Correlations between SAR arc intensity and solar and geomagnetic activity

We present a study of statistical relationships between SAR arc intensities acquired by the Paci®c Northwest Laboratory Photometer Network during 1978±1988 and solar and geomagnetic activity indices Dst; F 10:7, and Kp by use of the method of multiple regression analysis. We found signi®cant correlations between intensity and all of the indices involved. In the present work we show for the ®rst time that the partial correlation coecients depend on the time o€set, t, between the time of SAR arc intensity observations and the onset of the geomagnetic storm recovery phase, with the largest correlations being observed when 8 h t 16 h. It is also shown that there are signi®cant di€erences between partial correlation coecients calculated for SAR arcs associated with strong …Dstmin ÿ100 nT† and weak …Dstmin > ÿ100 nT† geomagnetic storms. We observe also that the multiple correlation coecients for strong storms are much larger than for weak ones. We found that the variations in the electron temperature, Te, in the SAR arc region are not mainly produced by variations in the electron density of the ionosphere but are strongly driven by the additional heating of the electron gas due to an interaction of the ring current ions and the plasmaspheric electrons. As a result, variations of Te in the SAR arc region with characteristic time scales from several minutes to several hours are stipulated by time variations of ring current parameters. Key words. Atmospheric composition and structure (airglow and aurora) á Ionosphere (ionosphere± atmosphere interactions; ionosphere±magnetosphere interactions). 1 Introduction During the recovery phase of large magnetic storms characterized by a geomagnetic index Kp of 5 or greater, stable auroral red (SAR) arcs are observed optically on the equatorward edge of the midlatitude ionospheric trough (Rees and Roble, 1975; Kozyra et al., 1997). The characteristic spectral emission at 630 nm that identi®es a SAR arc is associated with the atomic oxygen transition OI …P2 ÿ D2†. The centre of the SAR arc emission is usually observed in the region between 350 and 450 km in altitude. The ring current is a primary energy source for maintenance of SAR arcs (Kozyra et al., 1997). As a result, the observed SAR arc intensity is in ̄uenced by a number of factors associated with the energy exchange and transformation in the ring currentplasmasphere-atmosphere system. Using a small data set of bright SAR arc observations performed during the solar cycle maximum in 1956±1960, Rees and Akasofu (1963) and Roach and Roach (1963) found that there are correlations of the SAR arc intensities with Dst and Kp. It was also noted that the frequency of SAR arc occurrence is greatest during the solar cycle maximum and this frequency is diminished greatly as the Sun proceeded through the solar cycle minimum (Rees and Roble, 1975; Rees and Akasofu, 1963; Slater and Kleckner, 1989; Kozyra et al., 1997). The main goal of the present study is to ®nd the statistical relationships between SAR arc intensities and solar and geomagnetic activity measured in terms of indices Dst, F 10:7, and Kp, using the data acquired by the Paci®c Northwest Laboratory Photometer Network during 1978±1988 (more than 1000 h of measurements) (Slater and Kleckner, 1989). The statistical study of large data sets can shed some light on mechanisms of energy transfer from the ring current to the plasmasphere and ionosphere and investigate which formulas relating SAR arc intensities with the solar and geomagCorrespondence to: V. V. Lobzin Ann. Geophysicae 17, 770±781 (1999) Ó EGS ± Springer-Verlag 1999