Strategies for the efficient estimation of soil organic carbon at the field scale with vis-NIR spectroscopy: Spectral libraries and spiking vs. local calibrations

Seidel, Michael; Hutengs, Christopher; Ludwig, Bernard; Thiele‐Bruhn, Sören; Vohland, Michael

doi:10.1016/j.geoderma.2019.07.014

Cited by 66 publications

(56 citation statements)

References 47 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…In addition, the predictive ability of the CSSL model on TA2 pre was also reduced when 10 spiking samples were added, compared with 5 spiking samples, as shown in Figure 5 e,f (the RPD decreased from 2.70 to 2.64). Seidel et al found a similar result; when the number of spiking samples reached 15–20, the predictive ability of the model declined due to saturation [ 31 ].…”

Section: Resultsmentioning

confidence: 81%

“…Guerrero et al used the KS algorithm and selected 10% of the samples from the target area as the spiking samples according to the principle of uniform distribution of the principal component scores of the spectral data [ 30 ]. Seidel et al used the KS algorithm and selected 12.5% of the samples from the target area as the spiking samples according to the principle of maximum differentiation based on the Mahalanobis distance of spectral data [ 31 ]. The results showed that these spiking samples had greatly improved the prediction accuracy of the model.…”

Section: Resultsmentioning

confidence: 99%

See 1 more Smart Citation

Prediction of Soil Organic Carbon in a New Target Area by Near-Infrared Spectroscopy: Comparison of the Effects of Spiking in Different Scale Soil Spectral Libraries

Jia

2020

Sensors

View full text Add to dashboard Cite

Near-infrared (NIR) spectroscopy is widely used to predict soil organic carbon (SOC) because it is rapid and accurate under proper calibration. However, the prediction accuracy of the calibration model may be greatly reduced if the soil characteristics of some new target areas are different from the existing soil spectral library (SSL), which greatly limits the application potential of the technology. We attempted to solve the problem by building a large-scale SSL or using the spiking method. A total of 983 soil samples were collected from Zhejiang Province, and three SSLs were built according to geographic scope, representing the provincial, municipal, and district scales. The partial least squares (PLS) algorithm was applied to establish the calibration models based on the three SSLs, and the models were used to predict the SOC of two target areas in Zhejiang Province. The results show that the prediction accuracy of each model was relatively poor regardless of the scale of the SSL (residual predictive deviation (RPD) < 2.5). Then, the Kennard-Stone (KS) algorithm was applied to select 5 or 10 spiking samples from each target area. According to different SSLs and numbers of spiking samples, different spiked models were established by the PLS. The results show that the predictive ability of each model was improved by the spiking method, and the improvement effect was inversely proportional to the scale of the SSL. The spiked models built by combining the district scale SSL and a few spiking samples achieved good prediction of the SOC of two target areas (RPD = 2.72 and 3.13). Therefore, it is possible to accurately measure the SOC of new target areas by building a small-scale SSL with a few spiking samples.

show abstract

Section: Resultsmentioning

confidence: 81%

Section: Resultsmentioning

confidence: 99%

Prediction of Soil Organic Carbon in a New Target Area by Near-Infrared Spectroscopy: Comparison of the Effects of Spiking in Different Scale Soil Spectral Libraries

Jia

2020

Sensors

View full text Add to dashboard Cite

show abstract

“…They recommended performing a calibration before each measurement campaign due to the sensitivity of the method to small changes in the field conditions at sampling (e.g., moisture content, roughness, and vegetation). It has also been emphasized that bias should be calculated to evaluate model performance due to the common occurrence of a high correlation between measured and predicted values but consistent over-or underestimation (Bellon-Maurel & McBratney, 2011;Seidel, Hutengs, Ludwig, Thiele-Bruhn, & Vohland, 2019).…”

Section: Core Ideasmentioning

confidence: 99%

Robustness of visible near‐infrared and mid‐infrared spectroscopic models to changes in the quantity and quality of crop residues in soil

Greenberg

Linsler

Vohland

et al. 2020

Soil Science Soc of Amer J

Self Cite

View full text Add to dashboard Cite

The robustness of soil organic carbon (SOC) and total nitrogen (TN) content prediction accuracy by visible near-infrared spectroscopy (visNIRS) and mid-infrared spectroscopy (MIRS) models after a change in the quantity or quality of crop residues requires investigation. Arable soils (0-20 cm) from 20 locations across Germany were collected, and 0, 2, 4, or 8 g C kg soil −1 of wheat straw (C/N ratio, 54) or clover (C/N ratio, 13) were added. Before and after a 56-d incubation, dried and ground samples were measured for SOC and TN content and by visNIRS and MIRS. The complete dataset (n = 280) was subdivided into calibration and validation datasets to test the robustness of partial least squares regression models to changes in crop residue quantity and quality in soil. Noise-reducing data pretreatments included region selection, moving averages, resampling every second data point, and the Savitzky-Golay algorithm. The MIRS estimates for SOC (7.4-33 g kg −1) had lower root mean squared error of validation (RMSE V = 0.9-2.9 g kg −1) compared with visNIRS (RMSE V = 1.6-7.1 g kg −1). Total N estimates (0.7-2.8 g kg −1) were more comparable for MIRS (RMSE V = 0.1-0.3 g kg −1) and visNIRS (RMSE V = 0.1-1.0 g kg −1). Loadings of partial least squares regression components suggested the predictive mechanisms for SOC and TN were more similar for visNIRS than for MIRS. Differing crop residue quantity or quality in calibration versus validation resulted in biased SOC and TN estimates by visNIRS and MIRS models. However, calibration with a global residue model containing all soils and crop residue quantities and qualities lowered RMSE V for SOC and TN prediction with visNIRS and MIRS, demonstrating the usefulness of this approach. Abbreviations: MIR, mid-infrared; MIRS, mid-infrared spectroscopy; RPIQ, ratio of performance to interquartile distance; RPIQ V , ratio of performance to interquartile distance of validation; SOC, soil organic carbon; TN, total nitrogen; PLSR, partial least squares regression; RMSE V , root mean squared error of validation; RPD, ratio of performance to deviation; RPD V , ratio of prediction to deviation of validation; visNIR, visible near-infrared; visNIRS, visible near-infrared spectroscopy. This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.

show abstract

“…The operational value of SSLs lies in the ability to pull representative information (either the actual soil spectra or learnt model "rules") from them, requiring less new local samples and laboratory analysis. These methods can be summarized as augmenting SSLs with local soil samples, or spiking (Shepherd and Walsh, 2002;Brown, 2007;Wetterlind and Stenberg, 2010;Seidel et al, 2019), memory-or instance-based learning (Ramirez-Lopez et al, 2013;Gholizadeh et al, 2016), subsetting (Araújo et al, 2014;Lobsey et al, 2017) or transfer learning (Padarian et al, 2019a), which is the process of sharing intra-domain information and rules learnt by general models to a local domain (Pan and Yang, 2010).…”

Section: Introductionmentioning

confidence: 99%

Comment on soil-2020-93

Sanderman¹

2021

View full text Add to dashboard Cite

Traditional laboratory methods of acquiring soil information remain important for assessing key soil properties, soil functions and ecosystem services over space and time. Infrared spectroscopic modelling can link and massively scale up these methods for many soil characteristics in a cost-effective and timely manner. In Switzerland, only 10 % to 15 % of agricultural soils have been mapped sufficiently to serve spatial decision support systems, presenting an urgent need for rapid quantitative soil characterization. The current Swiss soil spectral library (SSL; n = 4374) in the mid-infrared range includes soil samples from the Biodiversity Monitoring Program (BDM), arranged in a regularly spaced grid across Switzerland, and temporally-resolved data from the Swiss Soil Monitoring Network (NABO). Given the relatively low representation of organic soils and their organo-mineral diversity in the SSL, we aimed to develop both an efficient calibration sampling scheme and accurate modelling strategy to estimate soil carbon (SC) contents of heterogeneous samples between 0 m to 2 m depth from 26 locations within two drained peatland regions (HAFL dataset; n = 116). The focus was on minimizing the need for new reference analyses by efficiently mining the spectral information of SSL instances and their target-feature representations.We used partial least square regressions (PLSR) together with a 5 times repeated, grouped by location, 10-fold cross validation (CV) to predict SC ranging from 1 % to 52 % in the local HAFL dataset. We compared the validation performance of different calibration schemes involving local models (1), models using the entire SSL spiked with local samples (2) and subsets of local and SSL samples using the RS-LOCAL algorithm (3). Using local and RS-LOCAL calibrations with at least 5 local samples, we achieved similar validation results for predictions of SC up to 52 % (R 2 = 0.94 − 0.96, bias = −0.6 − 1.5, RMSE = 2.6 % to 3.5 % total carbon). However, calibrations of representative SSL and local samples using RS-LOCAL only required 5 local samples for very accurate models (RMSE = 2.9 % total carbon), while local calibrations required 50 samples for similarly accurate results (RMSE < 3 % total carbon). Of the three approaches, the entire SSL spiked with local samples for model calibration led to validations with the lowest performance in terms of R 2 , bias and RMSE. Hence, we show that a simple and comprehensible modelling approach using RS-LOCAL together with a SSL is an efficient and accurate strategy when using 1

show abstract

Strategies for the efficient estimation of soil organic carbon at the field scale with vis-NIR spectroscopy: Spectral libraries and spiking vs. local calibrations

Cited by 66 publications

References 47 publications

Prediction of Soil Organic Carbon in a New Target Area by Near-Infrared Spectroscopy: Comparison of the Effects of Spiking in Different Scale Soil Spectral Libraries

Prediction of Soil Organic Carbon in a New Target Area by Near-Infrared Spectroscopy: Comparison of the Effects of Spiking in Different Scale Soil Spectral Libraries

Robustness of visible near‐infrared and mid‐infrared spectroscopic models to changes in the quantity and quality of crop residues in soil

Comment on soil-2020-93

Contact Info

Product

Resources

About