Said Nawar

Modeling and Mapping of Soil Salinity with Reflectance Spectroscopy and Landsat Data Using Two Quantitative Methods (PLSR and MARS)

The monitoring of soil salinity levels is necessary for the prevention and mitigation of land degradation in arid environments. To assess the potential of remote sensing in estimating and mapping soil salinity in the El-Tina Plain, Sinai, Egypt, two predictive models were constructed based on the measured soil electrical conductivity (ECe) and laboratory soil reflectance spectra resampled to Landsat sensor’s resolution. The models used were partial least squares regression (PLSR) and multivariate adaptive regression splines (MARS). The results indicated that a good prediction of the soil salinity can be made based on the MARS model (R2 = 0.73, RMSE = 6.53, and ratio of performance to deviation (RPD) = 1.96), which performed better than the PLSR model (R2 = 0.70, RMSE = 6.95, and RPD = 1.82). The models were subsequently applied on a pixel-by-pixel basis to the reflectance values derived from two Landsat images (2006 and 2012) to generate quantitative maps of the soil salinity. The resulting maps were validated successfully for 37 and 26 sampling points for 2006 and 2012, respectively, with R2 = 0.72 and 0.74 for 2006 and 2012, respectively, for the MARS model, and R2 = 0.71 and 0.73 for 2006 and 2012, respectively, for the PLSR model. The results indicated that MARS is a more suitable technique than PLSR for the estimation and mapping of soil salinity, especially in areas with high levels of salinity. The method developed in this paper can be used for other satellite data, like those provided by Landsat 8, and can be applied in other arid and semi-arid environments.

Digital Mapping of Soil Properties Using Multivariate Statistical Analysis and ASTER Data in an Arid Region

Modeling and mapping of soil properties has been identified as key for effective land degradation management and mitigation. The ability to model and map soil properties at sufficient accuracy for a large agriculture area is demonstrated using Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) imagery. Soil samples were collected in the El-Tina Plain, Sinai, Egypt, concurrently with the acquisition of ASTER imagery, and measured for soil electrical conductivity (ECe), clay content and soil organic matter (OM). An ASTER image covering the study area was preprocessed, and two predictive models, multivariate adaptive regression splines (MARS) and the partial least squares regression (PLSR), were constructed based on the ASTER spectra. For all three soil properties, the results of MARS models were better than those of the respective PLSR models, with cross-validation estimated R2 of 0.85 and 0.80 for ECe, 0.94 and 0.90 for clay content and 0.79 and 0.73 for OM. Independent validation of ECe, clay content and OM maps with 32 soil samples showed the better performance of the MARS models, with R2 = 0.81, 0.89 and 0.73, respectively, compared to R2 = 0.78, 0.87 and 0.71 for the PLSR models. The results indicated that MARS is a more suitable and superior modeling technique than PLSR for the estimation and mapping of soil salinity (ECe), clay content and OM. The method developed in this paper was found to be reliable and accurate for digital soil mapping in arid and semi-arid environments.

Rapid prediction of total petroleum hydrocarbons concentration in contaminated soil using vis-NIR spectroscopy and regression techniques

Visible and near infrared spectrometry (vis-NIRS) coupled with data mining techniques can offer fast and cost-effective quantitative measurement of total petroleum hydrocarbons (TPH) in contaminated soils. Literature showed however significant differences in the performance on the vis-NIRS between linear and non-linear calibration methods. This study compared the performance of linear partial least squares regression (PLSR) with a nonlinear random forest (RF) regression for the calibration of vis-NIRS when analysing TPH in soils. 88 soil samples (3 uncontaminated and 85 contaminated) collected from three sites located in the Niger Delta were scanned using an analytical spectral device (ASD) spectrophotometer (350-2500nm) in diffuse reflectance mode. Sequential ultrasonic solvent extraction-gas chromatography (SUSE-GC) was used as reference quantification method for TPH which equal to the sum of aliphatic and aromatic fractions ranging between C10 and C35. Prior to model development, spectra were subjected to pre-processing including noise cut, maximum normalization, first derivative and smoothing. Then 65 samples were selected as calibration set and the remaining 20 samples as validation set. Both vis-NIR spectrometry and gas chromatography profiles of the 85 soil samples were subjected to RF and PLSR with leave-one-out cross-validation (LOOCV) for the calibration models. Results showed that RF calibration model with a coefficient of determination (R2) of 0.85, a root means square error of prediction (RMSEP) 68.43mgkg-1, and a residual prediction deviation (RPD) of 2.61 outperformed PLSR (R2=0.63, RMSEP=107.54mgkg-1 and RDP=2.55) in cross-validation. These results indicate that RF modelling approach is accounting for the nonlinearity of the soil spectral responses hence, providing significantly higher prediction accuracy compared to the linear PLSR. It is recommended to adopt the vis-NIRS coupled with RF modelling approach as a portable and cost effective method for the rapid quantification of TPH in soils.

Comparison between Random Forests, Artificial Neural Networks and Gradient Boosted Machines Methods of On-Line Vis-NIR Spectroscopy Measurements of Soil Total Nitrogen and Total Carbon

Accurate and detailed spatial soil information about within-field variability is essential for variable-rate applications of farm resources. Soil total nitrogen (TN) and total carbon (TC) are important fertility parameters that can be measured with on-line (mobile) visible and near infrared (vis-NIR) spectroscopy. This study compares the performance of local farm scale calibrations with those based on the spiking of selected local samples from both fields into an European dataset for TN and TC estimation using three modelling techniques, namely gradient boosted machines (GBM), artificial neural networks (ANNs) and random forests (RF). The on-line measurements were carried out using a mobile, fiber type, vis-NIR spectrophotometer (305–2200 nm) (AgroSpec from tec5, Germany), during which soil spectra were recorded in diffuse reflectance mode from two fields in the UK. After spectra pre-processing, the entire datasets were then divided into calibration (75%) and prediction (25%) sets, and calibration models for TN and TC were developed using GBM, ANN and RF with leave-one-out cross-validation. Results of cross-validation showed that the effect of spiking of local samples collected from a field into an European dataset when combined with RF has resulted in the highest coefficients of determination (R2) values of 0.97 and 0.98, the lowest root mean square error (RMSE) of 0.01% and 0.10%, and the highest residual prediction deviations (RPD) of 5.58 and 7.54, for TN and TC, respectively. Results for laboratory and on-line predictions generally followed the same trend as for cross-validation in one field, where the spiked European dataset-based RF calibration models outperformed the corresponding GBM and ANN models. In the second field ANN has replaced RF in being the best performing. However, the local field calibrations provided lower R2 and RPD in most cases. Therefore, from a cost-effective point of view, it is recommended to adopt the spiked European dataset-based RF/ANN calibration models for successful prediction of TN and TC under on-line measurement conditions.

Almost 25 years of chromatographic and spectroscopic analytical method development for petroleum hydrocarbons analysis in soil and sediment: state-of-the-art, progress and trends

ABSTRACT This review provides a critical insight into the selection of chromatographic and spectroscopic techniques for semi-quantitative and quantitative detection of petroleum hydrocarbons in soil and sediment matrices. Advantages and limitations of both field screening and laboratory-based techniques are discussed and recent advances in chemometrics to extract maximum information from a sample by using the optimal pre-processing and data mining techniques are presented. An integrated analytical framework based on spectroscopic techniques integration and data fusion for the rapid measurement and detection of on-site petroleum hydrocarbons is proposed. Furthermore, factors influencing petroleum hydrocarbons analysis in contaminated samples are discussed and recommendations on how to reduce their influence provided.

Evaluation of vis-NIR reflectance spectroscopy sensitivity to weathering for enhanced assessment of oil contaminated soils

This study investigated the sensitivity of visible near-infrared spectroscopy (vis-NIR) to discriminate between fresh and weathered oil contaminated soils. The performance of random forest (RF) and partial least squares regression (PLSR) for the estimation of total petroleum hydrocarbon (TPH) throughout the time was also explored. Soil samples (n = 13) with 5 different textures of sandy loam, sandy clay loam, clay loam, sandy clay and clay were collected from 10 different locations across the Cranfield Universitys Research Farm (UK). A series of soil mesocosms was then set up where each soil sample was spiked with 10 ml of Alaskan crude oil (equivalent to 8450 mg/kg), allowed to equilibrate for 48 h (T2 d) and further kept at room temperature (21 °C). Soils scanning was carried out before spiking (control TC) and then after 2 days (T2 d) and months 4 (T4 m), 8 (T8 m), 12 (T12 m), 16 (T16 m), 20 (T20 m), 24 (T24 m), whereas gas chromatography mass spectroscopy (GC-MS) analysis was performed on T2 d, T4 m, T12 m, T16 m, T20 m, and T24 m. Soil scanning was done simultaneously using an AgroSpec spectrometer (305 to 2200 nm) (tec5 Technology for Spectroscopy, Germany) and Analytical Spectral Device (ASD) spectrometer (350 to 2500 nm) (ASDI, USA) to assess and compare their sensitivity and response against GC-MS data. Principle component analysis (PCA) showed that ASD performed better than tec5 for discriminating weathered versus fresh oil contaminated soil samples. The prediction results proved that RF models outperformed PLSR and resulted in coefficient of determination (R2) of 0.92, ratio of prediction deviation (RPD) of 3.79, and root mean square error of prediction (RMSEP) of 108.56 mg/kg. Overall, the results demonstrate that vis-NIR is a promising tool for rapid site investigation of weathered oil contamination in soils and for TPH monitoring without the need of collecting soil samples and lengthy hydrocarbon extraction for further quantification analysis.

Optimal sample selection for measurement of soil organic carbon using on-line vis-NIR spectroscopy

Abstract The selection of samples for modelling of visible and near infrared (vis-NIR) spectra for prediction of soil organic carbon (OC) is a crucial step for improving model prediction performance. This paper aims at comparing three soil sample selection methods coupled with spiking technique for improving on-line prediction performance of OC. Sample selection methods included random selection (RS), Kennard-Stone (KS) algorithm and similarity analysis (SA). Soil vis-NIR spectra was measured with an on-line fibre-type vis-NIR spectrophotometer (tec5 Technology for Spectroscopy, Germany), with a spectral range of 305–2200 nm. A multiple field sample set (268 samples) was merged with samples (148 samples) collected from one target field, and the resulted sample set was subjected to the three sample selection methods. After dividing spectra into calibration and prediction sets, partial least squares regression (PLSR) was run on the calibration set to develop calibration models for OC, and resulted models were validated using samples of the prediction set. Results show that SA performed generally better than its competitors, especially when there were 58 spiked samples used in the calibration set (54% of total spiked samples of 106), with the best residual prediction deviation (RPD) and root mean squares error of prediction (RMSEP) of 2.14–2.54 and 0.16–0.15% for laboratory and on-line prediction. KS and RS performed similarly, but depending on the size of the calibration set, KS produced slightly better models. This indicates that the proposed SA coupled with spiking holds great potential in the optimization of a calibration set size and may serve as a novel and efficient tool for balancing the cost and quality of vis–NIR calibrations for estimating OC.