Multiple regression analysis of changes in upper atmosphere characteristics based on observations in Eastern Siberia in solar cycle 24

We present the results of the analysis of changes in the parameters of the upper neutral atmosphere and ionosphere in the Eastern Siberia region in solar cycle 24 with using long-term data from the instrument complex of the Institute of Solar-Terrestrial Physics SB RAS. The analysis is based on data from spectrometric and radiophysical measurements in 2008-2020. We used the data on the temperature of the mesopause region Tm from spectrometric measurements of the hydroxyl emission (band (6-2), 834.0 nm, the height of the maximum emission is ~87 km) at the ISTP SB RAS Geophysical Observatory (51.8° N, 103.1° E, Tory) and data on the peak electron density NmF2 from vertical sounding data on the Irkutsk ionosonde DPS-4 (52.3° N, 104.3° E). The rotational temperature of the hydroxyl molecule obtained from the recorded spectra corresponds the atmospheric temperature at the mesopause altitudes. For the analysis we used the multiple regression method. The analysis was implemented on the basis of multiple linear regressions of the upper atmosphere characteristics on the indices of solar and geomagnetic activity, as well as indices of various atmospheric oscillations. It was found, that the year-to-year changes in the average annual NmF2 are mainly due to changes in solar activity (F10.7). Both solar (F10.7) and geomagnetic (Ap) activities affect variations in the average annual values of σNmF2 variability. Unlike ionospheric characteristics, the year-to-year changes in the mesopause temperature and its variability weakly correlate with the F10.7 and Ap indices. For further analysis of the mesopause temperature behavior, the North Atlantic (NAO), Arctic (AO) and Southern (SOI) atmospheric oscillation indices were involved, which led to a significant increase in the determination coefficients. Thus, for year-to-year changes in the mesopause temperature, the determination coefficient for regression on the AO and F10.7 indices reached 71.8, and for the NAO and F10.7 indices was 71.4, which is ~8 times higher than the determination coefficient for regression on the F10.7 and Ap indices. For temperature variability σTm, the highest determination coefficients were obtained for multiple regression on SOI and Ap. The obtained results may indicate a significant impact of processes in the lower atmosphere on the temperature regime of the mid-latitude mesopause region in the analyzed time interval.

The study was supported by the grant of the Russian Science Foundation No. 25-17-00187, https://rscf.ru/project/25-17-00187.