Predicting soil organic carbon and total nitrogen using mid- and near-infrared spectra for Brookston clay loam soil in Southwestern Ontario, Canada

Abstract

Xie, H. T., Yang, X. M., Drury, C. F., Yang, J. Y. and Zhang, X. D. 2011. Predicting soil organic carbon and total nitrogen using mid- and near-infrared spectra for Brookston clay loam soil in Southwestern Ontario, Canada. Can. J. Soil Sci. 91: 53–63. Mid-infrared (MIR) and near-infrared (NIR) spectroscopy of soils have been tested to estimate soil organic carbon (SOC) and total N (TN) concentrations at local, regional and national scales. However, these methods have rarely been used to assess SOC and TN concentrations of the same soil under different management practices. The objective of this study was to determine if models developed from infrared spectra of Brookston clay loam soils under different management practices could be used to estimate SOC, and TN concentrations and the C:N ratio. Soils used for model calibration included 217 samples from a long-term fertilization and crop rotation study and a long-term compost study, whereas 78 soil samples from a long-term tillage study on the same soil type were used for model validation. Soil organic carbon and TN concentrations of all samples were also analyzed using dry combustion techniques. Soil samples were scanned from 4000 to 400 cm−1 (2500–25 000 nm) for MIR spectra and from 8000 to 4000 cm−1 (1250–2500 nm) for NIR spectra. Partial least squares regression (PLSR) analysis was used for the calibration dataset to build prediction models for SOC, TN and C:N ratio. The SOC and TN concentrations determined using dry combustion techniques were compared with the prediction from the models using the calibration datasets. The predictions of SOC and TN concentrations by the PLSR method using infrared spectra were statistically sound, with high coefficient of determination with the calibration dataset (R2cal, SOCMIR=0.99 and SOCNIR=0.97, TNMIR=0.98 and TNNIR=0.97) and the validation dataset (R2val, SOCMIR=0.96 and SOCNIR=0.95, TNMIR=0.96 and TNNIR=0.95) and low root mean square error (RMSEPcal, SOCMIR=0.93 and SOCNIR=1.60, TNMIR=0.08 and TNNIR=0.12; RMSEPval, SOCMIR=1.40 and SOCNIR=1.75, TNMIR=0.11 and TNNIR=0.12). The predictions of SOC and TN concentrations in the 5 to 30 cm depth were better than the predictions for either the surface (0 to 5 cm) soils or for soils from lower depths (>30 cm). The models could be used as an alternative method for determining SOC and TN concentrations of Brookston clay loam soils; however, larger sample populations and improved model algorithms could further improve predictions.

Application of visible and near infrared reflectance spectroscopy to predict total nitrogen in soil

“…As can be seen in Table 3, 2 R regression value was found as 0,793 and residual prediction 2 deviation value as 2.0802. R and the RPD values calculated in this study for the prediction of total nitrogen were lower than that of previous studies (Xie et al, 2011;Wenjun et al, 2014). Many researchers have reportedsuccessful application of thevis-NIR spectroscopy for the prediction of various soil characteristics.…”

Monitoring organic carbon, total nitrogen, and pH for reclaimed soils using field reflectance spectroscopy

“…The results of the Cubist models for predicting SOC and TN under field conditions were satisfactory and compared favourably with results from studies using dried and homogenized soil samples in the laboratory (Chang et al 2001;Martin et al 2002;Xie et al 2011). Our results compare favourably with a study in Australia that used a portable spectrometer to measure SOC contents in the field, as we were able to obtain a RPD value of 2.2 compared with 1.92 in the study in Australia (Gomez et al 2008b).…”

“…While work conducted in the Canadian provinces of Manitoba and Ontario has successfully shown that reflectance spectroscopy can be used to predict SOC and TN (Martin et al 2002;Xie et al 2011), there is a lack of information on the use of reflectance spectroscopy on Canadian soils and more specifically, Alberta soils. The processing of soil spectral data has focused on using partial least squares regressions (e.g., Gomez et al 2008aGomez et al , 2008bXie et al 2011;Kinoshita et al 2012) and using signal processing techniques such as Savitzky-Golay smoothing, derivatives, and multiplicative scatter correction (e.g., He et al 2007;Minasny et al 2009;Doetterl et al 2013;Nawar et al 2016). Recently, two data processing techniques have been applied to spectral data and have improved the accuracy of model predictions.…”

Rapid determination of soil organic matter quality indicators using visible near infrared reflectance spectroscopy

“…Luce et al, 2012;Xie et al, 2011;Zornoza et al, 2008). The VNIRS was also used for prediction of MBN (Coûteaux et al, 2003;Ludwig et al, 2002;Reeves and McCarty, 2001;Terhoeven-Urselmans et al, 2008).…”