Comparing Ionospheric MUF using IRI16 Model with Mid-Latitude Ionosonde Observations and Associated with Strong Geomagnetic Storms

High frequency (HF) radio wave propagation depends on the ionosphere status which is changed with the time of day, season, and solar activity conditions. In this research, ionosonde observations were used to calculate the values of maximum usable frequency (MUF) the ionospheric F2layer during strong geomagnetic storms (Dst ≤ -100 nT) which were compared with the predicted MUF for the same layer by using IRI-16 model. Data from years 2015 and 2017, during which five strong geomagnetic storms occurred, were selected from two Japanese ionosonde stations (Kokubunji and Wakkanai) located at the mid-latitude region. The results of the present work do not show a good correlation between the observed and predicted MUF values for F2layer during the selected events of strong geomagnetic storms at these stations. Thus, there is a further need to improve the IRI-16 model for better matching with the observations during strong geomagnetic storms.


1.
Introduction The HF (3-30 MHz) radio wave is very important in communications because it has low cost and can be sent over long distances, with stations that are easy to set up. However, the propagation of this wave through the medium, such as the ionosphere layers, becomes complicated because of issues of absorption and interference [1][2][3]. The F2 layer of the ionosphere is responsible for radio communications due to its presence throughout the day. Therefore, this layer carries approximately all nighttime radio wave propagation all over the world for long distances, depending on the height of the layer [4,5]. There are daily, seasonally, and annually changes in the ionization of this layer, leading to variations in electron density (Ne) and its height. Also, the ionization in this layer varies with the geographic coordination, depending mainly on the latitude [6,7]. Several researchers have made attempts to explore the relationship between solar activity and ionosphere parameters [8][9][10]. Other researchers studied the ionosphere conditions with geomagnetic storms. Lakshmi et al., in 1997, found a rapid collapse in midnight ionospheric F layer electron density during severe geomagnetic storms [11]. Kouris and Fotiadis [12][13][14] conducted many types of research related to daily and hourly variability of ionospheric parameters and their variation with the latitude.
Najat, in 2009, compared some observed ionospheric parameters with the predicted values obtained from the international reference ionosphere (IRI) model for Japan's mid-latitudes region and reported no correlation relations [15]. Kotova et al., in 2016, studied the ionospheric variation with space weather changes and found an influence of stratosphere heating on the radio wave propagation, leading to attenuation in the HF signal in the daytime for the ionospheric equator region [16]. An international project by the committee on space research (COSPAR) was released to develop the IRI model output. The cross-correlation error represents the difference between a predicted value (P) and an observed value (O). The statistics used in this study include Spearman's rank correlation coefficient (r s ), expressed in equation (2). This coefficient has a value between 1 and -1. A value of 1 indicates a perfect positive correlation, a value of -1 indicates a perfect negative correlation, and a value near zero indicates poor correlation. The equation for r s is given below [33,34] OO PP In the above equations, n is the total number of pairs (predicted and observed) of values, and (i) refers to a specific value within a group.

Data selection and analysis
The primary source of observations data in this research is the ionosonde stations in Japan site (http://wdc.nict.go.jp) for ionosphere parameters. These are the long-range data recorded for the hourly and daily values of foF 2 Figure-1 shows the data of the hourly geomagnetic storm index Dst (nT) for the two years of study. A description of the five events that occurred during these years is provided in Table-2. Figure-2 presents data of the solar cycle 24, during which the two years selected for this study occurred.

Results and Discussion
In the initial step, attempts were made to calculate the values of MUF from observations, which were verified with the predicted MUF values using IRI-16 model (http://omniweb.gsfc.nasa.gov/for/dxl.html). Figures-(3-7) show the hourly predicted (blue) and observed (red) MUF values, using the hourly Dst-index, for the time intervals of two days before, during, and two days after the storm. Since the values of MUF depend mainly on those of foF2, as confirmed by their direct relation in Equation 1, the observations demonstrated that their day values were greater than night values. The propagation of HF varied through the seasons. The atmosphere was denser and colder in winter, e.g. December, as observed in event 3. This makes the height of the ionosphere layers lower to the Earth, with greater ionization or higher electron density (Ne). Also in winter, the Sun is closer to the Earth in the daytime, which causes more ionization and higher electron density in the ionosphere. Because of that, the daytime value of MUF in December was higher than that in the other months during the selected events. At night time, the recombination became faster in the F2 layer, causing a decrease in the values of MUF. Because of longer evenings in winter (December), the F2 layer has a longer time to lose its ionizations. When the day approached the sunrise and before dawn, the MUF fell to its lowest value (5 MHz) or lower in the quiet conditions. By observing the MUF values of 2015 and 2017, no clear difference was found, because the two years are in the descent of the solar cycle 24, as shown in Figure-2. In the minimum solar cycle (minimum sunspot), the Sun's chromosphere became quiet and the UV emissions became low, causing a decreased ionization of the ionospheric F2 layer and, accordingly, it's a decreased value of MUF. The results show a mismatch between observed and predicted values of MUF during the day, in which the strong geomagnetic storms occurred, and for the two latitudes selected (Kokubunji and Wakkanai). We also found non-linear relations between the observed and predicted values along the five events taken. For comparison, a statistical method was applied, which is the most important step in this research because it showed the validity or efficiency of the IRI-16 model in the mid-latitude region during strong geomagnetic storms. By using equation (2), the cross correlation between the observed and predicted values calculated for the five events and two latitudes indicated that the best correlation appears in events 2 and 3, reaching approximately 0.9. However, the lowest correlation was recorded in events 1 and 4. Therefore, it is necessary to improve the model to match the observations during strong geomagnetic storms.

Conclusions
The purpose of using the theoretical model (IRI) is to secure HF communication if it is blacked out, especially during geomagnetic storms, in which the ionospheric F2 layer is disturbed. From the results of the model, we can conclude that there is no linear relation between the observed and predicted values of MUF during strong geomagnetic storms and for the two mid-latitudes selected. The statistical cross-correlation between predicted and observed MUF values shows a low correlation. Thus, it is necessary to enhance the model to be suited with the observations during storm time.