Elsevier

Advances in Space Research

Volume 57, Issue 12, 15 June 2016, Pages 2452-2463
Advances in Space Research

A long-term study on the deletion criterion of questionable electron density profiles caused by ionospheric irregularities – COSMIC radio occultation technique

https://doi.org/10.1016/j.asr.2016.03.034Get rights and content

Highlights

  • A reliable deletion criterion of questionable density profiles retrieved using the COSMIC RO technique is discussed.

  • Statistics of seven continuous years are presented.

  • Computed gradient and the slope of density profiles have shown solar activity dependency.

Abstract

The crucial assumption made in the retrieval of radio-occultated atmospheric parameters is the spherical symmetry of the atmospheric refractive index, which implies that no horizontal gradient of the refractive index exists along the spherical shell. Nevertheless, the presence of density irregularities will lead to scintillation and multipath effects that often create highly fluctuating and random electron density profiles. In this study, it is proposed a reliable data quality control (QC) approach to remove questionable electron density profiles (due to the presence of ionospheric irregularities) retrieved using the COSMIC radio occultation (RO) technique based on two parameters, namely, the gradient and fluctuation of the topside density profile. Statistics of seven years density profiles (July 2006–May 2013) are presented by determining the aforementioned parameters for every density profile. The main advantage of this data QC is that it uses COSMIC RO electron density profiles retrieved from the slant total electron content (TEC) that is estimated from the excess phases of the GPS L1 and L2 frequencies only to delete the questionable profiles, instead of relying on any model and other observations. A systematic criterion has been developed based on the statistics to relinquish the so-called questionable density profiles. The computed gradients and fluctuations of the topside ionosphere electron density profiles have shown a few important features including, solar activity dependency and pronounced variations in between around +40° and −40° latitudes. After the removal of questionable profiles, both peak densities and heights of the ionosphere F layer are presented globally in different seasons of years during 2007 and 2012 that revealed several important features.

Introduction

The efforts made by the University Corporation for Atmospheric Research (UCAR), Boulder, USA in 1993 to organize a proof-of-concept Global Positioning System (GPS) Meteorology, in short GPS/MET (Ware et al., 1996), experiment on a 735 km low Earth orbit (LEO) MicroLab-1 satellite to receive GPS signals to study the Earth’s atmosphere and ionosphere by radio occultation (RO) technique had, later, led to initiate several GPS RO missions including, CHAllenging Mission Payload (CHAMP) (Wickert et al., 2001) by GFZ German Research Centre for Geosciences, Satellite de Aplicaciones Cientificas-C (SAC-C) (Hajj et al., 2004) by Argentina, Gravity Recovery and Climate Experiment (GRACE) by Germany-USA (Beyerle et al., 2005) and a six-satellite constellation of Constellation Observing System for Meteorology, Ionosphere, and Climate/third major project of the Formosa satellite series (COSMIC/FORMOSAT-3) (hereafter only COSMIC) micro satellites jointly by USA and Taiwan (Cheng et al., 2006). In GPS RO technique, in order to retrieve terrestrial atmosphere parameters the bending angle of radio wave transmitted by a very stable source situated at Earth’s orbit such as GPS satellite (acts as a transmitter) and received by a LEO satellite (acts as a receiver) on Earth’s orbit will be transformed into a corresponding atmospheric refractive index under a number of assumptions. Once the atmospheric refractive index is retrieved via the famous Abel (after Niels Henrick Abel) integral transform, it is possible to estimate the lower atmospheric temperature, humidity, and ionospheric electron density at the tangent point of the ray path piercing through the atmosphere (Schreiner et al., 1999).

Except for the atmospheric refractive index, under the straight line assumption of the radio ray path, the height variation of the ionospheric electron density can also be retrieved from calibrated total electron content (TEC) in accordance with the Abel transformation, which can be estimated from the phase path difference between L1 (=1.57542 GHz) and L2 (=1.22760 GHz) frequencies of GPS signals (Hajj and Romans, 1998, Schreiner et al., 1999). It can generally be understood from the fact that the ray bending in the ionosphere is too meager and the error introduced at tangent point is less than a few kilometers (Schreiner et al., 1999). Lei et al. (2007) have thoroughly described the detailed Abel retrieval procedure used by the COSMIC Data Analysis and Archival Center (CDAAC) at the UCAR for the inversion of COSMIC RO observations.

Near real-time processed RO data have been used by the data assimilation models in most weather centers all over the world (Anthes, 2011). Decrease in either the data volume or data quality occurrence will have negative effects in both weather prediction and space weather monitoring and this effect is more serious if both space weather and extreme weather events including Typhoon and solar radio burst (SRB) occur simultaneously (Yue et al., 2013). In climate study, the above effects should be considered and mitigated too. Due to the solar activity dependency of space weather occurrence, the negative effects of them on RO should be more significant during solar maximum, and, hence more attentions need to be paid by individual research as well as ensuing radio occultation mission sponsored groups in various countries including, USA (CDAAC) and Germany (GFZ German Research Centre for Geosciences) and others in order to provide more efficient and reliable data to the user community by developing efficient QC schemes.

The crucial assumptions in GPS RO techniques are the spherical symmetry of atmosphere refractive at the locality of occultations (Hajj et al., 2000, Kursinski et al., 2000) and geometrical optics assumption (Kursinski et al., 2000), although the assumption of spherical symmetry has been almost never true (Schreiner et al., 1999, Yue et al., 2010, Yue et al., 2011b). More specifically, the spherical symmetry assumption implies that no horizontal gradients of refractive index exist along the spherical shell. If there are no irregular electron density distributions in the GPS ray path, the occultation-retrieved electron density will be a smooth curve without random fluctuations superimposed on the curve. In reality, nonetheless, the presence of ionospheric irregularities is obvious, particularly in equatorial, ionization anomaly crest (during local sunset times) and at high-latitude regions (even during daytime hours), and during space weather events including geomagnetically disturbed and solar radio burst (SRB) (Yue et al., 2013) epochs. As a result, the retrieved electron density profile will be highly fluctuating which, eventually, gives rise to a large uncertainty in the estimation and impairment of the data reliability (Yang et al., 2009, Potula et al., 2011).

It is important to mention here that in the COSMIC RO ionospheric processing there are several other important error sources, due to: (i) spherical symmetry assumption (ii) cycle slips, and (iii) ionospheric irregularities. Due to the spherical symmetry assumption, the retrieved electron density profile causes some systematic errors. For instance, it has been reported that the spherical symmetry assumption often creates plasma caves, which are the regions with reduced electron density underneath the equatorial ionization anomaly (EIA) crests (Yue et al., 2010), and a pseudo reversal phase of the large-scale wave structure in the E and F1 layer (Yue et al., 2010, Yue et al., 2011a, Yue et al., 2011b, Yue et al., 2012). Secondly, cycle slips are loss of signal lock in the GPS receiver tracking loop and these are observed as a discontinuity in the phase integer ambiguity. Cycle slips can occur under ionospheric scintillation conditions when high phase dynamics (exceeding receiver tracking loop bandwidths) arise from ionospheric irregularities. According to the information provided by CDAAC a small part of the COSMIC electron density profiles are affected by cycle slips in the GPS phase data (Zakharenkova et al., 2012), meaning that the cycle slip of the raw phase is not identified in the processing. This will, therefore, certainly cause some unreliable profiles (sudden jump of the profiles). Finally, ionospheric irregularities will cause some small scale fluctuations of the profile, as will be shown in this paper.

A few individual research groups have developed their own ‘questionable electron profile data deletion criteria’ schemes during the early launch period of the COSMIC RO technique. For instance, in order to obtain a correct electron density profile, Lei et al. (2007) have used scale height at F-layer peak heights and correlation coefficient between fitted and observed profiles to serve as the rejection criteria to screen-out the COSMIC electron density data and used only the best quality data to compare with the international reference ionosphere (IRI) and thermosphere ionosphere electrodynamics general circulation model (TIEGCM) predictions. Liu et al. (2008) fitted individual electron density profiles between 160 and 600 km with a Chapman-α function to determine the peak height (hmF2), its density (NmF2), and scale height (Hm) in a least squares procedure and had discarded significantly deviated individual density profiles by comparing them with the determined ionospheric parameters. Yang et al. (2009) have also calculated mean deviations and slope of the topside electron density profile to screen-out questionable density profiles, however, for a mere eight month period (June 2006–January 2007). Recently, Liu et al. (2010) have made direct comparisons between COSMIC electron density profiles and the ones measured with incoherent scatter radars (ISR) at Jicamarca (11.9°S, 76°W) and Millstone Hills (42.6°N, 71.5°W) and in order to determine hmF2, the COSMIC retrieved electron density profiles were fitted with a two-layer Chapman function. Liu et al. (2010) have screened-out the profiles based on the fitting errors that the profiles having errors of less (greater) than 10% have been used (deleted) for further analysis. With the incorporation of large database in the present research, we could find several interesting features, particularly the solar-activity dependent aspect of mean deviations and the slope of topside density profiles. The organization of this article is as follows: In Section 2, few typical examples of unreliable electron density profiles are presented that show highly irregular shapes at lower and upper parts of the ionosphere and positive slopes in the upper ionosphere. It is discussed about the adopted methodology to relinquish unreliable and questionable density data in Section 2.1. The global trends of mean deviations and topside slopes are provided in Section 3.1, while Section 3.2 discusses global features of peak densities and peak heights of the ionosphere along with discussion part. Section 3.3 presents the typical comparisons between unprocessed and processed density profiles and peak heights during the spring equinox season of 2009. Section 4 contains conclusions.

Section snippets

Methodology

Before discussing our QC criteria in a detailed manner, we have presented here a few profiles that come under unreliable and questionable category that need to be deleted before doing any further analysis. Two panels in Fig. 1 show electron density profiles measured by the same satellite (FM2), which received the GPS signals transmitted from GPS satellite of numbers 19 (left panel) and 08 (right panel) respectively, but geographically left (right) electron density profile belongs to southern

Mean deviations and topside ionosphere slopes

In order to verify the global behaviors of mean deviations and slopes associated with processed profiles, seasonal trends of them during 2007 and 2012 are presented in the following way. Fig. 4 depicts the global (geographic longitude vs. geographic latitude) distributions of magnitudes of MDs in different seasons, including the March equinox (March, April and May), June solstice (June, July and August), September equinox (September, October and November), and December solstice (December,

Conclusions

We set up the thresholds for the quality control of radio occultation electron density profiles to screen-out questionable data, in which the mean deviation and the slope of the topside electron density profile are anomalous.

With Abel transformation, the COSMIC-measured electron density profile over the GPS ray tangent point is retrieved from the slant TEC that is estimated from the excess phases of the GPS L1 and L2 frequencies. If there are electron density irregularities with scales much

Acknowledgements

COSMIC data were obtained from the Taiwan Analysis Center for COSMIC (TACC)/COSMIC Data Analysis and Archive Centre (CDAAC). Dr. G. Uma is very much grateful to Department of Science & Technology (DST), Government of India for providing a Women Scientist Scheme-A (SR/WOS-A/EA-1015/2015).

References (39)

  • D.N. Anderson

    A theoretical study of the ionospheric, F region equatorial anomaly: I. Theory

    Planet. Space Sci.

    (1973)
  • R.A. Duncan

    The equatorial F region of the ionosphere

    J. Atmos. Terr. Phys.

    (1960)
  • J. Aarons et al.

    VHF scintillation activity over polar latitudes

    Geophys. Res Lett.

    (1981)
  • J. Aarons et al.

    GPS phase fluctuations in the equatorial region during solar minimum

    Radio Sci.

    (1997)
  • R.A. Anthes

    Exploring earth’s atmosphere with radio occultation: contributions to weather, climate and space weather

    Atmos. Meas. Tech.

    (2011)
  • E.V. Appleton

    Two anomalies in the ionosphere

    Nature

    (1946)
  • S. Basu et al.

    Characteristics of plasma structuring in the cusp/cleft region at Svalbard

    Radio Sci.

    (1998)
  • G. Beyerle et al.

    GPS radio occultation with GRACE: atmospheric profiling utilizing the zero difference technique

    Geophys. Res. Lett.

    (2005)
  • P.S. Brahmanandam et al.

    Global S4 index variations observed using FORMOSAT-3/COSMIC GPS RO technique during a solar minimum year

    J. Geophys. Res.

    (2012)
  • C.-Z.F. Cheng et al.

    Satellite constellation monitors global and space weather

    EOS Trans. AGU

    (2006)
  • C. Coker et al.

    High-latitude plasma structure and scintillation

    Radio Sci.

    (2004)
  • B.G. Fejer et al.

    Average vertical and zonal F region plasma drifts over Jicamarca

    J. Geophys. Res.

    (1991)
  • B.G. Fejer et al.

    Quiet time equatorial F region vertical plasma drift model derived from ROCSAT-1 observations

    J. Geophys. Res.

    (2008)
  • G.A. Hajj et al.

    Ionospheric electron density profiles obtained with the Global Positioning System: results from the GPS/MET experiment

    Radio Sci.

    (1998)
  • G.A. Hajj et al.

    COSMIC GPS ionospheric sensing and space weather

    Terr. Atmos. Oceanic Sci.

    (2000)
  • G.A. Hajj et al.

    CHAMP and SAC-C atmospheric occultation results and inter-comparisons

    J. Geophys. Res.

    (2004)
  • C.P. Ko et al.

    COSMIC/FORMOSAT-3 observations of equatorial F region irregularities in the SAA longitude sector

    J. Geophys. Res.

    (2010)
  • E.R. Kursinski et al.

    The GPS radio occultation technique

    Terr. Atmos. Oceanic Sci.

    (2000)
  • J. Lei

    Comparison of COSMIC ionospheric measurements with ground-based observations and model predictions: preliminary results

    J. Geophys. Res.

    (2007)
  • Cited by (21)

    • An aided Abel inversion technique assisted by artificial neural network-based background ionospheric model for near real-time correction of FORMOSAT-7/COSMIC-2 data

      2021, Advances in Space Research
      Citation Excerpt :

      The GIRO NmF2 and hmF2 data were obtained from the autoscaled ionogram echos using ARTIST-5 software (Galkin et al., 2008). To eliminate the noisy RO profiles, thorough quality control is applied on the data as proposed in Yang et al. (2009), Potula et al. (2011), and Uma et al. (2016). The GPS-RO profiles with an average standard deviation greater than 1.5 are omitted from the study.

    • Ionospheric responses to the 21 August 2017 great American solar eclipse – A multi-instrument study

      2020, Advances in Space Research
      Citation Excerpt :

      A constellation of six low earth orbiting (LEO) micro-satellites, named as COSMIC (Constellation Observing System for Meteorology, Ionosphere, and Climate), were deployed at 512 km on 15 April 2006; they later dispersed to an 800 km altitude (Cheng et al., 2006) to receive radio signals from Global Positioning System (GPS) using the RO technique (Kursinski et al., 1997). The COSMIC RO, a space-based method, has been proved to be one of the effective and accurate methods in providing lower (Brahmanandam et al., 2010; Anthes, 2011) and ionospheric (Potula et al., 2011; Brahmanandam et al., 2011; Brahmanandam et al., 2012; Uma et al., 2016; Uma et al., 2019) data products including temperature, water vapor, relative humidity, and ionospheric electron density with unprecedented resolutions. In this study, we used post-processed Ne(h) (data product name – “ionPrf”) that were collected from the CDAAC (COSMIC Data Analysis and Archive Center) website, www.cdaac.cosmic.ucar.edu.

    • The influence of the spatial and temporal collocation windows on the comparisons of the ionospheric characteristic parameters derived from COSMIC radio occultation and digisondes

      2019, Advances in Space Research
      Citation Excerpt :

      While with the technology of GPS RO, the ionosphere is detected using a GPS receiver on a low Earth orbit (LEO) satellite to observe the GPS satellite (Schreiner et al., 2007; Yue et al., 2011; Arras et al., 2008; Arras and Wickert, 2018). An electron density profile (EDP) is retrieved from the measured GPS L1 and L2 phase path differences when an ionospheric RO event occurs using the Abel inversion under the assumptions and approximations including straight-line signal propagation, spherical symmetry of electron density, circular satellite orbits and first-order estimation of electron density at the top (Lei et al., 2007; Wu et al., 2009; Yue et al., 2010; Brahmanandam et al., 2012; Xu et al., 2010; Uma et al., 2016). It is reasonable that differences exist in the comparison results of the above studies considering that besides the differences in the collocation criteria, differences in other factors, such as the time periods of the data, the locations of the digisonde stations, and the quality control and calibration methods of GPS RO and digisonde data, may also lead to the discrepancies in the results.

    • A complete solar cycle (2006–2016) studies of scale heights derived using COSMIC radio occultation retrieved electron density profiles

      2019, Journal of Atmospheric and Solar-Terrestrial Physics
      Citation Excerpt :

      The F10.7, as can be seen in the bottom panel of Fig. 1, is characterized by extremely low values, specifically between 2006 and 2008, during that period various researchers have conducted a few important global ionosphere studies using COSMIC RO retrieved ionosphere data products (Liu et al., 2008; Potula et al., 2011; Brahmanandam et al., 2012; Uma et al., 2012). Prior to the data processing, COSMIC RO retrieved density profiles have gone through a data quality control (QC) approach (Potula et al., 2011) to restrict a good number of unreliable and questionable density profiles (Uma et al., 2016). This QC approach has been developed by considering mean deviation (MD) and the vertical gradient (slope) of the topside electron density profiles.

    View all citing articles on Scopus
    View full text