Comparison between Variable-Selection Algorithms in PLS Regression with Near-Infrared Spectroscopy to Predict Selected Metals in Soil.

Soil is one of the Earth's most important natural resources. The presence of metals can decrease environmental quality if present in excessive amounts. Analyzing soil metal contents can be costly and time consuming, but near-infrared (NIR) spectroscopy coupled with chemometric tools can offer an alternative. The most important multivariate calibration method to predict concentrations or physical, chemical or physicochemical properties as a chemometric tool is partial least-squares (PLS) regression. However, a large number of irrelevant variables may cause problems of accuracy in the predictive chemometric models. Thus, stochastic variable-selection techniques, such as the Firefly algorithm by intervals in PLS (FFiPLS), can provide better solutions for specific problems. This study aimed to evaluate the performance of FFiPLS against deterministic PLS algorithms for the prediction of metals in river basin soils. The samples had their spectra collected from the region of 1000-2500 nm. Predictive models were then built from the spectral data, including PLS, interval-PLS (iPLS), successive projections algorithm for interval selection in PLS (iSPA-PLS), and FFiPLS. The chemometric models were built with raw data and preprocessed data by using different methods such as multiplicative scatter correction (MSC), standard normal variate (SNV), mean centering, adjustment of baseline and smoothing by the Savitzky-Golay method. The elliptical joint confidence region (EJCR) used in each chemometric model presented adequate fit. FFiPLS models of iron and titanium obtained a relative prediction deviation (RPD) of more than 2. The chemometric models for determination of aluminum obtained an RPD of more than 2 in the preprocessed data with SNV, MSC and baseline (offset + linear) and with raw data. The metals Be, Gd and Y failed to obtain adequate models in terms of residual prediction deviation (RPD). These results are associated with the low values of metals in the samples. Considering the complexity of the samples, the relative error of prediction (REP) obtained between 10 and 25% of the values adequate for this type of sample. Root mean square error of calibration and prediction (RMSEC and RMSEP, respectively) presented the same profile as the other quality parameters. The FFiPLS algorithm outperformed deterministic algorithms in the construction of models estimating the content of Al, Be, Gd and Y. This study produced chemometric models with variable selection able to determine metals in the Ipojuca River watershed soils using reflectance-mode NIR spectrometry.

Distribution of rare earth elements in soils of contrasting geological and pedological settings to support human health assessment and environmental policies.

Establishing quality reference values (QRVs) for rare earth elements (REEs) in soils is essential for the screening of these emergent contaminants. Currently, Brazil has the second-largest reserve of REEs, but data regarding background concentrations and distributions in soils remain scarce. The aim of this study was to establish the QRVs and assess the spatial distribution of REEs in soils, including REE fractionations and anomalies in (Piauí) state (251,529.186 km2), northeastern Brazil. This study reports the most detailed data on REE geochemistry in Brazilian soils. A total of 243 composite soil samples was collected at 0-20 cm depth. The mean background concentrations in soils followed the abundance of the earth's upper crust: Ce > La > Nd > Pr > Sm > Dy > Gd > Er > Yb > Eu > Tb > Lu. The ∑REEs (mg kg-1) showed the following order based on the individual mesoregions of Piauí state: Southeast (262.75) > North and Central-North (89.68) > Southwest (40.33). The highest QRVs were observed in the Southeast mesoregion. The establishment of QRVs based on the mesoregion scale improves data representativeness and the monitoring of natural REE values by identifying hot spots. Geostatistical modeling indicated significant local variability, especially in the Southeast mesoregion. The levels of these elements in this spatial zone are naturally higher than the other values across Piauí state and the mesoregion itself and indicate a high potential to exceed the QRVs. Our approach provides much needed data to help strengthen policies for both human health and environmental protection.

Near-infrared spectroscopy for the prediction of rare earth elements in soils from the largest uranium-phosphate deposit in Brazil using PLS, iPLS, and iSPA-PLS models.

The largest uranium-phosphate deposit in Brazil also contains considerable levels of rare earth elements (REEs), which allows for the co-mining of these three ores. The most common methods for REE determination are time-consuming and demand complex sample preparation and use of hazardous reagents. Thus, the development of a safer and faster method to predict REEs in soil could aid in the assessment of these elements. We investigated the efficiency of near-infrared (NIR) spectroscopy to predict REEs in the soil of the uranium-phosphate deposit of Itataia, Brazil. We collected 50 composite topsoil samples in a well-distributed sampling grid along the deposit. The NIR measures in the soils ranged from 750 to 2500 nm. Three partial least squares regressions (PLSR) were selected to calibrate the spectra: full-spectrum partial least squares (PLS), interval partial least squares (iPLS), and successive projections algorithms for interval selection in partial least squares (iSPA-PLS). The concentrations of REEs were measured by inductively coupled plasma optical emission spectroscopy (ICP-OES). In addition to raw spectral data, we also used spectral pretreatments to investigate the effects on prediction results: multiplicative scatter correction (MSC), Savitzky-Golay derivatives (SG), and standard normal variate transformation (SNV). Positive results were obtained in PLS for La and ΣLREE using MSC pretreatment and in iSPA-PLS for Nd and Ce using raw data. The accuracy of the measurements was related to the REE concentration in soil; i.e., elements with higher concentrations tended to present more accurate results. The results obtained here aim to contribute to the development of NIR spectroscopy techniques as a tool for mapping the concentrations of REEs in topsoil.

Combining geospatial analyses to optimize quality reference values of rare earth elements in soils.

High pedological and geological variability can trigger the formation of REE hotspots, causing a need to optimize the establishment of quality reference values (QRVs). Thus, we determined the background concentrations of REEs in the soils of an emerging Brazilian state and used a combination of Moran's I and indicator kriging to identify REE hotspots and determine QRVs. A total of 100 composite soil samples was collected at a 0.20 m depth to establish background concentrations, QRVs, and spatial distribution and to elaborate probability maps for REEs. The QRVs established for soils were the following (mg kg-1): La (27.21), Ce (57.26), Pr (10.49), Nd (24.29), Sm (4.75), Eu (0.90), Gd (4.22), Tb (0.82), Dy (1.54), Ho (0.38), Er (1.23), Yb (1.07), Lu (0.24), Y (10.65), and Sc (3.70). It was possible to draw attention to the Northwest and Southwest regions of the Rio Grande do Norte (RN) state, due to the formation of REE hotspots, indicated by Moran's I, and a high tendency to exceed the QRVs, confirmed by the indicator kriging. The high background concentrations and geochemical patterns for REEs showed that a single QRV for each REE and the entire state can neglect specific environmental characteristics and misrepresent the natural geochemistry of the soil. Thus, specific QRVs were established to optimize the monitoring of natural REE values by identifying hotspot areas. The criteria established here may be useful for other groups of potentially toxic elements, provided that observations meet the requirements of the spatial autocorrelation and kriging analyses. Graphical abstract.

Thorium content in soil, water and sediment samples and fluvial sediment-associated transport in a catchment system with a semiarid-coastal interface, Brazil.

Thorium (Th) is one of the main sources of natural radiation to ecosystems. However, data regarding Th concentrations in rocks, soil, water and sediments are currently scarce. Accordingly, this study aimed to establish background concentrations and quality reference values (QRVs) for Th in the environmentally impacted Ipojuca River catchment in Brazil, where the weathering of granites releases Th into the environment. Additionally, the study aimed to calculate Th fluxes in water, and both bed and suspended sediment. The mean Th concentration in the study catchment soils was 28.6 mg kg-1. The QRV for Th was estimated to be 21 mg kg-1 and 86.3 Bq kg-1. Bed and suspended sediment-associated concentrations ranged from 2.8 to 32.9 mg kg-1. Suspended sediment-associated discharge (3.42 t year-1) accounted for more than 99% of the total Th flux, while the dissolved phase transport was negligible in comparison. At the downstream cross section in the study catchment, suspended sediment samples exhibited Th concentrations similar to those observed in rivers impacted by mining activities. The discharge of sediment to the ocean from the study area is mainly triggered by soil erosion processes in the hotspot region (middle-inferior course). It is essential to identify Th hotspots before establishing environmental policies regarding human health and environmental protection.

Watershed scale assessment of rare earth elements in soils derived from sedimentary rocks.

Despite the rare earth elements (REEs) being considered as emerging contaminants, their natural values and possible anthropogenic enrichments in soils have not been studied well in Brazil. The intensive use of conditioners and fertilizers in agricultural frontiers from Brazilian Cerrado can increase the concentration of REE in soils of the region. In this context, the objectives of this study were to determine the natural content and establish quality reference values (QRV) for REEs in soils of a watershed from Brazilian Cerrado composed of sedimentary rocks and to evaluate the influence of agricultural cultivation and the spatial variability of these elements. Thirty and twenty-six composite soil samples were collected under native vegetation and soybean cultivation, respectively. The background concentrations followed the order (mg kg-1) Ce > Nd > La > Pr > Sm > Yb > Er > Eu > Dy. The QRVs established were as follows (mg kg-1): La (1.76), Ce (5.20), Pr (0.74), Nd (1.35), Sm (0.38), Eu (0.06), Dy (0.15), Er (0.12), and Yb (0.14). Lantanium, Ce, and Er exhibited strong spatial dependence, while Eu, Dy, and Yb showed weak or total absence of spatial dependence. The spherical model was most suitable for the spatial characteristics of REEs. The parent material, mainly characterized by soils derived from sedimentary rocks (i.e., sandstone), was the primordial source of REEs for soils and that there was no or little effect of agricultural practices on these levels. Our data reinforced the need for geochemical mapping at the watershed scale, since they are important conservation units.