programs so that help can be on the scene within hours rather than days.Scientific impactThis project has achieved significant results regarding changes in the electromagnetic response of the atmosphere and ionosphere to increases in seismic activity. These results have been made publically available to the scientific community through the publication of papers in international, open source, peer reviewed journals and conference presentations. These results will form the basis for future scientific studies to investigate electromagnetic phenomena associated with seismic activity. The results have also highlighted several parameters that may be used as indicators of an impending earthquake since their onset occurred several days before the main event, thus providing more potential inputs to studies of the forecasting of earthquakes. In terms of the original proposal, this project has contributed to the expected scientific impact listed in the work program:'Proposals are expected to contribute significantly to the joint use of Russian satellite capabilities and European satellite for enhancing GMES Earth observation, in particular in areas related to earthquake precursors ' Within SEMEP, data from the satellites Cosmos-900 (Russia), Demeter (France), Cluster (ESA), and THEMIS (NASA) have been used to investigate ionospheric effects resulting from increased seismic activity.'Projects should also result in enhanced collaboration between Russian and European actors in this field'. The results of this project have provided the participants with increased expertise in the analysis an understanding of the response of the electromagnetic environment of the Earth to increases in the level of seismic activity. In particular, it has enhanced the collaborations between the European and Russian participants and further stimulated the interest in all aspects of electromagnetic phenomena associated with the precursory phases of major earthquakes. In particular, the USFD team have procured funding for the installation of possibly two LF / VLF receivers to add to the complement that are currently used to monitor signal propagation over the seismically active regions in Southern Europe.DisseminationThe main dissemination activities undertaken during the project relate to the publication of the scientific results produced by the project's RTD WPs. So far, there have been a total of 18 papers published in peer reviewed scientific journals with several others currently in the review process. As these papers are published they will be added to the list on the project website. This work has also been included in three chapters of books. The scientific results have formed the basis for 34 presentations at geophysical conferences such as EGU, AGU, EMSEV, etc. A list of papers, book chapters, and conference presentations may be found on the project website. Information regarding the project is available through the project website.WebsiteThe project website may be found at the address http://www.ssg.group.shef.ac.uk/semep/ The website contains descriptions of the project, its aims and objectives, and details regarding the individual work packages. There are also links to all reports generated by the project that are not private to the project as well as links to papers published in peer-reviewed scientific journals.
