Deep transfer learning of global spectra for local soil carbon monitoring

There is global interest in spectroscopy and the development of large and diverse soil spectral libraries (SSL) to model soil organic carbon (SOC) and monitor, report, and verify (MRV) its changes. The reason is that increasing SOC can improve food production and mitigate climate change. However, ‘g...

Full description

Bibliographic Details
Main Authors: Shen, Zefang, Ramirez-Lopez, Leonardo, Behrens, Thorsten, Cui, Lei, Zhang, Mingxi, Walden, Lewis, Wetterlind, Johana, Shi, Zhou, Sudduth, Kenneth, Baumann, Philipp, Song, Yongze, Catambay, Kevin, Viscarra Rossel, Raphael
Format: Journal Article
Published: Elsevier 2022
Online Access:http://hdl.handle.net/20.500.11937/88543
_version_ 1848765039288254464
author Shen, Zefang
Ramirez-Lopez, Leonardo
Behrens, Thorsten
Cui, Lei
Zhang, Mingxi
Walden, Lewis
Wetterlind, Johana
Shi, Zhou
Sudduth, Kenneth
Baumann, Philipp
Song, Yongze
Catambay, Kevin
Viscarra Rossel, Raphael
author_facet Shen, Zefang
Ramirez-Lopez, Leonardo
Behrens, Thorsten
Cui, Lei
Zhang, Mingxi
Walden, Lewis
Wetterlind, Johana
Shi, Zhou
Sudduth, Kenneth
Baumann, Philipp
Song, Yongze
Catambay, Kevin
Viscarra Rossel, Raphael
author_sort Shen, Zefang
building Curtin Institutional Repository
collection Online Access
description There is global interest in spectroscopy and the development of large and diverse soil spectral libraries (SSL) to model soil organic carbon (SOC) and monitor, report, and verify (MRV) its changes. The reason is that increasing SOC can improve food production and mitigate climate change. However, ‘global’ modelling of SOC with such diverse and hyperdimensional SSLs do not generalise well locally, e.g. at a field scale. To address this challenge, we propose deep transfer learning (DTL) to leverage useful information from large-scale SSLs to assist local modelling. We used one global, three country-specific SSLs and data from three local sites with DTL to improve the modelling and localise the SOC estimates in individual fields or farms in each country. With DTL, we transferred instances from the SSLs, representations from one-dimensional convolutional neural networks (1D-CNNs) trained on the SSLs, and both instances and representations to improve local modelling. Transferring instances effectively used information from the global SSL to most accurately estimate SOC in each site, reducing the root mean square error (RMSE) by 25.8% on average compared with local modelling. Our results highlight the effectiveness of DTL and the value of diverse, global SSLs for accurate local SOC predictions. Applying DTL with a global SSL one could estimate SOC anywhere in the world more accurately, rapidly, and cost-effectively, enabling MRV protocols to monitor SOC changes.
first_indexed 2025-11-14T11:28:54Z
format Journal Article
id curtin-20.500.11937-88543
institution Curtin University Malaysia
institution_category Local University
last_indexed 2025-11-14T11:28:54Z
publishDate 2022
publisher Elsevier
recordtype eprints
repository_type Digital Repository
spelling curtin-20.500.11937-885432022-06-22T04:49:04Z Deep transfer learning of global spectra for local soil carbon monitoring Shen, Zefang Ramirez-Lopez, Leonardo Behrens, Thorsten Cui, Lei Zhang, Mingxi Walden, Lewis Wetterlind, Johana Shi, Zhou Sudduth, Kenneth Baumann, Philipp Song, Yongze Catambay, Kevin Viscarra Rossel, Raphael There is global interest in spectroscopy and the development of large and diverse soil spectral libraries (SSL) to model soil organic carbon (SOC) and monitor, report, and verify (MRV) its changes. The reason is that increasing SOC can improve food production and mitigate climate change. However, ‘global’ modelling of SOC with such diverse and hyperdimensional SSLs do not generalise well locally, e.g. at a field scale. To address this challenge, we propose deep transfer learning (DTL) to leverage useful information from large-scale SSLs to assist local modelling. We used one global, three country-specific SSLs and data from three local sites with DTL to improve the modelling and localise the SOC estimates in individual fields or farms in each country. With DTL, we transferred instances from the SSLs, representations from one-dimensional convolutional neural networks (1D-CNNs) trained on the SSLs, and both instances and representations to improve local modelling. Transferring instances effectively used information from the global SSL to most accurately estimate SOC in each site, reducing the root mean square error (RMSE) by 25.8% on average compared with local modelling. Our results highlight the effectiveness of DTL and the value of diverse, global SSLs for accurate local SOC predictions. Applying DTL with a global SSL one could estimate SOC anywhere in the world more accurately, rapidly, and cost-effectively, enabling MRV protocols to monitor SOC changes. 2022 Journal Article http://hdl.handle.net/20.500.11937/88543 10.1016/j.isprsjprs.2022.04.009 http://creativecommons.org/licenses/by-nc-nd/4.0/ Elsevier fulltext
spellingShingle Shen, Zefang
Ramirez-Lopez, Leonardo
Behrens, Thorsten
Cui, Lei
Zhang, Mingxi
Walden, Lewis
Wetterlind, Johana
Shi, Zhou
Sudduth, Kenneth
Baumann, Philipp
Song, Yongze
Catambay, Kevin
Viscarra Rossel, Raphael
Deep transfer learning of global spectra for local soil carbon monitoring
title Deep transfer learning of global spectra for local soil carbon monitoring
title_full Deep transfer learning of global spectra for local soil carbon monitoring
title_fullStr Deep transfer learning of global spectra for local soil carbon monitoring
title_full_unstemmed Deep transfer learning of global spectra for local soil carbon monitoring
title_short Deep transfer learning of global spectra for local soil carbon monitoring
title_sort deep transfer learning of global spectra for local soil carbon monitoring
url http://hdl.handle.net/20.500.11937/88543