scholarly journals Estimating Soil Organic Carbon in Agricultural Gypsiferous Soils by Diffuse Reflectance Spectroscopy

Water ◽  
2020 ◽  
Vol 12 (1) ◽  
pp. 261 ◽  
Author(s):  
Maria Marques ◽  
Ana Álvarez ◽  
Pilar Carral ◽  
Iris Esparza ◽  
Blanca Sastre ◽  
...  
Keyword(s):  
Organic Carbon ◽  
Soil Organic Carbon ◽  
Near Infrared ◽  
Diffuse Reflectance Spectroscopy ◽  
Partial Least Square ◽  
Least Square ◽  
Partial Least Square Regression ◽  
Prediction Ability ◽  
Greenhouse Gas Inventories ◽  
Starting Point

Contents of soil organic carbon (SOC), gypsum, CaCO3, and quartz, among others, were analyzed and related to reflectance features in visible and near-infrared (VIS/NIR) range, using partial least square regression (PLSR) in ParLes software. Soil samples come from a sloping olive grove managed by frequent tillage in a gypsiferous area of Central Spain. Samples were collected in three different layers, at 0–10, 10–20 and 20–30 cm depth (IPCC guidelines for Greenhouse Gas Inventories Programme in 2006). Analyses were performed by C Loss-On-Ignition, X-ray diffraction and water content by the Richards plates method. Significant differences for SOC, gypsum, and CaCO3 were found between layers; similarly, soil reflectance for 30 cm depth layers was higher. The resulting PLSR models (60 samples for calibration and 30 independent samples for validation) yielded good predictions for SOC (R2 = 0.74), moderate prediction ability for gypsum and were not accurate for the rest of rest of soil components. Importantly, SOC content was related to water available capacity. Soils with high reflectance features held c.a. 40% less water than soils with less reflectance. Therefore, higher reflectance can be related to degradation in gypsiferous soil. The starting point of soil degradation and further evolution could be established and mapped through remote sensing techniques for policy decision making.

The interplay among analytical method, preprocessing, and modeling on soil organic carbon Vis-NIR-SWIR predictions

2021 ◽  
Author(s):  
Taciara Zborowski Horst-Heinen ◽  
Ricardo Simão Diniz Dalmolin ◽  
Alessandro Samuel-Rosa ◽  
Sabine Grunwald
Keyword(s):  
Organic Carbon ◽  
Soil Organic Carbon ◽  
Predictive Models ◽  
Analytical Method ◽  
Analytical Methods ◽  
Near Infrared ◽  
Partial Least Square ◽  
Least Square ◽  
Partial Least Square Regression ◽  
Spectral Models

<p>The relationship between visible-near-infrared (Vis-NIR-SWIR) spectra and soil organic carbon (SOC) and the effects of preprocessing techniques on SOC predictive models have been shown in several studies. However, little attention has been given to the effect of analytical methods used to produce the SOC data used to calibrate those models. The predictive performance of Vis-NIR spectral models depends not only on the preprocessing technique and machine learning method but also on the analytical method employed to produce the SOC data. Our hypothesis is that some combinations of preprocessing and models may be more sensitive to laboratory (measurement) error than others. To test this hypothesis, we evaluated the leave-one-out cross-validation performance of three predictive models (Random Forest (RF), Cubist, and Partial Least Square Regression (PLSR)) calibrated using SOC data produced via three analytical methods (dry combustion (DC) and wet combustion with quantification by titration (WCt) and colorimetry (WCc)) and three Vis-NIR spectra preprocessing techniques (smoothing (SMO), continuum removal (CRR), and Savitzky-Golay first derivative (SGD)). The prediction performance varied among the models. DC and WCt provided a higher correlation between SOC and spectra than WCc, and thus, resulted in higher accuracy. The Cubist+CRR was ranked the best performing model, with an average of R2 = 0.81 and RMSE = 0.81% among analytical methods. Cubist+CRR also minimized the accuracy differences resulting from SOC analytical methods employed. The RF model had low accuracy and was unable to explain more than 46% of the variance. Overall, the analytical method significantly affects SOC predictions, and its impact may be larger than the preprocessing. </p>

scholarly journals Error propagation in spectrometric functions of soil organic carbon

2019 ◽  
Author(s):  
Monja Ellinger ◽  
Ines Merbach ◽  
Ulrike Werban ◽  
Mareike Ließ
Keyword(s):  
Organic Carbon ◽  
Soil Organic Carbon ◽  
Near Infrared ◽  
Field Experiments ◽  
Model Building ◽  
Partial Least Square ◽  
Least Square ◽  
Partial Least Square Regression ◽  
Error Measure ◽  
Validation Procedure

Abstract. Soil organic carbon (SOC) plays a major role concerning the chemical, physical and biological soil properties and functions. To get a better understanding how soil management affects the SOC content, the exact monitoring of SOC on long-term field experiments (LTFE) is needed. Visible and near infrared (Vis-NIR) reflectance spectrometry is an inexpensive and fast possibility to enhance conventional SOC analysis and has often been used to predict SOC. For this study, 100 soil samples were collected at a LTFE in central Germany by two different sampling designs. Regression models were built using partial least square regression (PLSR). In order to build robust models, 10-fold cross-validation was used for model tuning and validation procedure. We analysed and discussed various aspects that influence the obtained error measure. A transparent and precise documentation of the model building and validation process, including the calculation of the error measure, is necessary in order to assess the model accuracy in a comprehensive way. This would be the first step to gain a standardized method for model building and validation procedure.

scholarly journals Prediction of Soil Organic Carbon for Ethiopian Highlands Using Soil Spectroscopy

2013 ◽  
Vol 2013 ◽  
pp. 1-11 ◽  
Author(s):  
Tadele Amare ◽  
Christian Hergarten ◽  
Hans Hurni ◽  
Bettina Wolfgramm ◽  
Birru Yitaferu ◽  
...  
Keyword(s):  
Organic Carbon ◽  
Soil Organic Carbon ◽  
Diffuse Reflectance ◽  
Principal Component ◽  
Partial Least Square ◽  
Least Square ◽  
Partial Least Square Regression ◽  
Coefficient Of Determination ◽  
Calibration Models ◽  
The Stability

Soil spectroscopy was applied for predicting soil organic carbon (SOC) in the highlands of Ethiopia. Soil samples were acquired from Ethiopia’s National Soil Testing Centre and direct field sampling. The reflectance of samples was measured using a FieldSpec 3 diffuse reflectance spectrometer. Outliers and sample relation were evaluated using principal component analysis (PCA) and models were developed through partial least square regression (PLSR). For nine watersheds sampled, 20% of the samples were set aside to test prediction and 80% were used to develop calibration models. Depending on the number of samples per watershed, cross validation or independent validation were used. The stability of models was evaluated using coefficient of determination (R2), root mean square error (RMSE), and the ratio performance deviation (RPD). The R2 (%), RMSE (%), and RPD, respectively, for validation were Anjeni (88, 0.44, 3.05), Bale (86, 0.52, 2.7), Basketo (89, 0.57, 3.0), Benishangul (91, 0.30, 3.4), Kersa (82, 0.44, 2.4), Kola tembien (75, 0.44, 1.9), Maybar (84. 0.57, 2.5), Megech (85, 0.15, 2.6), and Wondo Genet (86, 0.52, 2.7) indicating that the models were stable. Models performed better for areas with high SOC values than areas with lower SOC values. Overall, soil spectroscopy performance ranged from very good to good.

scholarly journals Predicting Soil Organic Carbon Content Using Hyperspectral Remote Sensing in a Degraded Mountain Landscape in Lesotho

2020 ◽  
Vol 2020 ◽  
pp. 1-11
Author(s):  
Freddy Bangelesa ◽  
Elhadi Adam ◽  
Jasper Knight ◽  
Inos Dhau ◽  
Marubini Ramudzuli ◽  
...  
Keyword(s):  
Random Forest ◽  
Organic Carbon ◽  
Soil Organic Carbon ◽  
Partial Least Square ◽  
Least Square ◽  
Partial Least Square Regression ◽  
Organic Carbon Content ◽  
Random Forest Regression ◽  
Least Square Regression ◽  
Field Spectroscopy

Soil organic carbon constitutes an important indicator of soil fertility. The purpose of this study was to predict soil organic carbon content in the mountainous terrain of eastern Lesotho, southern Africa, which is an area of high endemic biodiversity as well as an area extensively used for small-scale agriculture. An integrated field and laboratory approach was undertaken, through measurements of reflectance spectra of soil using an Analytical Spectral Device (ASD) FieldSpec® 4 optical sensor. Soil spectra were collected on the land surface under field conditions and then on soil in the laboratory, in order to assess the accuracy of field spectroscopy-based models. The predictive performance of two different statistical models (random forest and partial least square regression) was compared. Results show that random forest regression can most accurately predict the soil organic carbon contents on an independent dataset using the field spectroscopy data. In contrast, the partial least square regression model overfits the calibration dataset. Important wavelengths to predict soil organic contents were localised around the visible range (400–700 nm). This study shows that soil organic carbon can be most accurately estimated using derivative field spectroscopy measurements and random forest regression.

scholarly journals Prediction of Soil Texture Distributions by using PLSR and Reflectance Spectroscopy

2019 ◽  
Vol 8 (3) ◽  
pp. 7876-7881
Keyword(s):  
Statistical Analysis ◽  
Soil Texture ◽  
Near Infrared ◽  
Diffuse Reflectance Spectroscopy ◽  
Principal Component ◽  
Reflectance Spectroscopy ◽  
Soil Samples ◽  
Partial Least Square ◽  
Least Square ◽  
Partial Least Square Regression

The texture of soil i.e. Sand, Silt and Clay are the most important physical properties of soil for agricultural management. In the agricultural practices to increase the productivity of soil, moisture-holding capacity, aeration and to support the agronomic decisions the knowledge of soil texture is an essential task. For this purpose, the present research gives better results and fast acquisition of soil information with the use of Visible and Near Infrared (Vis- NIR) Diffuse Reflectance Spectroscopy. A total of 30 soil samples from two different locations from Aurangabad, Maharashtra, India were collected and analyzed for soil texture. To detect the soil texture the Vis-NIR DRS has shown levels of accurate results compared to the traditional laboratory method with less time, cost and effort. To measure the reflectance of soil the ASD FieldSpec4 Spectroradiometer (350-2500nm) was used. By the observation of captured spectra by using Spectroradiometer it showed that on the basis of different textural classes the soil samples could be spectrally separable. For database collection and pre-processing, we have used RS3 and ViewSpec Pro software respectively. The statistical analysis by using the combination of Principal Component Analysis (PCA) and Partial Least Square Regression method gives accurate results. To determine the texture of soil sample thirteen features were calculated. The main goal of this research was to determine the soil texture by using statistical methods and to test the performance of VNIR-SWIR reflectance spectroscopy by using the ASD FieldSpec4 Spectroradiometer for estimation of the texture of the soil. The results showed that R2 = 0.99 gives maximum accuracy for clay content and R2 = 0.988 for silt content and R2 = 0.989 for sand. The Root Mean Square Values (RMSE) for clay, silt, and sand are 0.02392, 0.02399 and 0.02289 respectively. With the use of reflectance spectroscopy and statistical analysis by using regression models we can determine the soil properties accurately in very less time.

scholarly journals Error propagation in spectrometric functions of soil organic carbon

SOIL ◽  
2019 ◽  
Vol 5 (2) ◽  
pp. 275-288
Author(s):  
Monja Ellinger ◽  
Ines Merbach ◽  
Ulrike Werban ◽  
Mareike Ließ
Keyword(s):  
Organic Carbon ◽  
Soil Organic Carbon ◽  
Error Propagation ◽  
Field Experiments ◽  
Model Building ◽  
Model Performance ◽  
Partial Least Square ◽  
Least Square ◽  
Partial Least Square Regression ◽  
Measurement Protocol

Abstract. Soil organic carbon (SOC) plays a major role concerning chemical, physical, and biological soil properties and functions. To get a better understanding of how soil management affects the SOC content, the precise monitoring of SOC on long-term field experiments (LTFEs) is needed. Visible and near-infrared (Vis–NIR) reflectance spectrometry provides an inexpensive and fast opportunity to complement conventional SOC analysis and has often been used to predict SOC. For this study, 100 soil samples were collected at an LTFE in central Germany by two different sampling designs. SOC values ranged between 1.5 % and 2.9 %. Regression models were built using partial least square regression (PLSR). In order to build robust models, a nested repeated 5-fold group cross-validation (CV) approach was used, which comprised model tuning and evaluation. Various aspects that influence the obtained error measure were analysed and discussed. Four pre-processing methods were compared in order to extract information regarding SOC from the spectra. Finally, the best model performance which did not consider error propagation corresponded to a mean RMSEMV of 0.12 % SOC (R2=0.86). This model performance was impaired by ΔRMSEMV=0.04 % SOC while considering input data uncertainties (ΔR2=0.09), and by ΔRMSEMV=0.12 % SOC (ΔR2=0.17) considering an inappropriate pre-processing. The effect of the sampling design amounted to a ΔRMSEMV of 0.02 % SOC (ΔR2=0.05). Overall, we emphasize the necessity of transparent and precise documentation of the measurement protocol, the model building, and validation procedure in order to assess model performance in a comprehensive way and allow for a comparison between publications. The consideration of uncertainty propagation is essential when applying Vis–NIR spectrometry for soil monitoring.

Sentinel-2 exposed soil composite for soil organic carbon prediction: the ‘greening-up’ method for detecting suitable images

2021 ◽  
Author(s):  
Klara Dvorakova ◽  
Bas van Wesemael
Keyword(s):  
Remote Sensing ◽  
Surface Roughness ◽  
Soil Moisture ◽  
Organic Carbon ◽  
Soil Organic Carbon ◽  
Satisfactory Result ◽  
Partial Least Square ◽  
Least Square ◽  
Partial Least Square Regression ◽  
Sentinel 2

<p>Pilot studies have demonstrated the potential for remote sensing techniques for soil organic carbon  (SOC) mapping in exposed croplands. However, the use of remote sensing for SOC prediction is often hindered by disturbing factors at the soil surface such as photosynthetic active and nonphotosynthetic active vegetation, variation in soil moisture or surface roughness. With the increasing amount of freely available satellite data, many studies have focused on stabilizing the soil reflectance by building image composites that are generated using a set of criteria. These composites tend to minimize and cancel out the disturbing effects. Here we aim to develop a robust method that allows selecting Sentinel-2 (S-2) pixels that are not affected by the following disturbing factors: crop residues, surface roughness and soil moisture. We selected all S-2 cloud-free images covering the Loam Belt of Belgium from January 2019 to December 2020 (in total 38 images). We then built four exposed soil composites based on four sets of criteria: (1) NDVI < 0.25, (2) NDVI < 0.25 & Normalized Burn Ratio (NBR2) < 0.07, (3) the ‘greening-up’ period of a crop and (4) the ‘greening-up’ period of a crop & NBR2 < 0.07. The ‘greening-up’ period was selected based on the NDVI timeline, where ‘greening-up’ is considered as the last date of acquisition where the soil is bare (NDVI < 0.25) before the crop develops (NDVI > 0.6).,We then built a partial least square regression (PLSR) model with 10-fold cross-validation to estimate the SOC content based on 137 georeferenced calibration samples on the four above described composites. We obtained a non-satisfactory result for composites (1) to (3): R² = 0.22, RMSE = 3.46 g C kg<sup>-1</sup> and RPD 1.12 for (1), R² = 0.19, RMSE = 3.43 g C kg<sup>-1</sup> and RPD 1.10 for (2) and R² = 0.15, RMSE = 2.74 g C kg<sup>-1</sup> and RPD 1.06 for (3). We, however, obtained a satisfactory result for composite (4): R² = 0.54, RMSE = 2.09 g C kg<sup>-1</sup> and RPD 1.68. Hence, the ‘greening-up’ method combined with a strict NBR2 threshold allows selecting the purest exposed soil pixels suitable for SOC prediction. The limit of this method might be the surface coverage, which for a two-year period reached 47% of croplands, compared to 89% exposure if only the NDVI threshold is applied.</p>

A Consensus Partial Least Squares Regression for Analysis of Near-Infrared Spectroscopy

2013 ◽  
Vol 765-767 ◽  
pp. 528-531
Author(s):  
Dan Peng ◽  
Qing Chen Nie
Keyword(s):  
Least Squares ◽  
Partial Least Squares ◽  
Near Infrared ◽  
Partial Least Square ◽  
Least Square ◽  
Partial Least Square Regression ◽  
Least Squares Regression ◽  
Prediction Ability ◽  
Successful Model ◽  
Consensus Strategy

To improve the prediction performance of partial least square regression algorithm (PLS), the consensus strategy was applied to develop the multivariate regression model using near-infrared (NIR) spectra and named as C-PLS. Coupled with the consensus strategy, this algorithm can take the advantage of reducing dependence on single model to obtain prediction precision and stability by randomly changing the calibration set. Through an optimization of the parameters involved in the model including criterion threshold and number of sub-models, a successful model was achieved by effectively combining many sub-models with different accuracy and diversity together. To validate the C-PLS algorithm, it was applied to measure the original extract concentration of beer using NIR spectra. The experimental results showed that the prediction ability and robustness of model obtained in subsequent partial least squares calibration using consensus strategy was superior to that obtained using conventional PLS algorithm, and the root mean square error of prediction can improve by up to 45.2%, indicating that it is an efficient tool for NIR spectra regression.

Sign in / Sign up

Export Citation Format

Share Document