Comparison between GPS radio occultation electron densities and in situ satellite observations

Abstract

Constellation Observing System for Meteorology, Ionosphere, and Climate (COSMIC) Global Positioning System (GPS) radio occultation (RO) electron densities are compared with collocated in situ observations from the CHAllenging Minisatellite Payload (CHAMP) and Communications/Navigation Outage Forecasting System (C/NOFS) satellites. The comparison is restricted to observations occurring within 2 ∘ latitude and longitude and 15 min local time. The in situ observations occur at altitudes of ∼300-800 km, and the results of the present study represent the first global comparison of COSMIC electron densities at altitudes ranging from near the F region peak to the topside ionosphere. The correlation coefficient between the COSMIC and in situ observations is greater than 0.90, indicating an overall good agreement between GPS RO electron densities and CHAMP and C/NOFS satellite observations. Furthermore, when averaged over all latitudes and local times, we find a near-zero mean bias and root-mean-square difference of typically less than ±30% between the COSMIC electron densities and collocated in situ observations. The overall good agreement demonstrates that the COSMIC GPS RO observations provide an accurate measure of electron density in the topside ionosphere. The results also reveal a systematic structure to the error in the equatorial and low-latitude daytime ionosphere. This structure is related to the equatorial ionization anomaly and is consistent with the error introduced by the Abel inversion spherical symmetry assumption used to retrieve the COSMIC electron density profiles. The present study thus provides direct observational evidence of the Abel inversion error on GPS RO electron densities.

Long‐term analysis between radio occultation and ionosonde peak electron density and height during geomagnetic storms

“…Thus, some comparative studies between RO and ground‐based observations preferred analyses during geomagnetically quiet conditions [e.g., Chu et al , ; Wu et al , ] to avoid errors related to the occurrence of ionospheric electron density irregularities. However, other studies have done validation case studies on one storm period [e.g., Krankowski et al , ], while significant literature that exists performs statistics based on large data sets without necessarily doing any isolation based on geomagnetic activity or occurrence of ionospheric irregularities [e.g., Yue et al , ; Chuo et al , ; Pedatella et al , , and references therein]. These studies give an indication that RO data may be able to follow ionospheric dynamics even during geomagnetic storms, although not in a detailed quantitative sense.…”

Adapting a climatology model to improve estimation of ionosphere parameters and subsequent validation with radio occultation and ionosonde data

“…Considering COSMIC data, the modified IRI‐2012 model leads to an improvement in estimating topside N e profiles. A recent global study comparing GPS radio occultation N e with in situ CHAMP and C/NOFS satellite observations for altitude range from near F 2 peak to topside ionosphere reported that the COSMIC is accurate at providing topside N e data [ Pedatella et al , ]. These authors also found errors within the low latitude and equatorial ionization anomaly regions resulting from the spherical symmetry assumption employed by the Abel inversion technique in deriving electron density profiles.…”

Improvement of the Ionospheric Radio Occultation Retrievals by Means of Accurate Global Ionospheric Maps

“…The f o F 2 values derived from EDPs will be compared with ionosonde measurements that will be taken as reference values to check the quality of the RO inversion. In this way, one can assess the error of the RO results by calculating the distribution of the relative difference with respect to ionosonde measurements: Since early studies, this relative difference has been used as a metric for assessing the accuracy of f o F 2 retrievals (e.g., Hajj & Romans, ; Hernández‐Pajares et al, ; Schreiner et al, ) and more recently in Pedatella et al () or Habarulema and Carelse ().…”

“…Since early studies, this relative difference has been used as a metric for assessing the accuracy of f o F 2 retrievals (e.g., Hajj & Romans, 1998;Hernández-Pajares et al, 2000;Schreiner et al, 1999) and more recently in Pedatella et al (2015b) or Habarulema and Carelse (2016).…”