Gradient boosting machine learning to improve satellite-derived column water vapor measurement error

Just, Allan C.; Liu, Yang; Sorek-Hamer, Meytar; Rush, Johnathan; Dorman, Michael; Chatfield, Robert; Wang, Yujie; Lyapustin, Alexei; Kloog, Itai

doi:https://doi.org/10.5194/amt-13-4669-2020

Articles | Volume 13, issue 9

https://doi.org/10.5194/amt-13-4669-2020

© Author(s) 2020. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/amt-13-4669-2020

© Author(s) 2020. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 13, issue 9

Research article

|

02 Sep 2020

Research article |

| 02 Sep 2020

Gradient boosting machine learning to improve satellite-derived column water vapor measurement error

Allan C. Just, Yang Liu, Meytar Sorek-Hamer, Johnathan Rush, Michael Dorman, Robert Chatfield, Yujie Wang, Alexei Lyapustin, and Itai Kloog

Download

Final revised paper (published on 02 Sep 2020)
Supplement to the final revised paper
Preprint (discussion started on 11 Dec 2019)
Supplement to the preprint

Interactive discussion

Status: closed

AC: Author comment | RC: Referee comment | SC: Short comment | EC: Editor comment

- Printer-friendly version

- Supplement

RC1: 'Review', Anonymous Referee #2, 29 Jan 2020
- AC1: 'Response for RC1', Allan Just, 12 Mar 2020
RC2: 'review', Anonymous Referee #3, 26 Feb 2020
- AC2: 'Response for RC2', Allan Just, 12 Mar 2020

Peer-review completion

AR: Author's response | RR: Referee report | ED: Editor decision

AR by Allan Just on behalf of the Authors (05 May 2020) Author's response Manuscript

ED: Publish as is (06 May 2020) by Alexander Kokhanovsky

AR by Allan Just on behalf of the Authors (10 May 2020)

Short summary

A flexible machine-learning model was fit to explain the differences between estimates of water vapor from satellites versus ground stations in Northeastern USA. We use nine variables derived from the satellite acquisition and ground characteristics to explain this measurement error. Our results showed overall good agreement, but data from the Terra satellite were drifting too high in recent summers. Our model reduces measurement error and works well in new locations in the northeast.