Preprints
https://doi.org/10.5194/amt-2024-63
https://doi.org/10.5194/amt-2024-63
03 May 2024
 | 03 May 2024
Status: this preprint is currently under review for the journal AMT.

Forward Model Emulator for Atmospheric Radiative Transfer Using Gaussian Processes And Cross Validation

Otto M. Lamminpää, Jouni I. Susiluoto, Jonathan M. Hobbs, James L. McDuffie, Amy J. Braverman, and Houman Owhadi

Abstract. Remote sensing of atmospheric carbon dioxide (CO2) carried out by NASA’s Orbiting Carbon Observatory-2 (OCO-2) satellite mission and the related Uncertainty Quantification (UQ) effort involves repeated evaluations of a state-of-the-art atmospheric physics model. The retrieval, or solving an inverse problem, requires substantial computational resources. In this work, we propose and implement a statistical emulator to speed up the computations in the OCO-2 physics model. Our approach is based on Gaussian Process (GP) Regression, leveraging recent research on Kernel Flows and Cross Validation to efficiently learn the kernel function in the GP. We demonstrate our method by replicating the behavior of OCO-2 forward model within measurement error precision, and further show that in simulated cases, our method reproduces the CO2 retrieval performance of OCO-2 setup with orders of magnitude faster computational time. The underlying emulation problem is challenging because it is high dimensional. It is related to operator learning in the sense that the function to be approximated is mapping high-dimensional vectors to high-dimensional vectors. Our proposed approach is not only fast but also highly accurate (its relative error is less than 1 %). In contrast with Artificial Neural Network (ANN) based methods, it is interpretable and its efficiency is based on learning a kernel in an engineered and expressive family of kernels.

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this preprint. The responsibility to include appropriate place names lies with the authors.
Otto M. Lamminpää, Jouni I. Susiluoto, Jonathan M. Hobbs, James L. McDuffie, Amy J. Braverman, and Houman Owhadi

Status: open (until 04 Jul 2024)

Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor | : Report abuse
  • RC1: 'Comment on amt-2024-63', Hristo Georgiev Chipilski, 04 May 2024 reply
Otto M. Lamminpää, Jouni I. Susiluoto, Jonathan M. Hobbs, James L. McDuffie, Amy J. Braverman, and Houman Owhadi
Otto M. Lamminpää, Jouni I. Susiluoto, Jonathan M. Hobbs, James L. McDuffie, Amy J. Braverman, and Houman Owhadi

Viewed

Total article views: 293 (including HTML, PDF, and XML)
HTML PDF XML Total BibTeX EndNote
199 86 8 293 10 6
  • HTML: 199
  • PDF: 86
  • XML: 8
  • Total: 293
  • BibTeX: 10
  • EndNote: 6
Views and downloads (calculated since 03 May 2024)
Cumulative views and downloads (calculated since 03 May 2024)

Viewed (geographical distribution)

Total article views: 284 (including HTML, PDF, and XML) Thereof 284 with geography defined and 0 with unknown origin.
Country # Views %
  • 1
1
 
 
 
 
Latest update: 29 May 2024
Download
Short summary
We develop and demonstrate a fast forward function emulator for remote sensing of greenhouse gases. These forward functions are computationally expensive to evaluate, and as such the key challenge for many satellite missions in their data processing is the time used in these evaluations. Our method is fast and accurate enough, less than 1 % relative error, so that it could be safely used in operational processing.