A deep learning approach to fast radiative transfer. (April 2022)

• Record Type: Journal Article
• Authors: Stegmann, Patrick G.; Johnson, Benjamin; Moradi, Isaac; Karpowicz, Bryan; McCarty, Will
• Is Part Of: Journal of quantitative spectroscopy & radiative transfer. Volume 280(2022)
• Publication Date: 2022-04
• Subjects: Machine learning -- Deep learning -- Radiative transfer -- Infrared radiation -- Transmittance; Spectrum analysis -- Periodicals; Radiation -- Periodicals; Analyse spectrale -- Périodiques; Rayonnement -- Périodiques; Radiation; Spectrum analysis; Periodicals; 543.0858
• DOI: 10.1016/j.jqsrt.2022.108088
• Languages: English
• ISSNs: 0022-4073

Abstract

Highlights: A statistical regression approach for fast transmittance modeling is proposed. The structure of the line-by-line radiative transfer regression process is outlined. Two deep learning, hidden-layer neural networks are compared as regression models according to their performance. Fitting accuracy and computation speed of the neural networks are shown to be excellent. Abstract: Due to the sheer volume of data, leveraging satellite instrument observations effectively in a data assimilation context for numerical weather prediction or for remote sensing requires a radiative transfer model as an observation operator that is both fast and accurate at the same time. Physics-based line-by-line radiative transfer (RT) models fulfil the requirement for accuracy, but are too slow and too costly in computational terms for operational applications. Therefore, fast methods were developed to be able to perform fast RT calculations using techniques such as spectral sampling or pre-computed look-up tables. The operational fast models currently calculate the absorption and scattering coefficients from the pre-computed regression coefficients and atmospheric state and cloud profiles. As a novel solution to this problem, this work investigates a deep learning approach to replace the regression coefficients in the fast RT models. A selection of hidden-layer neural network configurations is trained against atmospheric transmittance profile data computed by an accurate line-by-line model and their performance is evaluated and their advantages and disadvantages are discussed.