Journal cover Journal topic
Atmospheric Measurement Techniques An interactive open-access journal of the European Geosciences Union
Journal topic

Journal metrics

Journal metrics

  • IF value: 3.248 IF 3.248
  • IF 5-year value: 3.650 IF 5-year 3.650
  • CiteScore value: 3.37 CiteScore 3.37
  • SNIP value: 1.253 SNIP 1.253
  • SJR value: 1.869 SJR 1.869
  • IPP value: 3.29 IPP 3.29
  • h5-index value: 47 h5-index 47
  • Scimago H index value: 60 Scimago H index 60
Discussion papers
© Author(s) 2019. This work is distributed under
the Creative Commons Attribution 4.0 License.
© Author(s) 2019. This work is distributed under
the Creative Commons Attribution 4.0 License.

Research article 18 Jan 2019

Research article | 18 Jan 2019

Review status
This discussion paper is a preprint. It is a manuscript under review for the journal Atmospheric Measurement Techniques (AMT).

Neural Network Radiative Transfer for Imaging Spectroscopy

Brian D. Bue1, David R. Thompson1, Shubhankar Deshpande2, Michael Eastwood1, Robert O. Green1, Terry Mullen3, Vijay Natraj1, and Mario Parente3 Brian D. Bue et al.
  • 1Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA, USA
  • 2The Robotics Institute, Carnegie Mellon University, Pittsburgh, PA, USA
  • 3University of Massachusetts, Amherst, MA, USA

Abstract. Visible/Shortwave InfraRed imaging spectroscopy provides valuable remote measurements of Earth's surface and atmospheric properties. These measurements generally rely on inversions of computationally-intensive Radiative Transfer Models (RTMs). RTMs' computational expense makes them difficult to use with high volume imaging spectrometers, and forces approximations such as lookup table interpolation and surface/atmosphere decoupling. These compromises limit the accuracy and flexibility of the remote retrieval; dramatic speed improvements in radiative transfer models could significantly improve the utility and interpretability of remote spectroscopy for Earth science. This study demonstrates that nonparametric function approximation with neural networks can replicate Radiative Transfer calculations over a relevant range of surface/atmosphere parameters. Incorporating physical knowledge into the network structure provides improved interpretability and model efficiency. We evaluate the approach in atmospheric correction of data from the PRISM airborne imaging spectrometer, and demonstrate accurate emulation of radiative transfer calculations which run several orders of magnitude faster than first-principles models. These results are particularly amenable to iterative spectrum fitting approaches, providing analytical benefits including statistically rigorous treatment of uncertainty and the potential to recover information on spectrally-broad signals.

Brian D. Bue et al.
Interactive discussion
Status: open (until 15 Mar 2019)
Status: open (until 15 Mar 2019)
AC: Author comment | RC: Referee comment | SC: Short comment | EC: Editor comment
[Subscribe to comment alert] Printer-friendly Version - Printer-friendly version Supplement - Supplement
Brian D. Bue et al.
Model code and software

dsmbgu8/isofit: Neural Network Radiative Transfer Emulation for ISOFIT D. R. Thompson and B. D. Bue

Brian D. Bue et al.
Total article views: 249 (including HTML, PDF, and XML)
HTML PDF XML Total BibTeX EndNote
189 55 5 249 1 2
  • HTML: 189
  • PDF: 55
  • XML: 5
  • Total: 249
  • BibTeX: 1
  • EndNote: 2
Views and downloads (calculated since 18 Jan 2019)
Cumulative views and downloads (calculated since 18 Jan 2019)
Viewed (geographical distribution)  
Total article views: 174 (including HTML, PDF, and XML) Thereof 174 with geography defined and 0 with unknown origin.
Country # Views %
  • 1
No saved metrics found.
No discussed metrics found.
Latest update: 22 Feb 2019
Publications Copernicus
Short summary
Imaging spectrometers provide valuable remote measurements of Earth's surface and atmosphere. These measurements rely on computationally-expensive Radiative Transfer Models (RTMs). Spectrometers produce too much data to process with RTMs directly, requiring approximations that trade accuracy for speed. We demonstrate that neural networks can quickly emulate RTM calculations more accurately than current approaches, enabling the application of more sophisticated RTMs than current methods permit.
Imaging spectrometers provide valuable remote measurements of Earth's surface and atmosphere....