Investigating the Compatibility of IRI and ASAPS Models in Predicting the foF2 Ionospheric Parameter over the Mid Latitude Region

In this research, an investigation for the compatibility of the IRI-2016 and ASAPS international models was conducted to evaluate their accuracy in predicting the ionospheric critical frequency parameter (foF2) for the years 2009 and 2014 that represent the minimum and maximum years of solar cycle 24. The calculations of the monthly average foF2 values were performed for three different selected stations distributed over the mid-latitude region. These stations are Athens Greece (23.7 o E, 37.9 o N), El Arenosillo Spain (-6.78 o E, 37.09 o N), and Je Ju South Korea (124.53 o E, 33.6 o N). The calculated values using the two tested models were compared with the observed foF2 datasets for each of the three selected locations. The results showed that the two tested models gave good and close results for all selected stations compared to the observed data for the studied period of time. At the minimum solar cycle 24, the ASAPS model showed in general better values than the IRI-2016 model at Athens, El Arenosillo and Je Ju stations for all tested methods. At maximum solar cycle 24, the IRI-2016 model showed higher and closer values to the observed data at Athens and El Arenosillo stations, while the ASAPS model showed better values at Je Ju station.


Introduction
The ionosphere is one of the layers of the Earth's upper atmosphere, extending from about 50 km to 1000 km and higher, which constitutes less than 1% of the mass of the atmosphere that exceeds 100 km. The ionosphere is an electrically neutral layer which is ionized when solar radiation strikes the components of chemical substances to the atmosphere by displacing their electrons from atoms and molecules [1]. This process occurs on the illuminating side of the Sun towards the Earth. The shorter wavelengths of solar radiation (ultraviolet photons (EUV) and shorter X-rays) have sufficient energy to produce this ionization. The presence of these charged particles makes the upper atmosphere an electrical conductor that supports electric currents and affects radio waves [2]. According to the density of the electron and ionization, the ionosphere is classified into two main regions, the "topside region", which extends over the surface of the earth upwards from about 500 km to 1000 km, and the "bottom side region", which extends from 50 to 500 km above the surface of the earth. The bottom side of the ionosphere is divided into three specific regions according to the height and distribution of ions, which are regions D, E, and F. Each region is split into layers, called D, E, Es, F1, and F2 layers. The D layer, which extends over the surface of the earth approximately about 50-90 km, is mainly responsible for the partial absorption of high frequency radio waves [3]. The E layer is extending about 90-150 km. These layers can only reflect radio waves that have frequencies below 5 MHz [4]. Also, there is an unexpected layer, known as E-Sporadic (Es), with a height of 80 to 120 km [5,6]. One of the most ionized layers of the ionosphere is the F layer, and it usually ranges about 140-500 km. The light coming from the sun causes this layer to split into two distinct layers; F1, located at an altitude of 150-250 km, and F2, which is the highest layer of the ionosphere and is located at an altitude of 250-400 km [7]. In 2014, Hadi et al. studied the variation of the ionospheric critical frequency of the F2-layer (f o F2) over Athens city for the monthly period of the years 2011, 2012, 2013. An analytical investigation was conducted and a relationship between the monthly average f o F2 values and the hourly time factor was expressed as a suggested mathematical formula [8]. Jeon et al. (2016) used the mean and standard deviation to analyze the seasonal and annual changes of f o F2 as well as the relationships of F2 layer height at two sites in South Korea. The median and spring for the study of the ionosphere were used to ensure a more accurate analysis [9]. Mohammed (2016) studied the accuracy of predicting the hourly f o F2 values using IRI-2012 and VOACAP models for three Iraqi cities during high solar activity. The results indicated that the accuracy of them increases for all hours during Spring and Summer and decreases during Winter and Autumn, especially at hours near to sunrise. Both models were shown to have the same accuracy. f o F2 values predicted by VOACAP model were reported to be higher than those predicted by IRI-2012 model for all seasons [10].

Ionospheric Critical Frequency Parameter
The ionosphere is characterized by a set of different parameters. One of the most important parameters and the most frequently used is the critical frequency parameter, which is considered to describe the state of change of the ionosphere. If an ionospheric layer possesses a distinct maximum in ionization, a radio frequency capable of just penetrating to this height is called the critical frequency of the layer. The critical frequency is the maximum frequency of each layer of the ionosphere at which radio waves can be sent vertically and refracted back to the Earth. The foF2 is an important parameter for describing the state of ionospheric variation and defined as the highest frequency signal that will reflect directly back to its transmission location depending on the time of day and day of the sunspot cycle. It is related to the maximum electron density of F2 layer (NmF2), according to the following equation [11] [12]: where: f o F2: critical frequency of the F2 layer. N m F2: max. electron density of the F2 layer. e: electron charge. ε o : vacuum permittivity. M: mass of electron.

International Ionospheric Models
In this research, the Advanced Stand Alone Prediction System (ASAPS6) model and the International Reference Ionosphere (IRI-2016) model were selected as international models to verify the compatibility of the accuracy of predicting the ionospheric critical frequency parameter which will be generated using the adopted models with the observed data for Athens, El Arenosillo and Je Ju stations. IRI model defines the monthly averages of critical F2-layer frequencies in the existing ionosphere altitude range of 50Km to 1500km [13]. ASAPS provides forecasting of sky-wave communication system performance in the highfrequency (HF) radio spectrum (1 to 30MHz) and basic surface wave performance in the medium frequency (300kHz-3MHz) and low-HF (3-5MHz) range. It is based on the ionosphere model developed by the Space Weather Services and ITU-R/CCIR models [14].

Test and Results
In this work, a comparative study between the ASAPS and IRI-2016 models was conducted by investigating the compatibility of predicting the ionospheric critical frequency parameter generated using the two tested models for three different stations distributed on the midlatitude region during the maximum and minimum years of the 24 th solar cycle. The f o F2 of the F2 ionosphere layer was adopted to make a comparison between the two selected models. The values of the critical frequency parameters (f o F2) were calculated for each of the three selected sites using the two tested models and compared with the observed f o F2 data values for the monthly times variations of 2009 and 2014, which represent the minimum and the maximum for the years of the solar cycle 24. The monthly calculations of the critical frequency parameter for the selected locations were made according to the available observational data within the study period. The three tested locations that spread over the midlatitude zone are Athens (Greece), El Arenosillo (Spain), and Je Ju (South Korea), for which the location and geographical coordinates are illustrated in Figure 1 and Table 1.  Table 1-Geographical coordinates of the selected stations distributed within the Mid-latitude region The implementation of the IRI-2016 and ASAPS models needs several input parameters, including the monthly sunspot number (SSN) of the tested years. In this work, daily sunspot numbers were used to calculate the daily variation of the critical frequency parameter using the tested models. Table (2) presents a daily sunspot numbers for the tow adopted years (2009 & 2014).        Samples of the monthly statistical correlation results between the observed and predicted f o F2 ionospheric parameter values that were generated using IRI-2016 and ASAPS models are shown in Figure 3. Also, samples of the statistical analysis resulted for the statistical methods of Difference Residual and Absolute Residual between observed and predicted f o F2 dataare presented in Figure 4.

Conclusions
The results of the conducted study showed that the two tested models gave good and close results for all selected stations compared to the observed data for the studied period of time.