Ionospheric parameters as tools for prediction of an earthquake have been adapted by many workers [1-7]. However, solar geomagnetic influences on ionosphere being significant, the processes of filtering the earthquake (EQ) induced effects from these parameters are difficult exercises. The situation becomes more complex at Appleton anomaly zone [8-11] where ionisation density is transported up by equatorial electric field (E) through EXB drift process. Therefore, at off equatorial stations i.e., round Appleton Anomaly region (±20˚ geomagnetic latitudes), the ionisation density and Total Electron Contents (TEC), have a strong component of migrating electron density. Such movement is also possible from epicentre region during EQ time electric field generated in its preparatory processes [12-14]. Added to these aspects, the anomaly zone undergoes changes with solar geomagnetic environments as well as on coupling dynamics between equatorial and magnetospheric processes [15-18], thereby making EQ time density changes on TEC or ionisation density around the anomaly as multifactor events [19-21]. There are also reports showing apparent modifications in ionisation density and on TEC by pre earthquake influences on lower atmosphere [22- 26].

All these activities make identification of earthquake related dynamics over extended latitudinal and longitudinal sectors very complex especially when anomaly feature is adapted as a tool of EQ precursor as the paper aims. Here, we analyze the growth and inhibition pattern of anomaly zone from global TEC data covering sectors from east Africa to west pacific; the feature so obtained is examined in association with major Japan EQ (JEQ) of March 11, 2011 epicentre at 38.322˚N, 142.369˚E (M = 8.9) and Indonesia EQ (IEQ, M = 8.6) of April 11, 2012 occurred at 2.311˚N, 93˚E. The two events are so selected that EQ time anomaly coupling dynamics could be brought in to focus at widely separated zones around and beyond equatorial anomaly region limited within ±20˚ geomagnetic latitudes. TEC data collected from the GPS receiving setup at Guwahati (26˚10'N, 91˚45'E) along with F-layer peak density (NmF2) data from relevant stations are used as supporting inputs.

A representative profile of TEC obtained from dual frequency GPS observation at Guwahati, an anomaly crest station (26˚10'N, 91˚45'E) is presented in Figure 1, to highlight a few features that will be utilized in the analysis to follow. The figure shows that the diurnal TEC reaches its maximum not at the local noon as expected (point A in the figure) but one or two hours from it (marked as B in the figure). Such enhancements in density after the local noon hours are generally associated with the sources at the equator from where electron density is pushed up by the EXB drift process, to be moved down along the field lines to off equatorial region, enhancing the total density at the recipient location. This is the Appleton anomaly phenomenon, described in brief in the introduction. The position of the peak density depends on magnitude of E field and also on the time of its appearance, although in general, the phenomenon is significant in post noon hours as shown in Figure 1. In this paper we would examine growth and inhibition of such anomaly zones from global TEC data over wide longitude sectors, the features so obtained will be associated with the two major EQs as described.

The explanations to the observed development of the EEA and shift of this anomaly zone towards the longitude of epicentre of the two most devastating recent EQs one at mid latitude and the other at the equator, are difficult to put forward with present understandings on equatorial anomaly, though the underlying physics and dynamics involved with the process are well established. The growth of such anomaly is coupled with conductivity of the ionosphere under an impressed electric field that becomes anisotropic in presence of a magnetic field. The result is generation of three conductivities namely Longitudinal, Pedersen and Hall conductivities. The Hall conductivity amongst the three is significant in this context. Because the Hall current flowing in a direction  perpendicular to both the electric and magnetic fields generates a force that moves the plasma upwards through the process E ´ B drift that causes transportation of ionization from the equator to regions off the equator, creating a minimum density over the magnetic equator, while causing maxima at latitudes to 20˚ (geomagnetic) north and south of the equator [8-11]. The effective enhancement of Ionization density or TEC caused by this process at off equatorial stations is generally obtained just after local noon as displayed in Figure 1. One can also see formation of such density maximum zones, i.e. Normal Equatorial Anomaly (NEA) around local noon (at 20˚E longitude) at 12 UT, in the global TEC map of Figure 3(a). But the development process of NEA is dissipated prior to the two strong EQs of this study. This apparent inhibition of NEA is replaced by an anomalous development of a high density zone at the near equator in the night sector; a phenomenon we name as EEA. The coincidence of the formation of the EEA with the longitudenal sector of the epicentre, as revealed through analysis of these two major EQs is significant. This observation may not be explained only through the proposals offered by many workers while examining relation between EQ events and modifications in ionization density or in the TEC [6,12,14,19,27,28]. One such hypothesis on EQ time changes in density peak at low-latitudes is due to manifestation of Appleton type E´B charge movement specially when the epicentre lies near to the equator and in such case E is the EQ induced electric field. It is shown by Devi et al. [6,28] that low-latitude earthquakes can also force similar changes to electron density. Devi et al. [28], have also observed that when model IRI is used to estimate earthquake time low latitude TEC, the input values on ‘F-region peak ionization density’, need to be modified appropriate to “earthquake time variations in TEC”, for achieving concordance between model output and TEC observed noontime values. For this purpose changes in F-peak density they considered are the result of dumping of ionisation to low latitude station from the epicentre through Appleton anomaly process, initiated by the EQ generated E-field. These observations and results thus suggest that E-field could contribute significantly in changing ionisation density around epicene zone. But to correlate association of the EEA i.e., a high density zone near the equator at the night sector with density modification near mid-latitude epicenter is difficult. To examine such relation we map the NmF2 data available during the period of JEQ, at I-Cheon (37.14˚N, 124˚E) near epicentre zone and also at Darwin (−12.4˚N, 130.9˚E) a location where EEA was formed. Figure 11 shows daily NmF₂ values and their monthly mean (presented by solid and dashed blue lines respectively) over Darwin (a station at the equatorial anomaly zone). The red and black portions of the curves in the Figure 11 indicate positive and negative ionospheric effects. One can see a large

accumulation of ionisation density especially on the night of March 8, (one day prior to EQ of March 9) at Darwin, and a well developed EEA in global TEC map (Figure 2 d) at similar latitude/longitude position. In I-Cheon, the station near to the epicentre, the increase in density is less intense by 50% than that of Darwin, suggesting a possible transportation from epicentre to the EEA zone. One can also note from Figure 10(a) the development of strong negative effect on the day of the major JEQ of March 11, especially at Darwin. In this aspect the presence of a magnetic storm as displayed by DST variations of Figure 11 that coincides with the March 11 EQ is to be noted. This negative density effect can be associated with the magnetic storm, a similar feature is also observed from Figure 10(a), for the storm of March 1 (Figure 11). However, the ionospheric storm time dynamics being not the subject of this paper, the negative density variation noted on the major EQ day i.e. on March 11 will not be further dealt with.

The IPS TEC global map shown here is produced with IRI model by taking real-time foF2 data obtained from the region. Therefore, the nighttime increase in NmF2 over Darwin supports the growth of EEA as reflected in the TEC global map. But how such anomaly is coupled to the major EQs occurring at latitude as high as of 35˚N, is a question. It is equally intriguing to note the adjustment of EEA longitude to epicenter longitudes. Considering all the observed features we can also look for the explanation to the hypotheses of EQ time equatorial/low latitude anomaly that suggest pumping of ionization from epicentre to off epicentre location by the EXB drift but possibly in this case the transportation process is significant during anomaly reversal time. It is well known that the day time anomaly caused by the eastward E field reverses after sunset with the development of westward electric filed. Therefore, EXB upward drift decreases during reversal time, goes to negative i.e. downward direction and reverse fountain effects become operational [29,30] and westward electric fields may result to nighttime enhancements in electron concentrations in mid latitude F-region [29], coupled with small increase in geomagnetic potential. Along with the mid-latitude increase in density, the downward compression of ionization due to large downward ExB drift can produce the nighttime increase in NmF2/TEC at the equator [31].

The large increase in nighttime TEC near equator prior to the JEQ i.e. the EEA, may be the result of such dumping from epicentre due to EQ induced westward E field, suggesting development of coupling processes between high and low latitude during EQ preparatory processes. The horizontal component of magnetic field at the location of the epicentre though weak, but the situation of zero declination contour lines [32,33] that touches epicentre to the near equatorial high density zone of Figure 2, supports possible coupling between the mid latitude epicentre and the near equatorial high density location within the same longitude zone. In case of the IEQ, the zero declination contour between epicentre and high density zone of Figure 3 supports such possibility of transportation to the epicentre longitude. Interestingly the zero-declination contour line follows the tectonic fault line of an active seismic zone, as displayed in Figure 11. Further, association of these major EQs with ongoing magnetic pole movement process that is related to inner core and tectonic status needs to be considered in the light of differential rotation of the core. Such rotation generating magnetic field and coupling process between modular structures of earth interior, resulting to a dynamo effect that may generate a two-way traffic of migrating electrons from one strong anomaly zone to another may be in a wide span of longitudes but in the same latitudinal situation, coupled by the similar magnetic declination contour. Such processes may result to dissipation of strong anomaly zone like at 20˚E (in this case) and development of anomaly in the night sector.

Along with this explanation, one has to look for other sources which are active prior to an impending EQ, specialty in the generation of a dynamical coupling from lower to atmosphere. In such an attempt Molchanov and Hayakawa [34] have shown that seismo-electromagnetic emissions mainly in lower frequencies propagate upwards, locally increasing the temperature of the ionosphere and may penetrate into the ionosphere/magnetosphere, resulting precipitation of magnetospheric highenergy charged particles into the lower ionosphere through wave-particle interactions. But seismogenic wave intensity being weak, such impact will be not effective and cannot be associated with EEA and more so in the background environment of a night sector.

One also looks for lower atmospheric perturbations trigged by seismic generated waves as another source in modulating density of ionosphere. The wavelike structures are generated in EQ environment due to modifications in atmospheric parameters like temperature, pressure, humidity. Because, the increase in temperature by the earthquake preparatory processes could develop a differential temperature situation in the steady nocturnal atmosphere thereby sharpening the density gradient or wind shear. In situations when earth’s gravity and the magnitude of stabilizing restoring force introduced by the density gradient are comparable, special structures or waves known as gravity waves are generated. They are quasi longitudinal waves and in presence of strong wind shear these waves break leading to Kelvin Helmholtz instability to generate wave structures. Development of such waves prior to one of the low altitude EQs was also seen by GU Sodar [35], as ramp type structures considered as manifestation of gravity waves [36,37]. Association of seismic triggered waves with modification in ionospheric electron density prior to an earthquake was also presented by a number of workers [38-41]. Hegai et al., [40] have computed possible effects of internal gravity waves in nighttime F layer density prior to a strong earthquake, considering origin of the waves with Joule heating due to the seismogenic electric field. They have shown that modification in ionisation density associated with wavelike character depends on the time of arrival of gravity waves at a distance r = 1000 km from the epicentre of an impending earthquake. Klomnikoto et al., [41] have presented through model computation the possibility in modulation of electron density by EQ generated gravity waves that could lead to redistribution of electron density from the epicenter zone to equatorial region, a result relevant to our observation. However, to support this mode of transportation of density no such wave signature could be seen in TEC or ionisation density in the data sets of the two case studies we have taken up here.

It is also important to consider the EQ time radon emission from lithosphere near to the epicentre that may trigger ionization in the atmosphere. The seismic fault zone being in severe stressed condition prior to a major EQ, may release radon gas. In ionosphere, the radon gas not only enhances the electron density, but could modulate the atmospheric electric field, leading to ionospheric perturbation [42,43]. In such situation the component of electric current caused by the Lithospher-AtmosphereIonosphere (LAI) coupling as has been suggested by Freund [43] and Sorokin et al., [44] may also be considered, where the group proposed that enhancement of  density is due to increase in DC electric field up to 10 mV/m in the ionosphere, thereby generating electric current into the atmosphere-ionosphere circuit. The current they said acts as source of an electromotive force in the ground-air layer through injection of charged aerosols with soil gases in the atmosphere during seismic activity period.

Finally out of all these possible sources we have zeroed down to two main factors in transporting charged particles for development of the EEA zone. One of the factors is the earthquake time downward ExB drift generated by westward electric field that leads to modification in nighttime increase in NmF2/TEC at the equator. Thus, E-H coupling mode in case of the two severe EQ events one in Japan ( lies in fault line, as marked in Figure 12), and the other in Indonesia (arrow head in Figure 12) could enhance TEC strength at locations as shown by the global TEC map. It is interesting to note that TEC values observed at Guwahati, a location touched by the fault line also recorded high density [45], prior to the JEQs. The coupling processes between mid and equatorial regions may thus develop before a strong EQ, the sig-

nature of which could be seen as an anomaly at near equatorial region in the epicentre longitude zone. This process of accumulation of ionisation density at the 100˚E - 120˚E longitudes is further enhanced by EM focusing effect generated at the EEA location due to the presence of the distortion of earth magnetic field in this sector. The focusing effect in this case can be explained through the simple theory that if the magnetic field converges towards a small region in an otherwise uniform magnetic field over an area, the flux density within the convergence zone would bring the electron current to a focus. The curve in the earth magnetic field (more so in the near horizontal component) at around110˚E and 15˚N, acts like a cone where charged particles if enter magnetic lines, may get trapped and thereby show increase in its density. Before the strong EQ cases as taken up here, the electrons with charge e moving downwards along the magnetic field line by EQ induced electric filed (as discussed above ) with drift velocity v that experience a force of magnitude e (v × B), resulting to a Hall force of H_f = −(v × B). Therefore, with the increase of Hall force, the density increases thereby strengthening the accumulation process of electrons by EQ preparatory processes at the apex of cone zone, in our case at 100˚E longitude and 15˚N latitude. Increase of drift value i.e. of Hall force with magnitude of EQ is the result of spreading of observation of TEC/foF2 far beyond epicentre zone (20, 27, 46, and 47). Here in this paper we consider that this process to be active during the EQ time along with the others and present that the magnetic distortion at special zone may lead to density modulation thereby acting as an identification signature of an EQ event.

As a possible aspect of the future study, we would use the results of this analysis as a basis for constructing quantitative models of the growth and inhibition of equatorial anomaly, to obtain a more reliable and accurate estimations of the EQ in connection with the observable quantities and their relationships.

The authors acknowledge the data support and suggestions received from the Australian meteorological department. The NmF2 data received from Ko-Ichiro Oyama, Institute of Space Science, National Central University, Taiyuan 32001, Taiwan are duly acknowledged.