Using deep learning for digital soil mapping

Padarian, José; Minasny, Budiman; McBratney, Alex B.

doi:https://doi.org/10.5194/soil-5-79-2019

Articles | Volume 5, issue 1

https://doi.org/10.5194/soil-5-79-2019

© Author(s) 2019. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/soil-5-79-2019

© Author(s) 2019. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 5, issue 1

Original research article

|

26 Feb 2019

Original research article |

| 26 Feb 2019

Using deep learning for digital soil mapping

José Padarian, Budiman Minasny, and Alex B. McBratney

Download

Final revised paper (published on 26 Feb 2019)
Preprint (discussion started on 03 Sep 2018)

Interactive discussion

Status: closed

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

- Printer-friendly version

- Supplement

RC1: 'review of "Using deep learning for Digital Soil Mapping"', Anonymous Referee #1, 30 Sep 2018
- AC1: 'Response to RC1', José Padarian, 08 Oct 2018
  - RC2: 'Further comments on authors' response', Anonymous Referee #1, 31 Oct 2018
    - AC2: 'Reply', José Padarian, 30 Nov 2018
RC3: 'Comments on Padarian et al. paper', Anonymous Referee #2, 25 Nov 2018
- AC3: 'Response to RC3', José Padarian, 30 Nov 2018

Peer-review completion

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

ED: Publish subject to minor revisions (review by editor) (10 Jan 2019) by Bas van Wesemael

AR by José Padarian on behalf of the Authors (18 Jan 2019) Author's response Manuscript

ED: Publish as is (05 Feb 2019) by Bas van Wesemael

ED: Publish subject to technical corrections (07 Feb 2019) by Kristof Van Oost (Executive editor)

AR by José Padarian on behalf of the Authors (08 Feb 2019) Author's response Manuscript

Short summary

Digital soil mapping has been widely used as a cost-effective method for generating soil maps. DSM models are usually calibrated using point observations and rarely incorporate contextual information of the landscape. Here, we use convolutional neural networks to incorporate spatial context. We used as input a 3-D stack of covariate images to simultaneously predict organic carbon content at multiple depths. In this study, our model reduced the error by 30 % compared with conventional techniques.