Correlation studies for B-spline modeled F2 Chapman parameters obtained from FORMOSAT-3/COSMIC data

摘要

The determination of ionospheric key quantities such as the maximum electrondensity of the F2 layer NmF2, the corresponding F2 peak heighthmF2 and the F2 scale height HF2 are of highrelevance in 4-D ionosphere modeling to provide information on the verticalstructure of the electron density (N_e). The N_edistribution with respect to height can, for instance, be modeled by thecommonly accepted F2 Chapman layer. An adequate and observation drivendescription of the vertical N_e variation can be obtained fromelectron density profiles (EDPs) derived by ionospheric radio occultationmeasurements between GPS and low Earth orbiter (LEO) satellites. For thesepurposes, the six FORMOSAT-3/COSMIC (F3/C) satellites provide an excellentopportunity to collect EDPs that cover most of the ionospheric region, inparticular the F2 layer. For the contents of this paper, F3/C EDPs have beenexploited to determine NmF2, hmF2 andHF2 within a regional modeling approach. As mathematical basefunctions, endpoint-interpolating polynomial B-splines are considered tomodel the key parameters with respect to longitude, latitude and time. Thedescription of deterministic processes and the verification of this modelingapproach have been published previously in Limberger et al. (2013), whereas this papershould be considered as an extension dealing with related correlationstudies, a topic to which less attention has been paid in the literature. Relationsbetween the B-spline series coefficients regarding specific key parameters aswell as dependencies between the three F2 Chapman key parameters are in themain focus. Dependencies are interpreted from the post-derived correlationmatrices as a result of (1) a simulated scenario without data gaps by takingdense, homogenously distributed profiles into account and (2) two real datascenarios on 1 July 2008 and 1 July 2012 including sparsely, inhomogeneouslydistributed F3/C EDPs. Moderate correlations between hmF2and HF2 as well as inverse correlations between NmF2 andHF2 are reflected from the simulation. By means of the real datastudies, it becomes obvious that the sparse measurement distribution leads toan increased weighting of the prior information and suppresses the parametercorrelations which play an important role regarding the parameterestimability. The currently implemented stochastic model is in need ofimprovement and does not consider stochastic correlations which consequentlycannot occur.