The ionosphere is an important atmospheric layer, lies between 100 to 1000 km above the Earth, is the greatest source of error for high-precision global navigation satellite system (GNSS) positioning. This error is measured in terms of total electron content (TEC) and its unit is known as TEC unit (TECU).
To deal with such type of problem, various research groups have attempted several forecasting models such as kriging, autoregressive moving average (ARMA) method, Holt-Winters method and empirical orthogonal functions etc. With the combined efforts of data from different parts of the globe, few more representative global numerical models have been made available like NeQuick, standard plasmasphere ionosphere model (SPIM), international reference ionosphere (IRI) and global ionosphere map (GIM) etc.
The present study examines the variations of TEC in Turkey from available 143 permanent GNSS stations network during the years 2009 and 2017. The corresponding vertical TEC (VTEC) predicted by kriging, ARMA, IRI-2012, IRI-2016, SPIM, GIM and NeQuick-2 models are tested to understand their efficacy over the country. The spatial, diurnal and seasonal activities of VTEC variations are modeled with the help of ordinary least square estimator (OLSE) using the long term of 9 years VTEC data.
The spatially modeled outcomes in latitude point out an inverse relationship (decreasing VTEC with increasing latitude) of VTEC during the whole year across the country while in longitude follows the same pattern during March equinox and June solstice only. The study confirms latitudinal gradient range of VTEC 0.1-0.2 TECU/degree in the daytime. The diurnal behavior of VTEC variation for the observed as well as modeled data depicts that it increases slowly during dawn and reaches a peak point between ~09.00 to ~12.00 UT.
After that, the VTEC start to decrease gradually and attains its minimum value at ~21.00 UT. The monthly results of VTEC variation indicate the lowest value in September while it shows maximum value in March months. The relative deviation monthly range of VTEC variation is noted from -1 to 4 units for all stations which confirm maximum difference 5 between negative and positive variability. The measured grand-mean VTEC intensity is in descending order during February, April, June, and September while in ascending order during March, May, August, and November months. The seasonal VTEC variation shows an increment of VTEC during March equinox followed by September equinox while December solstice has higher VTEC values comparison to June solstice. The overall pattern of VTEC variation is enhancing at all stations at the end of years contrast to mid-year due to the high solar activity.
The comparative analysis among the GNSS-VTEC, Kriging, NeQuick and the proposed mathematical model are evaluated with the help one way ANOVA test. The analysis shows that the null hypothesis of the models during the storm and quiet days is acceptable and suggesting that all models are statistically significantly equivalent to each other. The comparative analysis displays the prediction errors by OLSE, ARMA, and IRI-2016 varying from 0.23 to 1.17 %, 2.40 to 4.03 % and 24.82 to 25.79% respectively.
Additionally, the seasonal VTEC variation ties good agreement with OLSE as well as ARMA predicted results models but IRI prediction repeatedly underestimates the VTEC value at each location. Hence, predicted results by OLSE and ARMA models claim for further improvements in IRI model over the Turkish region. The analysis VTEC variation during the strong geomagnetic storm period (07-11 September 2015; SYM-H -120 nT) also studied by using SPIM and GIM models which predict the relative better result in comparison to the IRI-2012 model. Advanced analysis of GNSS data over Turkey may complement towards the future refinement of NeQuick and IRIs model over the lower mid-latitude region.
I believe the outcomes from this study would complement a better understanding of VTEC variation over the lower mid-latitude Turkish region and analogous latitudes over the globe.
References
- Ansari K and Corumluoglu O (2016) Ionospheric Observation over Turkey by using Turkish Permanent GPS Network, International Conference on Agricultural, Civil and Environmental Engineering (ACEE-16), Istanbul, Turkey: 32-36; doi:10.17758/URUAE. AE0416224
- Ansari K, Panda SK, Althuwaynee OF and Corumluoglu O (2017) Ionospheric TEC from the Turkish Permanent GNSS Network (TPGN) and comparison with ARMA and IRI models, Astrophysics and Space Science, 362:178, doi.org/10.1007/s10509-017-3159-z
- Ansari K, Corumluoglu O and Panda SK (2017) Analysis of Ionospheric TEC from GNSS observables over the Turkish region and predictability of IRI and SPIM models, Astrophysics and Space Science, 332:625,1-24,; doi.org/10.1007/s10509-017-3043-x
- Ansari K, Panda SK and Corumluoglu O (2018) Mathematical Modelling of Ionospheric TEC from Turkish permanent GNSS Network (TPGN) Observables during 2009-2017 and predictability of NeQuick and Kriging models, Astrophysics and Space Science (In Press)
These findings are described in the article entitled Ionospheric TEC from the Turkish Permanent GNSS Network (TPGN) and comparison with ARMA and IRI models, published in the journal Astrophysics and Space Science. This work was led by Kutubuddin Ansari from Izmir Katip Celebi University.