The VLF transmitters ’ radio wave anomalies related to 2010 1 Ms 7 . 1 Yushu earthquake observed by DEMETER satellite 2 and the possible mechanism 3

Earthquakes may disturb the lower ionosphere through various coupling mechanisms during their 8 seismogenic and coseismic periods. The VLF signal radiated from ground-based transmitters will get affected 9 when it penetrates the disturbed region in the ionosphere above the epicenter area, and this anomaly can be 10 recorded by low earth orbit satellite under certain conditions. In this paper, the temporal and spatial variation 11 of the Signal to Noise Ratio (SNR) of the VLF transmitter signal in the ionosphere over the epicenter of 2010 12 Yushu Ms 7.1 earthquake in China is analyzed. The results show that the SNR over the epicenter of Yushu 13 earthquake especially in the southwestern region decreased (or dropped)revealed by one satellite revisit period 14 before the main shock, which is consistent with the observed TEC anomaly at same time, implying that the 15 decrease of SNR might be caused by the enhancement of TEC. A full-wave method was used to study the 16 mechanism of the change of SNR before the earthquake. When the electron density in the lower ionosphere 17 increases by four times, the electric field will decrease about 1 dB, indicating that the disturbed electric field 18 decrease 20% compared with the original electric field and vice versa. It can be concluded that the variation of 19 electron density before earthquakes may be one important factor influence the variation of SNR. 20


8
seismogenic and coseismic periods. The VLF signal radiated from ground-based transmitters will get affected 9 when it penetrates the disturbed region in the ionosphere above the epicenter area, and this anomaly can be 10 recorded by low earth orbit satellite under certain conditions. In this paper, the temporal and spatial variation 11 of the Signal to Noise Ratio (SNR) of the VLF transmitter signal in the ionosphere over the epicenter of 2010 12 Yushu Ms 7.1 earthquake in China is analyzed. The results show that the SNR over the epicenter of Yushu 13 earthquake especially in the southwestern region decreased (or dropped)revealed by one satellite revisit period 14 before the main shock, which is consistent with the observed TEC anomaly at same time, implying that the 15 decrease of SNR might be caused by the enhancement of TEC. A full-wave method was used to study the 16 mechanism of the change of SNR before the earthquake. When the electron density in the lower ionosphere 17 increases by four times, the electric field will decrease about 1 dB, indicating that the disturbed electric field 22 23

24
The VLF (Very Low Frequency) radio waves radiated by the powerful ground-based VLF transmitters 25 have been utilizing for long distance communication and submarine navigation, because of the efficient 26 reflection within the earth-ionosphere waveguide. However, there is still a small fraction of the wave energy can 27 leak into the higher ionosphere and magnetosphere after being absorbed by the lower ionosphere. The signals  Since the seismo-ionospheric disturbance zone does not uniformly appear above the epicenter, the location of 51 the SNR abnormity in relation to the epicenter should be furtherly studied. Which factor influences the SNR and 52 its possible mechanism, that is also needed to be comprehensively illustrated. Therefore, in this paper we 53 investigate the temporal and spatial SNR variation of the VLF transmitter signal in the ionosphere over the 54 epicenter of 2010 Ms 7.1 Yushu earthquake, as well as the background variation of SNR in the same period of 55 2007-2010 to be distinguished whether the SNR reduction is caused by earthquake events or just ionospheric 56 background changes. The mechanism of how the seismo-ionospheric disturbance affect the variation of SNR 57 has been discussed in this paper.

58
As the mechanism of the VLF radio wave variations in the altitude of LEO satellite (presented as SNR 59 variation) before the earthquakes, Hayakawa (2007) and Píša et al. (2013) suggest the VLF anomalies exist 60 because the lower ionosphere is lowered before earthquake. Molchanov et al. (2006)

108
In this study, the night-time PSD data of electric field from the DEMETER's survey mode observations 109 within the region of epicenter ±10° (black square in Figure 1) were extracted, and the data within 600 km 110 circle (shown in Figure 3) over the epicenter was used to study the perturbations of the VLF radio waves from 111 the Russia transmitters before and after the Yushu earthquake. The spectrum data of the 1st re-visit period 112 (April 2 to14) before the earthquake within 600 km are averaged shown as Figure 2  119 (1) 120 where A(f0) is the amplitude of electric field spectrum at the central frequency, and A are the 121 spectrums at ∆ , where ∆ is the chosen frequency band. As can be seen in Figure 2, the 122 VLF radio signals radiated from ground based VLF transmitters has different transmitting frequency 123 bands. For the three Russian VLF transmitters, the f0 is set as three VLF radio waves frequency radiated 124 from NOV transmitters: 11.9/12.6/14.9 kHz, and the ∆ =300 Hz.
Where ω is the angular frequency, μ is the permeability of the medium (μ ≡ 1 for non-magnetic 135 medium), ε = ε I χ is dielectric tensor, and χ is electric susceptibility tensor (Yeh and Liu, 1972).
The electromagnetic field in each layer can be obtained in the k (wave vector) domain by solving equation (3) 147 recursively (Budden, 1985;Lehtinen and Inan, 2008). More details of full-wave method is described in 148 Lehtinen and Inan (2008).  the 12th day is greater or smaller than its previous 11-day-based UB or LB, a positive or negative abnormal of 186 SNR will be identified. Figure 5 shows the time series of SNR at 11.9, 12.6, 14.9 kHz respectively. The red,

247
The refractive index surface of the upgoing whistler mode at 120km is shown in Figure 8. A ducted 248 propagation is adopted at this L shell (Clilverd et al., 2008)

314
The change trend of SNR is the same in other revisit period both in 2010 and background time. In sum, it can be 315 concluded that the SNR over the epicenter of Yushu earthquake decreases abnormally in one satellite revisit 316 period before the earthquake, especially in the southwestern region of the earthquake, which is consistent with 317 the observed TEC anomaly before the earthquake. The decrease of SNR before the Yushu earthquake may be 318 due to the enhancement of electron density.

319
The electron density in the lower ionosphere may change abnormally before earthquake through some 320 coupling mechanisms. The full wave simulation result on NOV, which is the nearest transmitter next to Yushu 321 earthquake, indicates that the electric field at the altitude of satellite will change when we add a disturbance on 322 electron density in the lower ionosphere. That is to say that the SNR of electric field will also change when the 323 background noise is considered to be invariable a few days before the earthquake. The simulated results show 324 that the SNR of electric field will decrease with the increase of electron density in the lower ionosphere; the 325 SNR will increase with the decrease of electron density in the lower ionosphere. It can be concluded that the 326 variation of electron density before earthquakes may be one important factor influence the variation of SNR.