10 likes | 147 Vues
This study presents a method for reconstructing solar spectral irradiance, which is crucial for understanding the Earth's atmospheric and oceanic heating sources. Utilizing a backpropagation neural network, the model correlates irradiance variations with solar magnetic field evolution. By training the network with inputs from calibrated/uncalibrated time series from the LYRA instrument, solar magnetograms, and various solar indexes, we can effectively approximate solar irradiance observations, vital for global and regional climate modeling.
E N D
Reconstruction of solar spectral irradiance employing LYRA DataL.E.A. Vieira , T. Dudok de Wit, M. Kretzschmar, G. Cessateur • Context: The solar energy input is the main external heating source of the Earth’s atmospheric/oceanic system. Thus, variations of the solar emission produce corresponding changes of the weather and climate. • Aims: We present a reconstruction of the current solar spectral irradiance based on observations of solar parameters, which can be used to constrain the evolution of global and regional circulation models. • Method: We employ a backpropagation neural network to model the evolution of solar irradiance. The basic assumption of this model is that the variations of the irradiance are related to the evolution of the solar magnetic field. The input parameters and the corresponding targets are used to train a network until it can approximate the observations. • Input Parameters: • Calibrated/Uncalibrated time series of the four channels observed by LYRA instrument; • Solar disk magnetograms and continuum Images; • Solar indexes (e.g. Magnesium-II Index). • Targets: • SORCE/SOLSTICE and TIMED/SEE Observations; • Total Solar Irradiance (TIM/SORCE, VIRGO, ACRIM III).