Multivariate regression analysis of atmospheric transparency indicators during powerful magnetic storms in solar cycle 25

A multiple linear regression method was employed to assess the degree of influence of solar factors on changes in atmospheric transparency indicators during powerful magnetic storms in March and April 2023. Regression models were constructed using a two-factor linear regression equation, in which the influencing factors were the ionospheric electric potential (EP) contrast and cosmic rays (CRs). It was shown that the determination coefficient increases in the constructed regression models, i.e. the selected factors increase the proportion of explained dispersion in atmospheric transparency characteristics. We assessed the degree of influence associated with each explanatory factor included in the regression models. It was found that CR is the factor that explains the largest proportion of dispersion in cloud parameters in the maximum water vapor layer (700–500 hPa). It was established that the relative contribution of CRs to the increase in explained dispersion of cloud parameter is equivalent to the contribution of the EP contrast, which explains the largest proportion of the dispersion in the contrast of outgoing longwave radiation. The obtained results illustrate the role of CRs in the electrical mechanism of the nonlinear impact of solar activity on the climate system, whose theoretical foundations are being developed at ISTP SB RAS.