Effect of calibration set size on prediction at local scale of soil carbon by Vis-NIR spectroscopy

Abstract Predicting soil properties through visible and near-infrared (Vis–NIR) spectroscopy by a limited number of calibration samples can reduce the cost and time for physic-chemical analyses. This study was aimed to assess the influence of calibration set size on the prediction of total carbon (TC) in the soil by Vis–NIR spectroscopy. In a forested area of 33 ha in southern Italy (Calabria), 216 soil samples were analyzed for TC concentration, and reflectance spectra were measured in the laboratory. The whole data set was randomly split into calibration and validation sets (70% and 30%, respectively). To study the effect of the number of samples on TC prediction, ten calibration subsets of samples between 14 and 144 were selected. Three techniques including principal components regression (PCR), partial least squares regression (PLSR) and support vector machine regression (SVMR) were used to develop 84 calibration models, validated through the same independent data. The models were compared through the coefficient of determination (R 2 ), the root mean square error of prediction (RMSEP) and the ratio of the interquartile distance (RPIQ). Validation results showed that to obtain not significant differences with models based on the full calibration set, 29, 72 and 115 samples were required for PCR, SVMR and PLSR respectively. Although PCR appeared less sensitive than PLSR and SVMR to calibration sample size, SVMR outperformed PLSR and PCR with higher R 2 and RPIQ values and lower RMSEP. To obtain RMSEP not significantly different from the best model achieved in this study, the required minimum number of samples was 72 for SVMR and 130 for PLSR. All PCR model were significantly poorest than the best model.

[1]  A. Mouazen,et al.  Calibration of visible and near infrared spectroscopy for soil analysis at the field scale on three European farms , 2011 .

[2]  H. Ramon,et al.  Comparison among principal component, partial least squares and back propagation neural network analyses for accuracy of measurement of selected soil properties with visible and near infrared spectroscopy , 2010 .

[3]  A. Karnieli,et al.  Mapping of several soil properties using DAIS-7915 hyperspectral scanner data - a case study over clayey soils in Israel , 2002 .

[4]  W. Köppen.  Das geographische System der Klimate , 1936 .

[5]  R. Barnes,et al.  Standard Normal Variate Transformation and De-Trending of Near-Infrared Diffuse Reflectance Spectra , 1989 .

[6]  H. Levene.  Robust tests for equality of variances , 1961 .

[7]  Giorgio Matteucci,et al.  Visible and near infrared spectroscopy for predicting texture in forest soil: an application in southern Italy , 2014 .

[8]  Abdul Mounem Mouazen,et al.  Do we really need large spectral libraries for local scale SOC assessment with NIR spectroscopy , 2016 .

[9]  C. Hurburgh,et al.  Near-Infrared Reflectance Spectroscopy–Principal Components Regression Analyses of Soil Properties , 2001 .

[10]  N. Ziadi,et al.  Opportunities for, and limitations of, near infrared reflectance spectroscopy applications in soil analysis: a review. , 2009 .

[11]  Yiyun Chen,et al.  Estimating Soil Organic Carbon Using VIS/NIR Spectroscopy with SVMR and SPA Methods , 2014, Remote. Sens..

[12]  Abdul Mounem Mouazen,et al.  Influence of the number of samples on prediction error of visible and near infrared spectroscopy of selected soil properties at the farm scale , 2011 .

[13]  U. Schmidhalter,et al.  High resolution topsoil mapping using hyperspectral image and field data in multivariate regression modeling procedures , 2006 .

[14]  David J. Brown.  Using a global VNIR soil-spectral library for local soil characterization and landscape modeling in a 2nd-order Uganda watershed , 2007 .

[15]  L. Hoffmann,et al.  Measuring soil organic carbon in croplands at regional scale using airborne imaging spectroscopy , 2010 .

[16]  R. V. Rossel,et al.  Using data mining to model and interpret soil diffuse reflectance spectra. , 2010 .

[17]  Christian Walter,et al.  Regional predictions of soil organic carbon content from spectral reflectance measurements , 2009 .

[18]  Bernhard Schölkopf,et al.  Learning with kernels , 2001 .

[19]  R. Lal,et al.  Soil Carbon Sequestration Impacts on Global Climate Change and Food Security , 2004, Science.

[20]  G. Matteucci,et al.  Soil carbon stock in relation to soil properties and landscape position in a forest ecosystem of southern Italy (Calabria region) , 2016 .

[21]  Sabine Grunwald,et al.  Comparison of multivariate methods for inferential modeling of soil carbon using visible/near-infrared spectra , 2008 .

[22]  César Guerrero,et al.  Spiking of NIR regional models using samples from target sites: effect of model size on prediction accuracy. , 2010 .

[23]  K. Shepherd,et al.  Development of Reflectance Spectral Libraries for Characterization of Soil Properties , 2002 .

[24]  T. Næs,et al.  Ensemble methods and data augmentation by noise addition applied to the analysis of spectroscopic data , 2005 .

[25]  A. Paglionico,et al.  Stilo Unit and dioritic-kinzigitic unit in Le Serre (Calabria, Italy); geological, petrological, geochronological characters , 1976 .

[26]  Vladimir N. Vapnik,et al.  The Nature of Statistical Learning Theory , 2000, Statistics for Engineering and Information Science.

[27]  M. Conforti,et al.  Geomorphological map of the Crotone Province (Calabria, South Italy) , 2011 .

[28]  B. Kowalski,et al.  Partial least-squares regression: a tutorial , 1986 .

[29]  P. Miller,et al.  Validation requirements for diffuse reflectance soil characterization models with a case study of VNIR soil C prediction in Montana , 2005 .

[30]  C. Russell.  Sample Preparation and Prediction of Soil Organic Matter Properties by Near Infra-red Reflectance Spectroscopy , 2003 .

[31]  J. Niedźwiecki,et al.  Effect of the number of calibration samples on the prediction of several soil properties at the farm-scale , 2014 .

[32]  R. Poppi,et al.  Determination of organic matter in soils using radial basis function networks and near infrared spectroscopy , 2002 .

[33]  R. V. Rossel,et al.  Visible, near infrared, mid infrared or combined diffuse reflectance spectroscopy for simultaneous assessment of various soil properties , 2006 .

[34]  B. Stenberg,et al.  Near‐infrared spectroscopy for within‐field soil characterization: small local calibrations compared with national libraries spiked with local samples , 2010 .

[35]  A. McBratney,et al.  Critical review of chemometric indicators commonly used for assessing the quality of the prediction of soil attributes by NIR spectroscopy , 2010 .

[36]  Annamaria Castrignanò,et al.  Laboratory-based Vis–NIR spectroscopy and partial least square regression with spatially correlated errors for predicting spatial variation of soil organic matter content , 2015 .

[37]  Thorsten Behrens,et al.  Sampling optimal calibration sets in soil infrared spectroscopy , 2014 .

[38]  Tom Fearn,et al.  Comparing Standard Deviations , 1996 .

[39]  Grzegorz Siebielec,et al.  Near- and mid-infrared diffuse reflectance spectroscopy for measuring soil metal content. , 2004, Journal of environmental quality.

[40]  Budiman Minasny,et al.  A conditioned Latin hypercube method for sampling in the presence of ancillary information , 2006, Comput. Geosci..