Equatorial plasma bubbles (EPBs) are elongated plasma depletions that can occur in the nighttime ionospheric F region, causing scintillation in satellite navigation and communications signals. EPBs are believed to be Rayleigh-Taylor instabilities seeded by vertically propagating gravity waves. A necessary pre-condition for EPB formation is a threshold vertical ion drift from the E region, which is required to produce the vertical plasma gradients conducive to this instability. Factors affecting the variation of EPBs therefore include magnetic declination, the strength of the equatorial electojet, and the wind dynamo in the lower thermosphere controlling vertical plasma drifts. In most longitude zones, this results in elevated EPB occurrence rates during the equinoxes. The notable exception is over the central Pacific and African sectors, where EPB activity maximizes during solstice. \citet{tsunoda_jgr2015} hypothesized that the solstice maxima in these two sectors could be driven by a zonal wavenumber 2 atmospheric tide in the mesosphere and lower thermosphere. In this study, we find that the post-sunset electron density observed by FORMOSAT-3/COSMIC during the boreal summer from 2007 - 2012 does indeed exhibit a wave-2 zonal distribution, consistent with results expected from elevated vertical ion drift over the Central Pacific and African sectors. Numerical experiments are also carried out which found that forcing from the aforementioned tidal and stationary planetary wave (SPW) components produced wave-2 modulations on vertical ion drift, ion flux convergence, and midnight TEC. The relation between the vertical ion drift enhancements and the midnight TEC enhancements are consistent with the solstice maxima hypothesis.

This work explores the results of applying Square Euclidean and Correlation k-means cluster analysis on ion density irregularity profiles observed by the Advanced Ionospheric Probe (AIP) onboard the Formosat-5 (F-5) satellite from November 2017 to November 2020. The Square Euclidean cluster analysis yielded separate clusters each for ion density irregularities consistent with Equatorial Plasma Bubbles (EPBs) occurring over the southern low-latitude, northern low-latitude and equator. The Correlation k-means cluster analysis only yielded one cluster characterized by ion density irregularities consistent with EPBs occurring over the equator. Thus, this work suggests that a cluster analysis preferably a Square Euclidean Cluster analysis can be used to find and classify ion density irregularities consistent with EPBs. This work also shows that the F-5/AIP can perform multi-year observations of ion density irregularities at ~710 km altitude and at ~2230 pm local-time that are consistent with irregularities due to EPBs.