-
[show abstract]
[hide abstract]
ABSTRACT: 1] In this paper, the potential extrapolation capabilities and limitations of artificial neural networks (ANNs) are investigated. This is primarily done by generating total electron content (TEC) predictions using the regional southern Africa total electron content prediction (SATECP) model based on the Global Positioning System (GPS) data and ANNs with the aid of multiple inputs intended to enable the software to learn and correlate the relationship between their variations and the target parameter, TEC. TEC values are predicted over regions that were not covered in the model's development, although it is difficult to validate their accuracy in some cases. The SATECP model is also used to forecast hourly TEC variability 1 year ahead in order to assess the forecasting capability of ANNs in generalizing TEC patterns. The developed SATECP model has also been independently validated by ionosonde data and TEC values derived from the adapted University of New Brunswick Ionospheric Mapping Technique (UNB‐IMT) over southern Africa. From the comparison of prediction results with actual GPS data, it is observed that ANNs extrapolate relatively well during quiet periods while the accuracy is low during geomagnetically disturbed conditions. However, ANNs correctly identify both positive and negative storm effects observed in GPS TEC data analyzed within the input space.
J. Geophys. Res. 01/2011; 116.
-
IEEE International Geoscience & Remote Sensing Symposium, IGARSS 2009, July 12-17, 2009, University of Cape Town, Cape Town, South Africa, Proceedings; 01/2009
-
[show abstract]
[hide abstract]
ABSTRACT: The propagation of radio signals in the Earth’s atmosphere is dominantly affected by the ionosphere due to its dispersive nature. Global Positioning System (GPS) data provides relevant information that leads to the derivation of total electron content (TEC) which can be considered as the ionosphere’s measure of ionisation. This paper presents part of a feasibility study for the development of a Neural Network (NN) based model for the prediction of South African GPS derived TEC. The South African GPS receiver network is operated and maintained by the Chief Directorate Surveys and Mapping (CDSM) in Cape Town, South Africa. Vertical total electron content (VTEC) was calculated for four GPS receiver stations using the Adjusted Spherical Harmonic (ASHA) model. Factors that influence TEC were then identified and used to derive input parameters for the NN. The well established factors used are seasonal variation, diurnal variation, solar activity and magnetic activity. Comparison of diurnal predicted TEC values from both the NN model and the International Reference Ionosphere (IRI-2001) with GPS TEC revealed that the IRI provides more accurate predictions than the NN model during the spring equinoxes. However, on average the NN model predicts GPS TEC more accurately than the IRI model over the GPS locations considered within South Africa.
Advances in Space Research.
-
[show abstract]
[hide abstract]
ABSTRACT: This paper presents the results from a study designed to investigate the ability of a newly developed neural network (NN) based model to follow total electron content (TEC) dynamics over the Southern African region. The investigation is carried out by comparing results from the NN model with actual TEC data derived from Global Positioning System (GPS) observations and TEC values predicted by the International Reference Ionosphere (IRI-2007) model during magnetic storm periods over Southern Africa. The magnetic storm conditions chosen for the study presented in this paper occurred during the periods 16–21 April 2002, 1–6 October 2002, and 28 October–01 November 2003. A total of six South African GPS stations were used for the validation of the two models during these periods. A statistical analysis of the comparison between the actual TEC behaviour and that predicted by the two models is shown. In addition, ionosonde measurements from the South African Louisvale (28.5°S, 21.2°E) station, located close to one of the validation GPS stations used, are also considered during the Halloween storm period of 28–31 October 2003. The generalisation of TEC behaviour by the NN model is demonstrated by producing predicted TEC maps during magnetic storm periods over South Africa. Presented results demonstrate the ability of NNs in predicting TEC variability over South Africa during magnetically disturbed conditions, and highlight areas for improvement.
Journal of Atmospheric and Solar-Terrestrial Physics 72:509-520. · 1.60 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: Global positioning system (GPS) networks have provided an opportunity to study the dynamics and continuous changes in the ionosphere by supplementing ionospheric studies carried out using various techniques including ionosondes, incoherent scatter radars and satellites. Total electron content (TEC) is one of the physical quantities that can be derived from GPS data, and provides an indication of ionospheric variability. This paper presents a feasibility study for the development of a Neural Network (NN) based model for the prediction of South African GPS derived TEC. Three South African locations were identified and used in the development of an input space and NN architecture for the model. The input space included the day number (seasonal variation), hour (diurnal variation), Sunspot Number (measure of the solar activity), and magnetic index (measure of the magnetic activity). An analysis was done by comparing predicted NN TEC with TEC values from the IRI-2001 version of the International Reference Ionosphere (IRI), validating GPS TEC with ionosonde TEC (ITEC) and assessing the performance of the NN model during equinoxes and solstices. For this feasibility model, GPS TEC was derived for a limited number of years using an algorithm still in the early phases of validation. However, results show that NNs predict GPS TEC more accurately than the IRI at South African GPS locations, but that more good quality GPS data is required before a truly representative empirical GPS TEC model can be released.
Journal of Atmospheric and Solar-Terrestrial Physics.
