Ionospheric data assimilation: Recovery of strong mid-latitudinal density gradients

The September 1999 Caribbean Ionospheric Campaign (CIC99), spanning the period 15-17 September 1999, was a period of repeated moderate geomagnetic storms. Both incoherent scatter radar (ISR) and digisonde F-layer measurements were made in the vicinity of Arecibo, Puerto Rico by the Arecibo ISR and Ramey Digisonde, which is one of the Digital Ionospheric Sounding System instruments. These data sets showed significant day-to-day F-layer variability, especially in the evening sectors. Local ionospheric data assimilation with the Assimilation Ionospheric Model (AIM1.06L) was able to reproduce this ionospheric day-to-day weather. However, during this time the Arecibo ISR was able to determine that strong latitude density gradients existed in the evening sector. These latitude gradients ranged from 5% to over 10% per degree of latitude, increasing equatorward. The climatological ionospheric forecast model (IFM) predicts gradients of the correct sign but significantly less than 5% per degree. Using the observed fact that the strong local vertical plasma drift was produced by an eastward electric field, an enhanced equatorial electric field model was created. This model was used to drive the IFM-E model. The effect of the enhanced electric field is to drive plasma flux tubes to higher latitudes and effectively move the poleward shoulder of the equatorial anomalies to higher latitudes. In this simulation, the simulated Arecibo density gradients range from 10% to 20% per degree. In addition, the local densities at Arecibo are now higher than the observed ones in the evening sector. The irony of this study is that the local AIM1.06L assimilation provides the correct local densities in the F region but is based upon the assumption of corotating field lines. When the plasma flux tubes are free to be E × B/B² convected, as in the IFM equator model simulations, the local densities are too high. However, gradients in latitude are more realistic although somewhat larger than observed. This latter simulation is based on "better" physics than the AIM1.06L, but because the latitude (apex altitude) distribution of the eastward electric field driver is uncertain, it produces poor assimilation results. This study points to the need for more extended latitude coverage in the CIC campaigns in order to address the issue of a minimum data requirement for the assimilation.