Details
| Original language | English |
|---|---|
| Article number | 116724 |
| Number of pages | 14 |
| Journal | GEODERMA |
| Volume | 440 |
| Early online date | 1 Dec 2023 |
| Publication status | Published - Dec 2023 |
Abstract
Diffuse reflectance spectroscopy has been extensively employed to deliver timely and cost-effective predictions of a number of soil properties. However, although several soil spectral laboratories have been established worldwide, the distinct characteristics of instruments and operations still hamper further integration and interoperability across mid-infrared (MIR) soil spectral libraries. In this study, we conducted a large-scale ring trial experiment to understand the lab-to-lab variability of multiple MIR instruments. By developing a systematic evaluation of different mathematical treatments with modeling algorithms, including regular preprocessing and spectral standardization, we quantified and evaluated instruments' dissimilarity and how this impacts internal and shared model performance. We found that all instruments delivered good predictions when calibrated internally using the same instruments' characteristics and standard operating procedures by solely relying on regular spectral preprocessing that accounts for light scattering and multiplicative/additive effects, e.g., using standard normal variate (SNV). When performing model transfer from a large public library (the USDA NSSC-KSSL MIR library) to secondary instruments, good performance was also achieved by regular preprocessing (e.g., SNV) if both instruments shared the same manufacturer. However, significant differences between the KSSL MIR library and contrasting ring trial instruments responses were evident and confirmed by a semi-unsupervised spectral clustering. For heavily contrasting setups, spectral standardization was necessary before transferring prediction models. Non-linear model types like Cubist and memory-based learning delivered more precise estimates because they seemed to be less sensitive to spectral variations than global partial least square regression. In summary, the results from this study can assist new laboratories in building spectroscopy capacity utilizing existing MIR spectral libraries and support the recent global efforts to make soil spectroscopy universally accessible with centralized or shared operating procedures.
Keywords
- Calibration transfer, Chemometrics, Ring trial, Soil spectroscopy, Spectral standardization
ASJC Scopus subject areas
- Agricultural and Biological Sciences(all)
- Soil Science
Cite this
- Standard
- Harvard
- Apa
- Vancouver
- BibTeX
- RIS
In: GEODERMA, Vol. 440, 116724, 12.2023.
Research output: Contribution to journal › Article › Research › peer review
}
TY - JOUR
T1 - An interlaboratory comparison of mid-infrared spectra acquisition
T2 - Instruments and procedures matter
AU - Safanelli, José L.
AU - Sanderman, Jonathan
AU - Bloom, Dellena
AU - Todd-Brown, Katherine
AU - Parente, Leandro L.
AU - Hengl, Tomislav
AU - Adam, Sean
AU - Albinet, Franck
AU - Ben-Dor, Eyal
AU - Boot, Claudia M.
AU - Bridson, James H.
AU - Chabrillat, Sabine
AU - Deiss, Leonardo
AU - Demattê, José A.M.
AU - Scott Demyan, M.
AU - Dercon, Gerd
AU - Doetterl, Sebastian
AU - van Egmond, Fenny
AU - Ferguson, Rich
AU - Garrett, Loretta G.
AU - Haddix, Michelle L.
AU - Haefele, Stephan M.
AU - Heiling, Maria
AU - Hernandez-Allica, Javier
AU - Huang, Jingyi
AU - Jastrow, Julie D.
AU - Karyotis, Konstantinos
AU - Machmuller, Megan B.
AU - Khesuoe, Malefetsane
AU - Margenot, Andrew
AU - Matamala, Roser
AU - Miesel, Jessica R.
AU - Mouazen, Abdul M.
AU - Nagel, Penelope
AU - Patel, Sunita
AU - Qaswar, Muhammad
AU - Ramakhanna, Selebalo
AU - Resch, Christian
AU - Robertson, Jean
AU - Roudier, Pierre
AU - Sabetizade, Marmar
AU - Shabtai, Itamar
AU - Sherif, Faisal
AU - Sinha, Nishant
AU - Six, Johan
AU - Summerauer, Laura
AU - Thomas, Cathy L.
AU - Toloza, Arsenio
AU - Tomczyk-Wójtowicz, Beata
AU - Tsakiridis, Nikolaos L.
AU - van Wesemael, Bas
AU - Woodings, Finnleigh
AU - Zalidis, George C.
AU - Żelazny, Wiktor R.
N1 - Funding Information: Funding for this work came from USDA NIFA Award 2020-67021-32467 (Soil Spectroscopy for the Global Good). The authors would like to thank the staff at the NSSC-KSSL for facilitating this research and Matthew Jacques for preparing and shipping samples to all participants. We would like to thank Dr. Cathy Seybold and Dr. Jonathan Maynard for taking the time and effort necessary to review the manuscript as part of the USDA NRCS internal review, which helped us to improve the quality of the manuscript. In addition, C.M.B. received funding from the Grantham Foundation and support from the Analytical Resources Core RRID: SCR_021758, and thanks Olivia Hill for assistance with data collection; J.A.M.D. was supported by the Department of Soil Science, CNPq and FAPESP foundation 2014-22260-0, 2021-05129-8; L.G.G. J.H.B, and S.P. were supported by the Tree-Root-Microbiome program, which is funded by the New Zealand Ministry of Business, Innovation & Employment (MBIE) Endeavour Fund and in part by the New Zealand Forest Growers Levy Trust (C04X2002); M.H. was supported by USDA NRCS #NR193A750025C005; S.M.H. C.L.T. and J.H.A. were supported from the Growing Health Institute Strategic Programme [BB/X010953/1], funded by the Biotechnology and Biological Sciences Research Council of the United Kingdom (BBSRC); J.H. was supported by NSF Signals in the Soil Grant 2226568; J.D.J and R.M. were supported by the U.S. Department of Energy, Office of Science, Office of Biological and Environmental Research, Environmental System Science Program, under contract DE-AC02-06CH11357; A.M. was funded by USDA NIFA #2020-67021-32799, project accession 1024178, and USDA NIFA #2021-68012-35896, project accession 1027512; A.H.J.R was funded by the Rural & Environment Science & Analytical Services Division of the Scottish Government; P.R. was supported by the Strategic Science Investment Funding for Crown Research Institutes from the New Zealand Ministry of Business, Innovation and Employment's Science and Innovation Group; J.S. S.D. and L.S. received core funding provided by ETH Zurich to purchase and maintain MIR instrumentation; V.R.Ż. was supported by the Ministry of Agriculture of the Czech Republic, institutional support MZE-RO0423. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Trade names are solely used to provide information. Mention of trade names does not constitute a guarantee of the product nor does it imply endorsement over comparable products that are named or not named.
PY - 2023/12
Y1 - 2023/12
N2 - Diffuse reflectance spectroscopy has been extensively employed to deliver timely and cost-effective predictions of a number of soil properties. However, although several soil spectral laboratories have been established worldwide, the distinct characteristics of instruments and operations still hamper further integration and interoperability across mid-infrared (MIR) soil spectral libraries. In this study, we conducted a large-scale ring trial experiment to understand the lab-to-lab variability of multiple MIR instruments. By developing a systematic evaluation of different mathematical treatments with modeling algorithms, including regular preprocessing and spectral standardization, we quantified and evaluated instruments' dissimilarity and how this impacts internal and shared model performance. We found that all instruments delivered good predictions when calibrated internally using the same instruments' characteristics and standard operating procedures by solely relying on regular spectral preprocessing that accounts for light scattering and multiplicative/additive effects, e.g., using standard normal variate (SNV). When performing model transfer from a large public library (the USDA NSSC-KSSL MIR library) to secondary instruments, good performance was also achieved by regular preprocessing (e.g., SNV) if both instruments shared the same manufacturer. However, significant differences between the KSSL MIR library and contrasting ring trial instruments responses were evident and confirmed by a semi-unsupervised spectral clustering. For heavily contrasting setups, spectral standardization was necessary before transferring prediction models. Non-linear model types like Cubist and memory-based learning delivered more precise estimates because they seemed to be less sensitive to spectral variations than global partial least square regression. In summary, the results from this study can assist new laboratories in building spectroscopy capacity utilizing existing MIR spectral libraries and support the recent global efforts to make soil spectroscopy universally accessible with centralized or shared operating procedures.
AB - Diffuse reflectance spectroscopy has been extensively employed to deliver timely and cost-effective predictions of a number of soil properties. However, although several soil spectral laboratories have been established worldwide, the distinct characteristics of instruments and operations still hamper further integration and interoperability across mid-infrared (MIR) soil spectral libraries. In this study, we conducted a large-scale ring trial experiment to understand the lab-to-lab variability of multiple MIR instruments. By developing a systematic evaluation of different mathematical treatments with modeling algorithms, including regular preprocessing and spectral standardization, we quantified and evaluated instruments' dissimilarity and how this impacts internal and shared model performance. We found that all instruments delivered good predictions when calibrated internally using the same instruments' characteristics and standard operating procedures by solely relying on regular spectral preprocessing that accounts for light scattering and multiplicative/additive effects, e.g., using standard normal variate (SNV). When performing model transfer from a large public library (the USDA NSSC-KSSL MIR library) to secondary instruments, good performance was also achieved by regular preprocessing (e.g., SNV) if both instruments shared the same manufacturer. However, significant differences between the KSSL MIR library and contrasting ring trial instruments responses were evident and confirmed by a semi-unsupervised spectral clustering. For heavily contrasting setups, spectral standardization was necessary before transferring prediction models. Non-linear model types like Cubist and memory-based learning delivered more precise estimates because they seemed to be less sensitive to spectral variations than global partial least square regression. In summary, the results from this study can assist new laboratories in building spectroscopy capacity utilizing existing MIR spectral libraries and support the recent global efforts to make soil spectroscopy universally accessible with centralized or shared operating procedures.
KW - Calibration transfer
KW - Chemometrics
KW - Ring trial
KW - Soil spectroscopy
KW - Spectral standardization
UR - http://www.scopus.com/inward/record.url?scp=85178668878&partnerID=8YFLogxK
U2 - 10.1016/j.geoderma.2023.116724
DO - 10.1016/j.geoderma.2023.116724
M3 - Article
AN - SCOPUS:85178668878
VL - 440
JO - GEODERMA
JF - GEODERMA
SN - 0016-7061
M1 - 116724
ER -