Temporal Variations in Solar Irradiance Since 1947

The study of variations in total solar irradiance (TSI) and spectral irradiance is important for understanding how the Sun affects the Earth’s climate. A data-driven approach is used in this article to analyze and model the temporal variation of the TSI and Mg?ii index back to 1947. In both cases, observed data in the time interval of the satellite era, 1978?–?2013, were used for neural network (NN) model-design and testing. For this particular purpose, the evolution of the solar magnetic field is assumed to be the main driver for the day-to-day irradiance variability. First, we design a model for the Mg?ii index data from F10.7 cm solar radio-flux using the NN approach in the time span of 1978 through 2013. Results of Mg?ii index model were tested using various numbers of hidden nodes. The predicted values of the hidden layer with five nodes correspond well to the composite Mg?ii values. The model reproduces 94% of the variability in the composite Mg?ii index, including the secular decline between the 1996 and 2008 solar cycle minima. Finally, the extrapolation of the Mg?ii index was performed using the developed model from F10.7 cm back to 1947. Similarly, the NN model was designed for TSI variability study over the time span of the satellite era using data from the Physikalisch-Meteorologisches Observatorium Davos (PMOD) as a target, and solar activity indices as model inputs. This model was able to reproduce the daily irradiance variations with a correlation coefficient of 0.937 from sunspot and facular measurements in the time span of 1978?–?2013. Finally, the temporal variation of the TSI was analyzed using the designed NN model back to 1947 from the Photometric Sunspot Index (PSI) and the extrapolated Mg?ii index. The extrapolated TSI result indicates that the amplitudes of Solar Cycles 19 and 21 are closely comparable to each other, and Solar Cycle 20 appears to be of lower irradiance during its maximum.