Elsevier

Advances in Space Research

Volume 66, Issue 8, 15 October 2020, Pages 1937-1946
Advances in Space Research

Performance evaluation of model-assisted data inversion technique for ionospheric radio occultation

https://doi.org/10.1016/j.asr.2020.06.042Get rights and content

Abstract

Radio occultation (RO) has proved to be a powerful tool for weather forecasting and ionospheric modelling. It is a truly global technique for the observation in the lower as well as in the upper atmosphere. With the evolution of radio occultation technique, the data inversion algorithms used to extract data from the raw RO measurements also evolved over the period of last two decades. In this work, a thorough investigation has been conducted on the performance of model-assisted inversion technique proposed by Shaikh et al. (2017). A combination of low, mid and high latitude geographical locations was chosen to analyze the performance of the technique. It has shown that the proposed technique has significantly improved the data inversion results in high solar activity conditions when the horizontal gradients of the ionosphere are expected to be stronger than normal (for example 24% at King Salmon and 21% at Jeju in high solar activity). The longitudinal sector-wise analysis shows that the American longitudinal sector has slightly weaker performance measures for model-assisted inversion compared to the Asian longitudinal sector.

Introduction

Since the introduction of GNSS (Global Navigation Satellite System) based radio occultation (RO) technique for monitoring the Earth’s atmosphere (Anthes et al., 2000), there is a significant improvement in the weather forecasting for troposphere and modelling for the upper atmosphere. It is a powerful tool offering truly global and continuous measurements regardless of the geographical location. The technique involves a low Earth orbit (LEO) satellite receiving radio signals from a GNSS satellite. The signal has to pass through the atmosphere and get refracted along the way. The magnitude of the refraction depends on the temperature and water vapor concentration in the troposphere and integrated electron content, known as total electron content (TEC), in the ionosphere.

The standard technique used to invert the TEC data into vertical electron density (Ne(h)) profile in the ionosphere is based on the inverse Abel transform (Schreiner et al., 1999); hereinafter called ‘standard inversion’. In past two decades, several authors have suggested modification and potential improvement of this technique (Hernàndez-Pajares et al., 2000, Tsai and Tsai, 2004, Kulikov et al., 2011) due to its spherical symmetry problem that leads to erroneous electron density values in the bottomside of the inverted Ne(h) profile. Most of these techniques either used the data from global ionospheric maps (GIM) or already existing Ne(h) profile data to help improve the inversion process. Tulasi Ram et al. (2016) recently introduced a technique which is based on an asymmetry factor calculated using map of NmF2 values derived from a global observation of Abel inverted RO electron density data. However, this technique, like other techniques, also needs significant amount of data from the already existing database of vertical electron density profiles to perform the inversion of a single RO event. Yue et al. (2013) examined GIM based techniques thoroughly in an assessment study and concluded as follows:

  • (1)

    GIM-based techniques may be effective only if accurate GIMs are available;

  • (2)

    Retrieved Ne(h) profile error given by different authors differs mainly due to the quality of the selected GIMs;

  • (3)

    The problem of having data gaps over the oceans cannot be resolved using GIM based techniques.

Shaikh et al. (2017) has recently proposed a data inversion technique (hereinafter called ‘model-assisted inversion’) using NeQuick2 (Nava et al. 2008) as background ionospheric model to compensate for the horizontal gradients along the RO ray paths. The technique is based on the minimization of cost function involving experimental (RO data) and modelled (NeQuick2-derived) TEC data to calculate the effective local ionization parameter (F10.7 solar radio noise flux) which is then used to retrieve the Ne(h) profile along the tangent point (TP) locations using NeQuick2. Since no external or processed data is needed, this technique can be effectively used to improve the retrieval of Ne(h) profile by compensating for the spherical symmetry problem of standard inversion in near-real time. In this paper, the performance of the model-assisted inversion has been assessed processing RO data with TP located in different geographical regions. In the ‘Data’ section, a brief detail of the data used in this study has been presented. Results and analysis are presented in Section 3, followed by discussion in Section 4 and conclusion in Section 5.

Section snippets

Data

The radio occultation (RO) data used in this study has been acquired from the CDAAC (https://cdaac-www.cosmic.ucar.edu/cdaac/products.html) database. Only RO data from Constellation Observing System for Meteorology, Ionosphere and Climate-1 (COSMIC-1) for the period from 01/01/2010 to 31/12/2016 has been used. The reason for using only COSMIC-1 data is due to its complete coverage of the ionosphere’s bottomside as well as topside. In this work, only calibrated STEC values available from

Results and analysis

This study is a performance assessment of model-assisted technique proposed by Shaikh et al. (2017). It was proposed that an empirical model might help improve the radio occultation data inversion to account for horizontal electron density gradients in the ionosphere. The technique works by calculating a cost functions (Az) by minimizing the difference between RO observed slant-TEC (STEC) and the NeQuick2 modelled STEC (using the same geometry of RO event). A set of Az parameters is then

Discussion

Table 2 shows a summary of relative error mean for all stations listed in Table 1. It is clear from the individual station’s result that, in low solar activity, for Jicamarca, the mean relative error in calculating NmF2 using model-assisted inversion is similar to that of the standard inversion. All other geographical locations in low-mid latitude showed improvements with the best improvement achieved at Jeju (10%). For high latitudes, King Salmon, for example, suffered significant errors and

Conclusion

In this work, by implementing model-assisted inversion technique, we have shown that the ionospheric RO data inversion can be improved significantly by using an empirical model. Here, we have only discussed the improvement in the retrieval of NmF2 using model-assisted inversion technique proposed by Shaikh et al. (2017). The technique is based on the minimization of a cost function (Az) between experimental STEC (from RO data) and the modelled STEC to aid the RO data inversion for the retrieval

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgment

Author is thankful for the reviewer for the valuable comments that helped improve the quality of the paper.

Author is thankful for the following scientific data providing agencies for providing the data used in this study:

  • University Cooperation for Atmospheric Research (UCAR) for providing COSMIC mission data through CDAAC database: http://cdaacwww .cosmic.ucar.edu/cdaac/products.html;

  • Lowell GIRO Data Center (LGDC) for providing ionsonde data other than the Japanese stations through their

References (19)

There are more references available in the full text version of this article.

Cited by (6)

  • NeQuick2 validation using multi-satellite and ground data

    2023, Advances in Space Research
    Citation Excerpt :

    The comparative analysis shown is very encouraging although they were only able to improve a fraction of the TECU in low as well high solar activity conditions. After the analysis done in this paper, it is believed that suitable method of ingestion of all those different sources of data which were used in this study would further help improve the topside performance of the NeQuick2 model (Shaikh et al., 2017; Shaikh, 2020; Shaikh, 2021). Therefore, an attempt to ingest all the available experimental data to improve the overall performance of the NeQuick2 Ne(h) profile is foreseen as future work.

  • On the use of empirical models for high-latitude ionosphere

    2022, Journal of Atmospheric and Solar-Terrestrial Physics
    Citation Excerpt :

    Such small differences in rms show that NeQuick2, although mainly design to work with CCIR coefficients for which data is mostly obtained from the low latitudes, still performs very well for high-latitude ionosphere. Furthermore, previous studies show that NeQuick2 model is flexible enough to be used in data assimilation mode which may improve the performance of the model significantly at any latitudinal regions (for example, see Shaikh, 2021; Shaikh, 2020; Shaikh et al., 2017; Shaikh et al., 2018). This short paper comparatively analyze peak F2-layer peak parameters (foF2 and hmF2) calculated using two empirical models NeQuick2 and E-CHAIM.

  • DIURNAL AND SEASONAL VARIATIONS OF GNSS BASED IONOSPHERIC SLAB THICKNESS OVER ARABIAN PENINSULA

    2022, Proceedings of the International Astronautical Congress, IAC
  • Retrieval of Ne(H) profile for ionosonde by assimilating collocated radio occultation data into NeqUICK2

    2021, Proceedings of the 34th International Technical Meeting of the Satellite Division of the Institute of Navigation, ION GNSS+ 2021
View full text