Low Solar Activity Years Research Articles

Comparison of Performance of the IRI 2016, IRI‐Plas 2017, and NeQuick 2 Models During Different Solar Activity (2013–2018) Years Over South American Sector

AbstractThis study mainly focuses on the prediction of the variability of the vertical total electron content (VTEC) and validation of the new versions of the International Reference Ionosphere (IRI 2016), IRI extended to the plasmasphere (IRI‐Plas 2017), and NeQuick 2 models over the equatorial anomaly region during 2013–2018. The Global Positioning System (GPS)‐derived VTEC data inferred from the South American equatorial region during the relatively high (2013–2014), descending (2015–2016), and low (2017–2018) solar activity years have been considered as measurements to validate the models. The modeled (IRI 2016, IRI‐Plas 2017, and NeQuick 2) VTEC values are generally larger than the GPS‐derived VTEC values, with the highest overestimation being illustrated by the IRI‐Plas 2017 model. Small underestimations are also observed in employing the IRI 2016 and NeQuick 2 models, especially when the solar activity increases. The highest root‐mean‐square Deviations (RMSDs) reaching up to 10 TEC units (TECU) is observed in some hours while using the IRI‐Plas 2017 model during the high solar activity years. RMSDs less than 5 TECU are also observed on most of the hours while utilizing the model during all solar activity years. However, RMSD values less than 2 TECU (while using the NeQuick 2 model) and less than 1 TECU (while using the IRI 2016 model) are seen on most of the hours during 2013–2018. This shows that the performance of the IRI 2016 and NeQuick 2 is better than the IRI‐Plas 2017 with the IRI 2016 model consistently revealing the best in all solar activity years.

Semidiurnal Tidal Influence on the Occurrence of Postmidnight F Region FAI Radar Echoes

AbstractThe Equatorial Atmosphere Radar observations at Kototabang (0.2°S, 100.3°E) are used to study the possible semidiurnal tidal influence on the occurrence of postmidnight echoes from the field‐aligned irregularities (FAIs) due to spread F. It is found that the postmidnight FAI echoes show high percentage of occurrence (PO) during June–July and low PO in December–January of low solar activity years. As solar activity approaches minimum, the PO increase is extended to May, August, and September. The space‐time Fourier analysis on the temperature information obtained from the Sounding of Atmosphere by Broadband Emission Radiometry instrument onboard Thermosphere Ionosphere Mesosphere Energetics and Dynamics satellite reveals that at the low‐latitude upper mesospheric heights, though the migrating diurnal tide propagating westward with zonal wavenumber (k) 1 (DW1) dominates the migrating semidiurnal tide propagating westward with k = 2 (SW2) particularly during the equinox months, the SW2 tidal amplitudes are larger than DW1 during June–July. The postmidnight FAI echoes appear to be more frequent during the sudden stratospheric warming of the Solar Minimum Years 2018–2019 and 2008–2009 but not in the Solar Maximum Year 2013. The dominance of SW2 over DW1 is also noted during these years. This indicates that the SW2 plays a major role in the occurrence of postmidnight spread F. It is suggested that the dominant semidiurnal variation of zonal electric field can become eastward and lift the F layer to higher heights during midnight hours, which can favor the growth of the Rayleigh‐Taylor instability to cause delayed spread F around midnight.

Spatial Characteristics on the Occurrence of the Nighttime Midlatitude Medium‐Scale Traveling Ionospheric Disturbance at Topside Ionosphere Revealed by the Swarm Satellite

AbstractUsing 6 years collecting of electron density (Ne) data from the Swarm A satellite, this study portrays comprehensive maps of the occurrence of medium‐scale traveling ionospheric disturbance (MSTID) at middle latitudes, which is characterized by the in situ absolute Ne fluctuations above a certain threshold. Two interesting spatial preferences on the occurrence of MSTID as well as their dependence on local time and solar cycle are captured, that is, (1) regional preference near Weddell Sea Anomaly (WSA) to its west during December solstice, MSTID in this region appears even at very low latitudes (at least to 10°N in magnetic latitude) and (2) preference at middle latitudes with a wide zonal extension, which is similar to the recently disclosed nighttime midlatitude plasma band‐like enhancement (NMPBE) of the background plasma density (Zhong et al., 2019, https://doi.org/10.1029/2018JA026059). The first preference of MSTID occurring near WSA region can be understood as the effects of horizontal neutral wind contribution to the Perkins instability, with the principal requirement of sporadic E occurrence via ionospheric E‐F region coupling. The fundamental mechanism of the second preference of MSTID occurring includes the interhemispheric coupling that the MSTID can also develop in the winter hemisphere where the local growth rate of Perkins instability is low. Besides, the presence of NMPBE during low solar activity years sets up a favorable environment for the appearance of MSTID with considerable fluctuation amplitude.

Evaluation of ionospheric and solar proxy indices for IRI-Plas 2017 model over the East African equatorial region during solar cycle 24

This study evaluates the performance of Ionospheric and Solar proxy indices for the International Reference Ionosphere extended to the Plasmasphere (IRI-Plas 2017/SPIM) model in predicting the Total Electron Content (TEC) over the East African equatorial region during the high (2014) and low (2018) solar activity periods of solar cycle 24. TEC is derived from Global Positioning System (GPS) installed at Addis Ababa (adis, Geog: 9.04° N; 38.77° E and Geom: 5.28° N; 112.6° E) and Ambo (aboo, 8.99° N; 37.81° E and Geom: 5.43° N; 111.6° E). The results show that the SPIM model with all options of proxies overestimates the observed TEC for all seasons at both stations during the low solar activity year of 2018. However, in some cases, the SPIM model with IG index option showed a better agreement during the low solar activity period, while it displayed poor performance at both stations during the high solar activity period. The SPIM model with SSN proxy option was better in estimating seasonal and annual VTECs during the high solar activity period, while it was poor during low solar activity period over the study region. Results also showed that the SPIM model with Lyman-α proxy option performed poor for the seasonal and annual prediction of VTECs with maximum RMSE of 19 TECU during the high and low solar activity years. A good agreement is observed between the observed-VTEC and SPIM model with SSN proxy option. The correlation coefficients (r) between observed-VTEC and model output using SSN as proxy was good (r = 0.9566 at adis and 0.9494 at aboo stations) during the high solar activity year of 2014. Moreover, the model with all proxy options displayed a strong correlation (r > 0.96 at adis and r > 0.95 at aboo stations) during low solar activity year. Overall, the results in this study revealed that the SSN proxy option of the SPIM model was better in estimating VTEC during high solar activity period, while SPIM with the IG index was better during low solar activity period over the East African equatorial region. Such performance evaluation of a global model for different inputs/options is important for the space weather community at the study region where little or no ground based instruments are found.

TEC Map Completion Using DCGAN and Poisson Blending

AbstractBecause of the limited coverage of global navigation satellite system (GNSS) receivers, total electron content (TEC) maps are not complete. The processing to obtain complete TEC maps is time consuming and needs the collaboration of five international GNSS service (IGS) centers to consolidate final completed IGS TEC maps. The advance of deep learning offers powerful tools to perform certain tasks in data science, such as image completion (or inpainting). Among them, deep convolutional generative adversarial network (DCGAN) is capable of learning the properties of the objects and recovering missing data effectively. With years of IGS TEC maps for training, the combination of DCGAN and Poisson blending (DCGAN‐PB) is able to effectively learn the completion process of IGS TEC maps. Both random and more realistic masks are used to test the performance of DCGAN‐PB. The results with random masks (15–40% missing data) show that DCGAN‐PB can achieve better TEC map completion than DCGAN alone, and more training data can significantly improve its generalization. For the cross‐validation experiment using the realistic mask from Massachusetts Institute of Technology (MIT)‐TEC data (~52% missing data), DCGAN‐PB achieves the average root mean squared error about three absolute TEC units (TECu) for high solar activity years and less than two TECu for low solar activity years, which is about 50% reduction of means and more than 50% reduction on standard deviations compared to two conventional single‐image inpainting methods. The DCGAN‐PB model can lead to an efficient automatic completion tool for TEC maps by minimizing the manual work.

Validation of International Reference Ionosphere model (IRI-2016) for F-region peak electron density height (hmF2): Comparison with Incoherent Scatter Radar (ISR) and ionosonde measurements at Millstone Hill

In this report, we have evaluated the performance of International Reference Ionosphere (IRI-2016) model in predicting the diurnal variation of F-region peak electron density height (hmF2) at Millstone Hill (MH, 42.6°N, 288.5°E) during the solar cycle 24 for both high and low-solar activity years. The new methodology is adopted to derive hmF2 values from the Incoherent Scatter Radar (ISR) measurements. The hmF2 derived from ISR along with ionosonde measurements operating at MH has been used to evaluate the performance of IRI-2016 model. The IRI-2016 model has three options to predict hmF2, namely, BSE-1979, AMTB-2013, and SHU-2015. The observed hmF2 from ISR and ionosonde are compared with all the three options provided by IRI-2016 to predict hmF2. Results show that hmF2 from IRI-2016, ISR and ionosonde exhibit similar diurnal variation. The deviations between measurements and model predictions were found to increase during nighttime. The SHU-2015 model emerged out to be the best in predicting the hmF2 values over MH as compared to other two options. The BSE-1979 option also provides reasonably better prediction as compared to the relatively latest AMTB-2013 options. Based on this statistical analyses, we recommend SHU-2015 option of IRI-2016 hmF2 model over MH area.

Pattern of the variation of the TEC extracted from the GPS, IRI 2016, IRI-Plas 2017 and NeQuick 2 over polar region, Antarctica

As a result of the complex and typical behavior of the earth's magnetic field, the high latitude ionospheric parameters (such as the Vertical Total Electron Content, VTEC) are very dynamic. Thus, this study mainly focuses on the pattern of variation of the VTEC inferred from the GPS, IRI 2016, IRI-Plas 2017 and NeQuick 2 with different height over Antarctica during 2011–2017. It has been shown that the GPS-derived (GPS VTEC) and modelled (IRI 2016, IRI-Plas 2017 and NeQuick 2) VTEC do not show significant variations in the hourly values. This shows that the solar irradiance mostly responsible for the ionization of neutral molecules in the earth's upper atmosphere does not show significant hourly variation in the region. Moreover, the GPS VTEC values are larger than the modelled values on most of the hours with the highest Root-Mean-Square Deviations, RMSD (close to 14 TECU) resulting from underestimation of the VTEC by the IRI 2016 model being observed in the March equinox in 2014. On the other hand, relatively small RMSDs (less than 1 TECU) are observed on most of the hours in using all models, demonstrating that the models’ performance is good with the NeQuick 2 model showing the best followed by the IRI-Plas 2017. Similar and smooth VTEC variation patterns are also observed in the bottom-side, topside and plasmaspheric layers during both the high and low solar activity years. This shows the TEC variation does not have significant impact on the signal propagated through the ionosphere in the polar region. Besides, the highest and lowest contributions of the TEC have been seen in the topside and plasmaspheric layers, respectively. The topside TEC contribution generally increases when the solar activity (or solar irradiance) increases; while, the plasmaspheric TEC contribution enhances when the solar activity (or solar irradiance) drops. Moreover, the largest bottom-side contribution is observed in the December solstice during 2014–2017.

NIGHTTIME TEC VARIATION ANALYSIS OF KLOBUCHAR MODEL BASED ON IGS DATA IN CHINA

Abstract. In order to further study the variation characteristics of the nighttime ionospheric accuracy of the Klobuchar model and the correlation between the nighttime ionospheric accuracy of the Klobuchar model and solar activity intensity. In this paper, a method for comparative analysis on solar activity intensity in the high, medium and low solar activity years has been proposed. The GIM provided by the IGS Center from 2008 to 2018 are treated as reference values to analyze the nighttime accuracy of the Klobuchar model for annual, seasonal, and monthly changes in high solar activity, medium solar activity and low solar activity. The results are: (1) The nighttime accuracy of the Klobuchar model is best in the medium solar activity, followed in the low solar activity and the worst in the high solar activity. (2) The Bias and RMS of the Klobuchar model at nighttime had obviously semi-annual and seasonal variations in high and medium solar activity. (3) The nighttime accuracy of the Klobuchar model did not vary with Bias and RMS changes at nighttime.

Time of evening zonal drift reversal in F- region and its implications for post sunset ionosphere

The study analyses the delayed occurrence of the F- region ionospheric manifestations like peak vertical drift and Equatorial Spread F onset in summer solstice of low solar activity years, using the ionosonde data at the magnetic equatorial station Trivandrum (8.5° N, 76.9° E) and Republic of China Satellite-1 data. The analysis reveals the unambiguous role of the time of ionospheric F- region zonal plasma drift/wind reversal in modifying the above-mentioned phenomena. The direct control of equatorial Spread F onset by the F- region drift/wind reversal is clearly established on a day-to-day basis in summer season. The role of the time of zonal wind/drift reversal in modulating the ESF onset is examined quantitatively by exploring its effect on the growth rate of Rayleigh - Taylor instability. Further, the study brings out the solar activity variation of the time of F- region zonal plasma drift/wind reversal for all seasons. The results highlight the need to have systematic neutral wind measurements to better address the enigmatic and multi-dimensional problem of equatorial Spread F day-to-day variability.

Singular spectrum analysis of GPS derived ionospheric TEC variations over Nepal during the low solar activity period

Accurate modeling of ionospheric total electron content (TEC) is an important aspect for mitigating the threats of trans-ionospheric delay error in satellite communication, earth observation, space-based navigation, timing applications as well as space weather forecasting services. In recent years, singular spectrum analysis (SSA) has been proved to be a powerful technique giving a relatively accurate estimate in time-series analysis comparable to the contemporary methods. In the current study, the SSA has been implemented on the GPS-derived TEC during the low solar activity year of 2017 over Nepal region which locates itself almost in the vicinity of low-latitudes being sandwiched between India and Tibet, China. The country foresees an explicit investigation and modeling of ionospheric TEC variations and corresponding delay error to precisely accomplish the space-based trans-ionospheric applications. The semi-annual variability of TEC with higher magnitudes during equinoctial seasons and lower values during solstice seasons is clearly noticed in the diurnal plots which are further substantiated by the trajectory matrix of time-series. The decomposed modes in the principal component analysis (PCA) signifies diurnal (first), semidiurnal (second), semiannual (third), monthly (fourth) with higher orders representing associated noise errors in the signals. Correlation coefficients (CC) between the reconstructed and observed time-series demonstrates the SSA method could be a successful tool for forecasting the TEC series over the region. The results are compared with empirical global ionospheric maps (GIMs) and IRI-Plas 2017 models during different seasons, emphasizing the suitability of SSA technique for relatively better precise TEC forecasting over the region.

Spatio-temporal analysis of ionospheric disturbances for ground based augmentation systems over a midlatitude region

Forcings from above and below the ionosphere can cause disturbances that need to be detected and corrected for navigation systems. Ground Based Augmentation Systems (GBAS) are used to give corrections to aircraft navigation systems while landing. These systems use regional ionosphere monitoring algorithms to detect the anomalies in the ionosphere. The aim of this study is to understand occurrence of ionosphere anomalies and their trends over Turkey. A comprehensive analysis of spatio-temporal variability of ionosphere is carried out for a midlatitude GPS network using Slant Total Electron Content (STEC). Differential Rate Of TEC (DROT), which is a measure of the amount of deviation of temporal derivative of TEC from its trend, is used to detect and classify the level of such disturbances. The GPS satellite tracks are grouped into north, east, west and over directions. The 24 h is divided into six time intervals. The percentage occurrence of each DROT category and the deviation from STEC trend in magnitude are calculated and grouped into satellite track directions and time intervals for 2010 (low solar activity), 2011 and 2012 (medium solar activity). The highest level of disturbances is observed in north and west directions, and during sunrise and sunset hours. The dominant periods of percentage occurrences are diurnal (22–25 h), semidiurnal (12–13 h) and terdiurnal (8–9 h) followed by quasi two-day and quasi 16-day periods. Disturbances corresponding to 50% < DROT < 70% are mostly visible during low solar activity years with magnitudes from 1 to 2 TECU. Geomagnetic storms can cause aperiodic larger scale disturbances that are mostly correlated with DROT > 70%. In 2012, the magnitude of such disturbances can reach 5 TECU. The anisotropic and dynamic nature of midlatitude ionosphere is reflected in the spatio-temporal and spectral distributions of DROT, and their percentage occurrences. This study serves a basis for future studies about development of a regional ionosphere monitoring for Turkey.

Performance of NeQuick-2, IRI-Plas 2017 and GIM models over Ethiopia during varying solar activity periods

This paper investigates the performance of NeQuick-2, the latest version of International Reference Ionosphere IRI extended to the Plasmasphere (IRI-Plas 2017), and Global Ionospheric Map (GIM) models over four stations deployed in Ethiopia during medium, high, and low solar activity periods. The monthly and seasonal mean diurnal variations of Vertical Total Electron Content (VTEC) derived from NeQuick-2, IRI-Plas 2017 and GIM models were compared with the GPS-VTEC values in terms of Root-Mean-Square Errors (RMSE) and correlation coefficients (r). It was found that the monthly and seasonal mean diurnal VTEC variations agree fairly well with the GPS-VTEC at all stations although with some offsets. For the entire years of the study, the highest monthly mean VTEC variation is observed during equinox months, whereas minimum in the solstice months. The NeQuick-2 model monthly and seasonal mean VTEC output were generally better than IRI-Plas 2017 and GIM models output in most of the stations during medium solar activity year of 2012. For all models, the significant discrepancy (i.e, overestimations or underestimations) is observed during daytimes and seasons of high solar activities. In some cases, the IRI-Plas 2017 model displayed poor performance during medium and low solar activity periods, while it showed a better agreement for some stations (e.g. asos, armi, and bdmt) during high solar activity phase (especially in March equinox and December solstice). The performance of GIM model was good for most of the stations during low-solar activity year of 2016 and during the study years for adis, while it has poor performance during the periods of high solar activity year of 2014. All models correlated well, although GIM model was highly correlated with the GPS-VTEC value (r ≥0.99). The IRI-Plas 2017 model correlated better with the GPS-VTEC than NeQuick-2 model during the study years. Moreover, the performance of IRI-Plas 2017 model was better during high than low and medium periods of solar activity years. The GIM model has better performance for most stations (e.g. adis, asos, and bdmt) over the study region during low-solar activity period.

Modeling M(3000)F2 based on extreme learning machine

This paper presents an novel extreme learning machine (ELM)-based prediction model for the ionospheric propagation factor M(3000)F2 at Darwin station (12.4°S, 131.5°E; −44.5°dip) in Australia. The proposed ELM model is trained with hourly daily values of M(3000)F2 from the period 1998–2014 except 2001 and 2009. The hourly daily values of 2001 (high solar activity) and 2009 (low solar activity) are used for validating the prediction accuracy. The proposed ELM for modeling M(3000)F2 can achieve faster training process and similar testing accuracy compared with backward propagation neural network (BPNN). In addition, the performance of the ELM is verified by comparing the predicted values of M(3000)F2 with observed values and the international reference ionosphere (IRI −2016) model predicted values. Based on the error differences (the root mean square error (RMSE) and the M(3000)F2 percentage improvement values M(3000)F2IMP(%)), the result demonstrates the effectiveness of the ELM model compared with the IRI-2016 model at hourly, daily, monthly, and yearly in high (2001) and low (2009) solar activity years. The ELM also shows good agreement with observations compared with the IRI during disturbed magnetic activity.

Nonlinear dependence study of ionospheric F2 layer critical frequency with respect to the solar activity indices using the mutual information method

This paper analyzes the dependence of the ionospheric F2 layer critical frequency, foF2, on two widely used solar activity indices (12-month running mean of the sunspot number (R12) and the solar radio flux of 10.7 cm wavelength (F10.712)) by using mutual information method at the ionospheric station of Darwin (geographic 12.4°S, 131.5°E) in Australia. The results show that the foF2 have different responses to R12 and F10.712 in different solar activity years. In the high and low solar activity years, the foF2 is more dependent on R12 rather than F10.712. In the moderate solar activity years, the foF2 is more dependent on F10.712 rather than R12. We further verify the effectiveness of the mutual information on the selection of the solar activity index by developing the prediction models of the foF2. Additionally, the results are compared against the International Reference Ionosphere (IRI) model. The data of training period used in the model is from 1998 to 2015, except 2000, 2009 and 2013. Data for 2000, 2009 and 2013 are used to validate the prediction accuracy, for high, low and moderate solar activity years, respectively. The results demonstrate that the favorable prediction results can be obtained when the solar activity index with large mutual information is used as the input variable of the foF2 prediction model. The proposed method not only can quantitatively analyze the nonlinear relationship between the foF2 and solar activity indices but also provides an effective way of selecting the solar activity indices for the foF2 prediction.

Comparison of IRI-2016 model-predictions of F2-layer peak density height options with the ionosonde-derived hmF2 at the equatorial station during different phases of solar cycle

We have investigated the performance of 2016 International Reference Ionosphere (IRI) at equatorial Ouagadougou (12.4°N, 358.5°E; dip latitude: 1.5°N), Burkina Faso station, using 4 years 5 months of hourly monthly median M(3000)F2 values from January to May 1986 and from January 1987 to December 1990 using Shimazaki (1955) analysis technique. The deduced values of hmF2 are compared with the predictions of the three global IRI-2016 hmF2 model options for four different levels of solar activity with yearly averaged solar flux intensity F10.7 value of 85, 140, 214 and 190, corresponding to a year of 1987 solar minimum, 1988 moderate solar activity, 1989 very high solar activity and 1990 solar maximum, respectively. We find that the three IRI-hmF2 models: SHU-2015, AMTB-2013, and BSE-1979 reproduce the diurnal features of ionosonde-based hmF2 fairly well, showing daytime maximum and the prominent evening peak, except for solar cycle minimum where such feature is absent in Bilitza et al. (1979) and weakly expressed in Shubin (2015) hmF2 models. The magnitudes of the evening peaks in hmF2 for both data and models essentially correlate with solar activity but the occurrence times are independent of the phases of the solar cycle for all models. On average, the dayside maximum occurs between about noon and 1400 LT for both ionosonde-derived hmF2 and IRI models with typical values of 430 and 420 km, while the prominent evening peak in hmF2 occurs between about 1800–1900 LT with characteristic values of 390–450 km. We observed SHU-2015 persistently occur at about 1800 LT with a lower magnitude compared to ionosonde-based hmF2, the digisonde-based hmF2 model (AMTB), and Bilitza et al. (1979) model. We also show that SHU-2015 is consistently and significantly lower in values that the ionosonde hmF2, AMTB-2013 and BSE-1979 models. SHU-2015 deviates from data by about 2 times during low and median solar activity years but diverges by about 3–4 times during very and high solar maximum. Quantitative analyses indicate that the average absolute discrepancy between ionosonde-derived hmF2 and IRI model is 15%, 5.2% and 5.8% for SHU-2015, AMTB-2013, and BSE-1979, respectively. SHU-2015 model root-mean-squared error (RMSE) varies strongly with increasing solar activity with magnitudes roughly within the range ∼41–74 km, whereas AMBT-2013 and BSE-1979 RMSE vary between ∼19–36 km and ∼25–27 km, correspondingly. BSE-1979 RMSE does not exhibit any dependence on solar activity, while AMBT-2013 RMSE appears to decrease with an increase in solar activity. In contrast, SHU-2015 mean-relative-deviation (MRD) increases dramatically with increasing solar activity with values between ∼10–16%. AMTB-2013 and BSE-1979 both indicate a slight decrease with increasing solar activity, with values ranging between ∼3.3–8.9% and ∼4.8–6.7%, respectively. On the basis of evidence provided in this work, we conclude that AMTB-2013 and BSE-1979 have higher accuracy than SHU-2015. Although the hmF2 calculated from ionograms and from empirical formula of Shimazaki could differ rather substantially from the real hmF2 values, especially in the equatorial region. The results of this study show that the AMTB-2013 and BSE-1979 models describes better the hmF2 obtained from the M(3000)F2 values, consequently, we recommend AMTB-2013 and BSE-1979 for equatorial region over West African longitude sector.

RESPONSE OF F2- REGION MAXIMUM ELECTRON DENSITY (NmF2/foF2) TO SOLAR ACTIVITY USING PEARSON PRODUCT MOMENT CORRELATION

Introduction: The ionosphere owes its origin primarily to ultraviolet radiation from the Sun. The ionosphere is an essential part of the Earth’s upper atmosphere. It is ionized by solar radiation and influences transionospheric radio wave propagation. Maximum electron density of the F2- layer (NmF2) is an important parameter for studying the ionosphere. The ionospheric F2-region maximum electron density (NmF2) depends strongly on solar activity, it also suffers temporal and spatial variations. Aim: The aim of this paper is to investigate the response of NmF2 to solar activity during high solar activity (HSA), moderate solar activity (MSA) and low solar activity (LSA) years using correlation analysis. Materials and Methods: The data used in this work are the hourly NmF2 values derived from foF2 data observed at Jicamarca (Lat.11.9 oS, Long.76.8 oW) and Puerto Rico (Lat.18.5 oN, Long.67.2 oW) during high solar activity HSA (2002), moderate solar activity MSA (2011) and low solar activity LSA (2006) years. The NmF2 data were evaluated using the relation in equation 1 NmF2 = 1.24 x 1010 (foF2)2 (1) Where NmF2 is in el/m3 and foF2 is in MHz. Pearson Product Moment Correlation (PPMC) was used to further analyse the NmF2 data. Results: Our results revealed two unequal NmF2 peaks. The NmF2 peaks values at Jicamarca (60 - 240; 63– 204) x 1010 el/m3 are observed to be higher in values than those at Puerto Rico (63 – 187; 57 – 164) x 1010 el/m3. The highest NmF2 peak values of 240 and 187x 1010 el/m3 occurred during March equinox at 09:00 and 14:00 hours at Jicamarca and Puerto Rico respectively during HSA year. Conclusion: Correlation analysis for the three epochs of solar activity revealed that NmF2 showed positive correlation with sunspot number with highest correlation values of 0.904 and 0.976 at Jicamarca and Puerto Rico stations respectively during MSA year.

Morphology of GPS and DPS TEC over an equatorial station: validation of IRI and NeQuick 2 models

Abstract. We investigated total electron content (TEC) at Ilorin (8.50∘ N 4.65∘ E, dip lat. 2.95) for the year 2010, a year of low solar activity in 2010 with Rz=15.8. The investigation involved the use of TEC derived from GPS, estimated TEC from digisonde portable sounder data (DPS), and the International Reference Ionosphere (IRI) and NeQuick 2 (NeQ) models. During the sunrise period, we found that the rate of increase in DPS TEC, IRI TEC, and NeQ TEC was higher compared with GPS TEC. One reason for this can be attributed to an overestimation of plasmaspheric electron content (PEC) contribution in modeled TEC and DPS TEC. A correction factor around the sunrise, where our finding showed a significant percentage deviation between the modeled TEC and GPS TEC, will correct the differences. Our finding revealed that during the daytime when PEC contribution is known to be absent or insignificant, GPS TEC and DPS TEC in April, September, and December predict TEC very well. The lowest discrepancies were observed in May, June, and July (June solstice) between the observed values and all the model values at all hours. There is an overestimation in DPS TEC that could be due to extrapolation error while integrating from the peak electron density of F2 (NmF2) to around ∼1000 km in the Ne profile. The underestimation observed in NeQ TEC must have come from the inadequate representation of contribution from PEC on the topside of the NeQ model profile, whereas the exaggeration of PEC contribution in IRI TEC amounts to overestimation in GPS TEC. The excess bite-out observed in DPS TEC and modeled TEC indicates over-prediction of the fountain effect in these models. Therefore, the daytime bite-out observed in these models requires a modifier that could moderate the perceived fountain effect morphology in the models accordingly. The daytime DPS TEC performs better than the daytime IRI TEC and NeQ TEC in all the months. However, the dusk period requires attention due to the highest percentage deviation recorded, especially for the models, in March, November, and December. Seasonally, we found that all the TECs maximize and minimize during the March equinox and June solstice, respectively. Therefore, GPS TEC and modeled TEC reveal the semiannual variations in TEC.

A Prediction Model of Ionospheric foF2 Based on Extreme Learning Machine

AbstractThe highly nonlinear variation of the ionospheric F2 layer critical frequency (foF2) greatly limits the efficiency of communications, radar, and navigation systems that employ high‐frequency radio waves. This paper proposes an effective method to predict the foF2 using the extreme learning machine (ELM). Compared with the previous neural network model based on feedforward algorithm, the ELM model offers the advantages of faster training speed and less manual intervention. The ELM model is trained with the daily hourly values of foF2 at Darwin (12.4°S, 131.5°E) in Australia. The training data are selected from 1995 to 2012, except 1997 and 2000, which includes all periods of quiet and disturbed geomagnetic conditions. The foF2 data to verify model performance are selected in 1997, 2000, and 2013, which are low, high, and moderate solar activity years, respectively. The prediction results have shown that the proposed ELM model can achieve faster training process while maintaining the similar accuracy compared with BPNN. In addition, the proposed ELM model is compared with the International Reference Ionosphere model prediction. The ELM model predicts the foF2 values more accurately than the International Reference Ionosphere model in low (1997), moderate (2013), and high (2000) solar activity years, as clearly seen on the yearly root‐mean‐square error. As far as the author's knowledge, this is the first time that the ELM model is applied to predict foF2.

Investigation on slab-thickness and B0 over an equatorial station in Africa and comparison with IRI model

The present study investigates the simultaneous morphologies of slab-thickness (τ) and thickness parameter (Bo) over Ilorin (8.500 N, 4.680 E; dip lat. 2.95), an equatorial station, during a year of low solar activity (2010). The τ is deduced from global positioning system total electron content (GPS-TEC) while the F2 peak electron density (NmF2), from digisonde portable sounder (DPS). The use of measured TEC for this type of investigation takes care of the inclusion of plasmaspheric electron content (PEC). The PEC distributions on the topside and bottomside electron density (Ne) profile add considerably to the genuine signatures of τ and Bo. Therefore, the dynamic contributions of the PEC need to be emphasized for the accurate prediction of the ionospheric models. Apart from daytime signatures of the τ and Bo which are not primarily influenced by PEC, we found that between two and 3 h of pre-sunrise and dusk periods are mainly controlled by PEC which manifests as huge peaks in the τ and Bo. Also, our investigation reveals that the Bo profile is thicker than the τ profile during the pre-sunrise periods in June indicating partial flow or halt in PEC. The result revealed approximately the same values of τ and Bo around the sunrise period suggesting the absence or negligible PEC contributions. Our investigation also shows that there are maximum and minimum of the thickness in Bo and τ during the December solstice and June solstice, respectively. We also observed a moderate sunrise enhancement in τ that is not conspicuous in the Bo signature. On the relationship between the τ and Bo, we found a significant association between τ and Bo with the highest coefficient value observed during the June seasons indicating the possibility of predicting τ in the absence of Bo and vice-versa, especially during the June season. The validation of International Reference Ionosphere (IRI) model with observed τ and Bo revealed appreciable discrepancies between the model and observed values, particularly between the IRI-τ and observed τ values. Our result revealed the underestimation of IRI-τ with respect to observed τ value in November and December.

The Response Time of Equatorial Ionization Anomaly Crest: A Unique Precursor to the Time of Equatorial Spread F Initiation

AbstractThe time delay with which the magnitude and location of the total electron content (TEC) at the equatorial ionization anomaly (EIA) crest responds to the mean height of the F layer (hmF2) is examined seasonally using the ionosonde data at Trivandrum and the Global Positioning System data at different stations in the EIA region for low and high solar activity years. The study brings out the fact that for the low solar activity year, the crest TEC responds fastest in winter solstice compared to other seasons. Further, the response time of the crest is found to decrease with solar activity. The seasonal variation in the EIA response time is attributed to the modulation by meridional neutral winds, whereas the solar activity variation seems to be basically controlled by the diffusion times and the background ionization gradients in the respective epochs. The derived relationship between EIA crest location/magnitude and hmF2 for any season can be used for prediction purposes. Furthermore, this study for the first time establishes the relation of EIA response time to equatorial spread F (ESF) start time, for days when the F layer is in the neutral dynamically controlled domain. The deleterious effects of ESF irregularities on communication and navigation systems are well known, and the above result is significant in this context. The present study has the potential to be extended into a model to predict the ESF start time, from the EIA response time, earlier in the day.