Elsevier

Pedosphere

Volume 28, Issue 6, December 2018, Pages 856-872
Pedosphere

Transferability and Scalability of Soil Total Carbon Prediction Models in Florida, USA

https://doi.org/10.1016/S1002-0160(18)60048-7Get rights and content

Abstract

The applicability, transferability, and scalability of visible/near-infrared (VNIR)-derived soil total carbon (TC) models are still poorly understood. The objectives of this study were to: i) compare models of three multivariate statistical methods, partial least squares regression (PLSR), support vector machine (SVM), and random forest methods, to predict soil logarithm-transformed TC (logTC) using five fields (local scale) and a pooled (regional-scale) VNIR spectral dataset (a total of 560 TC spectral datasets), ii) assess the model transferability among fields, and iii) evaluate their up- and downscaling behaviors in Florida, USA. The transferability and up- and downscaling of the models were limited by the following factors: i) the spectral data domain, ii) soil attribute domain, iii) methods that describe the internal model structure of VNIR-TC relationships, and iv) environmental domain space of attributes that control soil carbon dynamics. All soil logTC models showed excellent performance based on all three methods with R2 > 0.86, bias < 0.01%, root mean squared error (RMSE) = 0.09%, residual predication deviation (RPD) > 2.70%, and ratio of prediction error to interquartile range (RPIQ) > 4.54. The PLSR method performed substantially better than the SVM method to scale and transfer the TC models. This could be attributed to the tendency of SVM to overfit models, while the asset of the PLSR method was its robustness when the models were validated with independent datasets, transferred, and/or scaled. The upscaled soil TC models performed somewhat better in terms of model fit (R2), RPD, and RPIQ, whereas the downscaled models showed less bias and smaller RMSE based on PLSR. We found no universal trend indicating which of the four limiting factors mentioned above had the most impact that constrained the transferability and scalability of the models. Given that several factors can impinge on the empirically derived soil spectral prediction models, as demonstrated by this study, more focus on their applicability and scalability is needed.

References (63)

  • M L McDowell et al.

    Soil total carbon analysis in Hawaiian soils with visible, near-infrared and mid-infrared diffuse reflectance spectroscopy

    Geoderma

    (2012)
  • B Minasny et al.

    Incorporating taxonomic distance into spatial prediction and digital mapping of soil classes

    Geoderma

    (2007)
  • B Minasny et al.

    Regression rules as a tool for predicting soil properties from infrared reflectance spectroscopy

    Chemometr Intell Lab Syst

    (2008)
  • B Minasny et al.

    Regional transferability of mid-infrared diffuse reflectance spectroscopic prediction for soil chemical properties

    Geoderma

    (2009)
  • A Morellos et al.

    Machine learning based prediction of soil total nitrogen, organic carbon and moisture content by using VIS-NIR spectroscopy

    Biosyst Eng

    (2016)
  • M Nocita et al.

    Predictions of soil surface and topsoil organic carbon content through the use of laboratory and field spectroscopy in the Albany Thicket Biome of Eastern Cape Province of South Africa

    Geoderma

    (2011)
  • R M Rees et al.

    The role of plants and land management in sequestering soil carbon in temperate arable and grassland ecosystems

    Geoderma

    (2005)
  • J Reeves et al.

    The potential of diffuse reflectance spectroscopy for the determination of carbon inventories in soils

    Environ Pollut

    (2002)
  • J B Sankey et al.

    Comparing local vs. global visible and near-infrared (VisNIR) diffuse reflectance spectroscopy (DRS) calibrations for the prediction of soil clay, organic C and inorganic C

    Geoderma

    (2008)
  • D V Sarkhot et al.

    Comparison and detection of total and available soil carbon fractions using visible/near infrared diffuse reflectance spectroscopy

    Geoderma

    (2011)
  • U Thissen et al.

    Comparing support vector machines to PLS for spectral regression applications

    Chemometr Intell Lab Syst

    (2004)
  • G M Vasques et al.

    Comparison of multivariate methods for inferential modeling of soil carbon using visible/near-infrared spectra

    Geoderma

    (2008)
  • R A Viscarra Rossel et al.

    Visible, near infrared, mid infrared or combined diffuse reflectance spectroscopy for simultaneous assessment of various soil properties

    Geoderma

    (2006)
  • R A Viscarra Rossel et al.

    A global spectral library to characterize the world's soil

    Earth Sci Rev

    (2016)
  • X Xiong et al.

    Scale-dependent variability of soil organic carbon coupled to land use and land cover

    Soil Till Res

    (2016)
  • F Yimer et al.

    Soil organic carbon and total nitrogen stocks as affected by topographic aspect and vegetation in the Bale Mountains, Ethiopia

    Geoderma

    (2006)
  • T Addiscott et al.

    Critical evaluation of models and their parameters

    J Environ Qual

    (1995)
  • G Blöschl et al.

    Scale issues in hydrological modelling: A review

    Hydrol Process

    (1995)
  • L Breiman

    Random forests

    Mach Learn

    (2001)
  • V P Chaudhary et al.

    Reflectance spectroscopy detects management and landscape differences in soil carbon and nitrogen

    Soil Sci Soc Am J

    (2012)
  • R N Clark

    Spectroscopy of rocks and minerals, and principles of spectroscopy

  • Cited by (0)

    View full text