Detection and analysis of short-period geomagnetic perturbations during increased solar activity and magnetic storms

The dynamics of geomagnetic field variations on the eve and during the periods of magnetic storms of 2015 and 2017 has been studied (data of the horizontal component of the Earth’s magnetic field of the terrestrial station network were used). The method developed by the authors based on wavelet transform and adaptive threshold functions was applied. Weak geomagnetic disturbances synchronously appearing at the stations and preceding the onset of strong magnetic storms was extracted. The correlation of the isolated geomagnetic disturbances with the AE-index is shown, both in the occurrence times and in intensities. On the basis of comparison with interplanetary environment data (interplanetary magnetic field data and the solar wind parameters were analyzed), and also based on the results of other author works [1, 2]. we assume that the isolated effects have solar nature. The research is supported by the grant of the Russian Science Foundation No. 14-11-00194.


Introduction
The work is aimed at developing methods for analyzing geomagnetic data and studying the processes in the magnetosphere during perturbed periods. The complexity of processing and analysis of geomagnetic data is associated with their complex nonstationary structure, with the presence of local features of different amplitude and duration. These features contain important information about the occurring processes in the magnetosphere. The use of traditional methods and approaches does not allow us to research the rapidly-variable structure of geomagnetic field variations in detail and leads to the loss of meaningful information. One of the most effective modern methods of data analysis is the wavelet transform [3][4][5][6][7][8].The wavelet transform is now successfully used for the tasks of denoising the geomagnetic data [4,6,8], extraction of the periodic components caused by the Earth's rotation [4 ,8], search for the precursors of intense solar flares [3], automatic detection of magnetic storm development [8], studying of the characteristics of solar daily variations based on data from ground-based magnetic stations [7], automation of geomagnetic activity indices calculation such as K-index [9,10], Dst-index and wavelet based index of storm activity «WISA» [4,8] as well as several other issues. This mathematical apparatus is taken as the basis for the study.
Earlier, the authors in the works [9,10] proposed a geomagnetic field variation model (FVM) based on wavelets. It allows describing both the characteristic variations and nonstationary short-period changes characterizing fast processes in the magnetosphere. On the basis of the FVM, computational algorithms for isolation and evaluation of local variations [11,12] have been developed, which allow us to study short-period small-scale (from a few seconds to tens of minutes) geomagnetic field parameters in detail. Application of these algorithms allowed us to extract anomalous changes in field variations (the data of meridionally located stations was used) arising on the eve and during magnetic storms [11][12][13]. This study is a continuation of the previous works. The paper presents the results of data processing and analysis from a geomagnetic station network in the northeast of Russia (Yakutsk «YAK», Magadan «MGD», Paratunka «PET», Khabarovsk «KHB») and from an equatorial station (Guam «GUA», USA). The results of the work confirmed the possibility of synchronous appearance of weak geomagnetic perturbations preceding the onset of strong magnetic storms. On the basis of data comparison with of the interplanetary environment parameters (the interplanetary magnetic field data and the solar wind parameters were analyzed), and also based on the results of other author works [1,2], we assume that isolated effects have solar nature.

Description of the method
The geomagnetic field variation can be represented as a combination of functions [9,14]: where the component (see Equations (1)) is described in [9,14]. The results of the method application for the «Paratunka» station (Kamchatka Region) data are given in the works [9][10][11][12], which show that On the basis of the component trend f , the authors developed a method for calculating the geomagnetic activity indices K, which makes it possible to reproduce the method of J. Bartels for the first time in an automatic mode [9,10]. In this work, the component trend f is not used.
To identify the set of indices I defining the model component pert f (see Equations (1)), a criterion was proposed [15]: , m is the sample average, v is the index of perturbed field variation, k is the index of quiet field variation,  is the some small positive number.
Assuming k j A is normally distributed with some mean (2) where U -is the threshold coefficient,

Experimental results and discussion
In the work, the geomagnetic data of minute resolution from the Russia's Northeast stations network and from the equatorial station GUA were processed (see Table 1 and Figure 1). In order to analyze geomagnetic perturbations in the auroral zone, we used the index of auroral electrojet (AE) (http://isgi.unistra.fr). Calculation of the AE index is based on the data of stations located in auroral and subauroral latitudes [17]. In order to analyze the equatorial current system, we used the Dst index (http://wdc.kugi.kyoto-u.ac. jp/dst_final/index.html), which is calculated using the data from the stations located near the Equator [18]. The results of our analysis were compared with the data of interplanetary magnetic field (IMF) and the solar wind parameters (http://www.srl.caltech.edu/ACE/ASC/index.html). The results of data processing during the magnetic storm on May 27, 2017 and July 16, 2017 are presented below in detail.   The second analyzed event on July 16, 2017 (see Figure 3) was caused by the CME on July 14 (the catalogue of ICMES by I. Richardson and H. Cane, http://www.srl.caltech.edu/ACE/ASC/DATA/level3/icmetable2.htm). On the eve of the storm, the solar wind speed was below the mean (< 360 km/s) [22], the IMF Bz component changed within +/-3nT. Analysis of the results of geomagnetic data processing shows that weak geomagnetic disturbances were observed at the analyzed stations on the eve of the event, during AE index increase (Figure 3c) (operation (5) (Figure 3 e). During the initial phase of the storm, the Dst-index increased to 51 nT (Figure 3 d). The auroral activity also increased to 700 nT (Figure 3 c). Note that during the main phase of the storm, the geomagnetic field variations structure differed at the high-latitude (YAK), mid-latitude (MGD, PET, KHB) and equatorial GUA stations that is probably due to their location. It can also be noted that within a few hours (approximately 10 hours) before the storm, the IMF Bz component turned to the south and small variations appeared in it (within +/-4 nT). At the same time, the auroral activity slightly increased (AE-index increased to 110 nT). During the same period, short-term insignificant increases in geomagnetic activity at the equatorial station GUA were observed (the time period in Figure 3 f is shown by dashed lines: on July 15 from 20:00 to 16 June, 02:30 UT). This gives us grounds to assume the connection of the detected perturbations with solar activity and auroral processes. Table 2 presents the processing results of the data obtained during periods of strong magnetic storms in 2015 and 2017. Processing details of the events on January 7, 2015 and March 17, 2015 are presented in [11][12][13]. Analysis of the results in Table 2 confirms the possibility of the appearance of weak short-period geomagnetic field perturbations on the eve of magnetic storms. These results agree with the results of the papers [1,2] where it was shown that increases in solar wind parameters and the following increases of geomagnetic activity (AE, Kp indices) can be observed prior to abrupt turns of IMF towards south which futher initiate magnetic storms [23].
The possibility of such anomalous effects was also shown earlier in [5,24] and noted in [15]. This allows us to make an assumption about their solar nature, and determines the applied significance of the study.

Conclusions
The processing and analysis of geomagnetic field data from meridianally located stations for the periods of magnetic storms of 2015 and 2017 have been performed. The results of the study showed the general nature of the processes in the analyzed locations. During the periods preceding the beginning of magnetic storms, weak geomagnetic perturbations synchronously appearing at the stations and correlating with the AE-index, both in occurrence times and in intensity, were detected. On the basis of comparison of the results with the interplanetary environment data (the interplanetary magnetic field data and the solar wind parameters were analyzed), with the results of other works [1,2], we can assume the connection of detected pre-storm geomagnetic perturbations with solar activity and auroral processes. Perturbations of maximum intensity were observed during the main phase of the storm, and the moments of their appearance correlated with the moments of the AE-index maxima. The largest perturbation amplitudes were observed at the northern station YAK.
The results of the experiments showed high sensitivity of the algorithms used and the possibility of its application for a detailed study of the dynamics and spatial-time distribution of geomagnetic disturbances. In the future, the authors plan to continue the study with the increase in statistics and the number of stations analyzed.