Substorms are a phenomenon in the magnetosphere–ionosphere system, which are characterised by brightening of an auroral arc and enhancement of electric currents in the polar ionosphere. Since substorms are difficult to predict, this study treats a substorm occurrence as a stochastic phenomenon and represents the substorm occurrence rate with a machine learning model. We then analyse the response of substorm activity to solar wind conditions by feeding synthetic solar wind data into the model.

The relationships between auroral activity and the solar-wind conditions are modeled with a machine-learning technique. The impact of various solar-wind parameters on the auroral activity is then evaluated by putting artificial inputs into the trained machine-learning model. One of the notable findings is that the solar-wind density effect on the auroral activity is emphasized under high solar-wind speed and weak solar-wind magnetic field.

Old diaries kept in Japan tell us a surprising fact. The 27-day solar rotational period in thunder and lightning activities had been persistent for the past 300 years. The intensity is found to be more prominent as solar activity increases. The physical mechanism of the Sun–Climate connection is yet uncertain, an important link surely exists between the solar activity and terrestrial climate even at a meteorological timescale.

Solar activity and climate show correlations over a wide range of timescales. It is important to understand the behavior of the 27-day solar rotational period in lightning activities because it provides an opportunity to understand how the sun influences weather and climate. We analyzed lightning data extracted from diaries written in Kyoto, Japan from the mid-17th to the mid-18th century. Lightning shows the signal of the 27-day period; however, it disappeared during the Maunder Minimum.

Detailed analyses of the 27-day solar rotational period in cloud and lightning activities may help in untangling the process of solar influence on weather and climate. We analyzed the lightning data in Japan for AD 1989–2015 and found that the 27-day solar rotational period is seen in wide-area lightning activity. The signal was stronger at the maxima of solar decadal cycles. It was also found that the signal of the 27-day period migrates from the southwest to the northeast in Japan.