Deep learning of the particulate and mineral-associated organic carbon fractions using a compositional transform and mid-infrared spectroscopy

We need soil organic carbon (SOC) and the SOC fractions, the particulate and mineral-associated organic carbon (POC, MAOC), to understand SOC dynamics. They have implications for soil management, carbon sequestration and climate change mitigation. However, conventional laboratory measurements of the...

Full description

Saved in:
Bibliographic Details
Main Authors: Mingxi Zhang, Zefang Shen, Lewis Walden, Farid Sepanta, Zhongkui Luo, Lei Gao, Oscar Serrano, Raphael A. Viscarra Rossel
Format: Article
Language:English
Published: Elsevier 2025-03-01
Series:Geoderma
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S001670612500045X
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850070111777456128
author Mingxi Zhang
Zefang Shen
Lewis Walden
Farid Sepanta
Zhongkui Luo
Lei Gao
Oscar Serrano
Raphael A. Viscarra Rossel
author_facet Mingxi Zhang
Zefang Shen
Lewis Walden
Farid Sepanta
Zhongkui Luo
Lei Gao
Oscar Serrano
Raphael A. Viscarra Rossel
author_sort Mingxi Zhang
collection DOAJ
description We need soil organic carbon (SOC) and the SOC fractions, the particulate and mineral-associated organic carbon (POC, MAOC), to understand SOC dynamics. They have implications for soil management, carbon sequestration and climate change mitigation. However, conventional laboratory measurements of the SOC fractions, which involve physical or chemical separations, are elaborate, time-consuming and expensive. Mid-infrared (MIR) spectroscopy combined with multivariate modelling can alleviate these limitations because the method can estimate SOC and its fractions rapidly, cost-effectively and accurately. Previous spectroscopic modelling has mostly ignored the compositional nature of the SOC fractions (i.e. SOC = ∑fractions), causing discrepancies in the estimation such that the sum of the fractions does not equal the total SOC. We recorded the MIR spectra (4000–450 cm−1) of 397 soil samples from across Australia and then performed a granulometric fractionation to derive three SOC fractions, the POC in the macroaggregates (250–2000μm, POCmac), POC in the micro-aggregates (50–250μm, POCmic), and MAOC (<50μm). We used the centred log ratio (CLR) method to transform the data compositionally and then modelled POCmac, POCmic, POC (POCmac + POCmic), and MAOC with the spectra, using convolutional neural networks (CNN) and cubist for benchmarking. We interpreted the models using the SHapley Additive exPlanations (SHAP) values and a land use classification of the data. Modelling the CLR-transformed SOC fractions with CNN maintained the composition of the fractions and improved the accuracy of the estimates (Lin’s concordance correlation coefficient (ρc) of 0.58, 0.86, and 0.94 for the POCmac, POCmic, and MAOC), compared to CLR with cubist (ρc of 0.49, 0.84, and 0.87 for the POCmac, POCmic, and MAOC) and cubist with no compositional transformation (ρc of 0.53, 0.85, and 0.88 for the POCmac, POCmic, and MAOC). The SHAP values reflected the compositional modelling and identified important organic and inorganic functional groups that differed by fraction and land use. Our approach can complement conventional physical SOC fractionations and improve the cost-effectiveness of the measurements, especially when there are many samples to measure, thus enhancing our understanding of SOC dynamics.
format Article
id doaj-art-c1bdc9e624d04c0aa47eb33c47f71fb1
institution DOAJ
issn 1872-6259
language English
publishDate 2025-03-01
publisher Elsevier
record_format Article
series Geoderma
spelling doaj-art-c1bdc9e624d04c0aa47eb33c47f71fb12025-08-20T02:47:37ZengElsevierGeoderma1872-62592025-03-0145511720710.1016/j.geoderma.2025.117207Deep learning of the particulate and mineral-associated organic carbon fractions using a compositional transform and mid-infrared spectroscopyMingxi Zhang0Zefang Shen1Lewis Walden2Farid Sepanta3Zhongkui Luo4Lei Gao5Oscar Serrano6Raphael A. Viscarra Rossel7Soil &amp; Landscape Science, School of Molecular &amp; Life Sciences, Faculty of Science &amp; Engineering, Curtin University, GPO Box U1987, Perth WA 6845, AustraliaSoil &amp; Landscape Science, School of Molecular &amp; Life Sciences, Faculty of Science &amp; Engineering, Curtin University, GPO Box U1987, Perth WA 6845, AustraliaSoil &amp; Landscape Science, School of Molecular &amp; Life Sciences, Faculty of Science &amp; Engineering, Curtin University, GPO Box U1987, Perth WA 6845, AustraliaSoil &amp; Landscape Science, School of Molecular &amp; Life Sciences, Faculty of Science &amp; Engineering, Curtin University, GPO Box U1987, Perth WA 6845, AustraliaCollege of Environmental and Resource Sciences, Zhejiang University, Hangzhou 310058, ChinaEnvironment Business Unit, Commonwealth Scientific and Industrial Research Organisation, Waite Campus, Urrbrae, SA 5064, AustraliaCentro de Estudios Avanzados de Blanes, Consejo Superior de Investigaciones Científicas, Blanes 17300, Spain; School of Science &amp; Centre for Marine Ecosystems Research, Edith Cowan University, Joondalup, WA 6027, AustraliaSoil &amp; Landscape Science, School of Molecular &amp; Life Sciences, Faculty of Science &amp; Engineering, Curtin University, GPO Box U1987, Perth WA 6845, Australia; Corresponding author.We need soil organic carbon (SOC) and the SOC fractions, the particulate and mineral-associated organic carbon (POC, MAOC), to understand SOC dynamics. They have implications for soil management, carbon sequestration and climate change mitigation. However, conventional laboratory measurements of the SOC fractions, which involve physical or chemical separations, are elaborate, time-consuming and expensive. Mid-infrared (MIR) spectroscopy combined with multivariate modelling can alleviate these limitations because the method can estimate SOC and its fractions rapidly, cost-effectively and accurately. Previous spectroscopic modelling has mostly ignored the compositional nature of the SOC fractions (i.e. SOC = ∑fractions), causing discrepancies in the estimation such that the sum of the fractions does not equal the total SOC. We recorded the MIR spectra (4000–450 cm−1) of 397 soil samples from across Australia and then performed a granulometric fractionation to derive three SOC fractions, the POC in the macroaggregates (250–2000μm, POCmac), POC in the micro-aggregates (50–250μm, POCmic), and MAOC (<50μm). We used the centred log ratio (CLR) method to transform the data compositionally and then modelled POCmac, POCmic, POC (POCmac + POCmic), and MAOC with the spectra, using convolutional neural networks (CNN) and cubist for benchmarking. We interpreted the models using the SHapley Additive exPlanations (SHAP) values and a land use classification of the data. Modelling the CLR-transformed SOC fractions with CNN maintained the composition of the fractions and improved the accuracy of the estimates (Lin’s concordance correlation coefficient (ρc) of 0.58, 0.86, and 0.94 for the POCmac, POCmic, and MAOC), compared to CLR with cubist (ρc of 0.49, 0.84, and 0.87 for the POCmac, POCmic, and MAOC) and cubist with no compositional transformation (ρc of 0.53, 0.85, and 0.88 for the POCmac, POCmic, and MAOC). The SHAP values reflected the compositional modelling and identified important organic and inorganic functional groups that differed by fraction and land use. Our approach can complement conventional physical SOC fractionations and improve the cost-effectiveness of the measurements, especially when there are many samples to measure, thus enhancing our understanding of SOC dynamics.http://www.sciencedirect.com/science/article/pii/S001670612500045XCarbon fractionsParticulate organic carbonMineral-associated organic carbonDeep learningLogratio transformationSoil spectroscopic modelling
spellingShingle Mingxi Zhang
Zefang Shen
Lewis Walden
Farid Sepanta
Zhongkui Luo
Lei Gao
Oscar Serrano
Raphael A. Viscarra Rossel
Deep learning of the particulate and mineral-associated organic carbon fractions using a compositional transform and mid-infrared spectroscopy
Geoderma
Carbon fractions
Particulate organic carbon
Mineral-associated organic carbon
Deep learning
Logratio transformation
Soil spectroscopic modelling
title Deep learning of the particulate and mineral-associated organic carbon fractions using a compositional transform and mid-infrared spectroscopy
title_full Deep learning of the particulate and mineral-associated organic carbon fractions using a compositional transform and mid-infrared spectroscopy
title_fullStr Deep learning of the particulate and mineral-associated organic carbon fractions using a compositional transform and mid-infrared spectroscopy
title_full_unstemmed Deep learning of the particulate and mineral-associated organic carbon fractions using a compositional transform and mid-infrared spectroscopy
title_short Deep learning of the particulate and mineral-associated organic carbon fractions using a compositional transform and mid-infrared spectroscopy
title_sort deep learning of the particulate and mineral associated organic carbon fractions using a compositional transform and mid infrared spectroscopy
topic Carbon fractions
Particulate organic carbon
Mineral-associated organic carbon
Deep learning
Logratio transformation
Soil spectroscopic modelling
url http://www.sciencedirect.com/science/article/pii/S001670612500045X
work_keys_str_mv AT mingxizhang deeplearningoftheparticulateandmineralassociatedorganiccarbonfractionsusingacompositionaltransformandmidinfraredspectroscopy
AT zefangshen deeplearningoftheparticulateandmineralassociatedorganiccarbonfractionsusingacompositionaltransformandmidinfraredspectroscopy
AT lewiswalden deeplearningoftheparticulateandmineralassociatedorganiccarbonfractionsusingacompositionaltransformandmidinfraredspectroscopy
AT faridsepanta deeplearningoftheparticulateandmineralassociatedorganiccarbonfractionsusingacompositionaltransformandmidinfraredspectroscopy
AT zhongkuiluo deeplearningoftheparticulateandmineralassociatedorganiccarbonfractionsusingacompositionaltransformandmidinfraredspectroscopy
AT leigao deeplearningoftheparticulateandmineralassociatedorganiccarbonfractionsusingacompositionaltransformandmidinfraredspectroscopy
AT oscarserrano deeplearningoftheparticulateandmineralassociatedorganiccarbonfractionsusingacompositionaltransformandmidinfraredspectroscopy
AT raphaelaviscarrarossel deeplearningoftheparticulateandmineralassociatedorganiccarbonfractionsusingacompositionaltransformandmidinfraredspectroscopy