[1] Since the FORMOSAT‐3/COSMIC satellites were launched in April 2006, ionospheric electron density profiles have been retrieved from the excess phase of the GPS signal by using the radio occultation technique and can be accessed from the Web site http://www.cosmic.ucar.edu/. On the basis of these electron density profiles and the use of data quality control criteria developed by Yang et al. [2009], Chu et al. [2009] investigated E region electron density morphology and showed that the general properties of the COSMIC‐ retrieved E region electron density are in good agreement with the predictions of the Chapman layer theory that was developed in accordance with photochemical process and controlled by solar zenith angle. Nevertheless, Chu et al. [2009] found existences of salient enhancements in the noontime E region electron density not only at the geomagnetic equator but also in the geomagnetic latitude regions ±15°−35°, which cannot be explained by the Chapman layer theory. In addition, they also provided compelling evidence to show the presence of longitudinal wave number 3 and 4 structures of the equatorial electron density in a height range of 100–200 km, which is in excellent agreement with longitudinal structures of equatorial electrojet intensity derived from equatorial magnetic field data obtained by the Orsted, CHAMP, and SAC‐C satellites during the years 1999–2006. [2] On the basis of the simulation result obtained by Yue et al. [2010], Lei et al. [2010] question the validity of the E region electron density retrieved by the GPS radio occultation technique. They argue that because of the presence of the ionospheric electron density gradient in the horizontal direction that violates the spherical symmetry assumption of the Abel transform for inverting ionospheric electron density profile from calibrated total electron content along the GPS raypath, the COSMIC‐measured E region electron density enhancements in midlatitude regions were caused by the retrieval error of the GPS radio occultation process. From Figure 1 of Lei et al. [2010] (identical to Figure 2 of Yue et al. [2010]), the simulation‐retrieved E region electron densities around 100 km in equinox season are approximately 50– 100% and 150–200% greater than the “true” model values at the geomagnetic equator and in geomagnetic latitude regions ±30°−50°, respectively. In addition, the former are about 150–200% smaller than the latter in geomagnetic latitude regions ±10°−30°. If the simulation‐retrieved results obtained by Yue et al. [2010] were true and able to be representative of the general GPS occultation‐retrieved results, the morphologies of the COSMIC‐measured E region electron density should be in accord with those of the simulation results. Namely, the E region electron density retrieved by COSMIC satellites should be much greater (smaller) than true measurement made by the ground‐based ionosonde in geomagnetic latitude regions ±30°–50° (±10°–30°). In order to validate the simulation‐retrieved results, we compare peak values of E layer electron density NmE between COSMIC retrieval and global ionosonde measurement in the different latitudinal regions for July 2006 to July 2009. The COSMIC data were selected for comparison if the separation between COSMIC occultation point and ionosonde station is 10 min in time and 2.5° in space. Figure 1 presents an example of latitudinal variation in histograms of percent errors of NmE between COSMIC retrieval and ionosonde measurement for spring (March–May) 2007–2009, in which the percent error (PE) is defined by
Read full abstract