Search

Quartz is one of the most common minerals in soils, mostly present in sand and silt fractions. Although quartz is basically formed of SiO2, other elements can be easily detected and assessed nowadays using a portable X-ray fluorescence (pXRF) spectrometer. Our study aims to evaluate the chemical composition of different quartz varieties, identifying their main elements, and relating them to soil attributes. Six quartz varieties (hyaline, amethyst, milky, rose, smoky, and ferriferous) were analyzed via pXRF and 13 oxides/trace elements were identified and used for quartz discrimination (Al2O3, CaO, P2O5, SiO2, Cl, Cr, Fe, K2O, Mn, Rb, S, Ti, and V). Hyaline quartz was characterized by the highest SiO2 and the lowest contents of other elements. Al2O3 was the second-highest compound present in all varieties of quartz, reaching 21,547 mg kg−1 in the smoky variety. S, P2O5, Cl, SiO2, and K2O were the main components determined by Random Forest algorithm for discriminating quartz varieties. Some elements detected may serve as a reserve of nutrients to plants to be released in soils along weathering, depending on quartz particle size, soil texture, leaching, and associated attributes. pXRF enhanced the information on chemical characterization of quartz varieties, without the generation of chemical pollutants. [ABSTRACT FROM AUTHOR]

Despite the increasing adoption of proximal sensors worldwide, rare works have coupled proximal with remotely sensed data to spatially predict soil properties. This study evaluated the contribution of proximal and remotely sensed data to predict soil texture and available contents of micronutrients using portable X-ray fluorescence (pXRF) spectrometry, magnetic susceptibility (MS), and terrain attributes (TA) via random forest algorithm. Samples were collected in Brazil from soils with high, moderate, and low weathering degrees (Oxisols, Ultisols, Inceptisols, respectively), and analyzed by pXRF and MS and for texture and available micronutrients. Seventeen TA were generated from a digital elevation model of 12.5 m spatial resolution. Predictions were made via: (i) TA; (ii) TA + pXRF; (iii) TA + MS; (iv) TA + MS + pXRF; (v) MS + pXRF; and (vi) pXRF; and validated via root mean square error (RMSE) and coefficient of determination (R2). The best predictions were achieved by: pXRF dataset alone for available Cu (R² = 0.80) and clay (R2 = 0.67) content; MS + pXRF dataset for available Fe (R2 = 0.68) and sand (R2 = 0.69) content; TA + pXRF + MS dataset for available Mn (R2 = 0.87) content. PXRF data were key to the best predictions. Soil property maps created from these predictions supported the adoption of sustainable soil management practices. [ABSTRACT FROM AUTHOR]

Many factors can influence the results obtained by portable X-ray fluorescence analysis (pXRF). The effect of soil organic matter on pXRF results is not satisfactory understood. Thus, we conducted this study to verify the effect of organic matter removal on oxide determination by pXRF in Oxisols. To obtain soil material with different organic matter contents and maintaining the same elemental composition from soil minerals, six contrasting Oxisols were heated in a muffle furnace for 30 min at the following temperatures (°C): 100; 200; 300; 400; 500 and 600. After heating, the soil samples were scanned using a pXRF Bruker® S1 Titan LE model (Dual Soil mode) for 60 s and the contents of SiO2, Al2O3, Fe2O3, TiO2, P2O5, and MnO were recorded. The soil organic matter presence underestimated the pXRF results for lightest oxides (Si and Al) compared to heaviest oxides (Fe, Ti, and Mn). These oxides are important for tropical soils classification and for many soil-related studies and pXRF technology has been a useful tool for soil chemical characterization. Our findings contribute to more suitable use of pXRF highlighting the possible effect of organic matter. [ABSTRACT FROM AUTHOR]

This work aimed to calibrate Mehlich 1, Mehlich 3, Bray 1, Olsen, and ion-exchange resin extraction methods with maize phosphorus (P) responses in a pot study with lowland and upland soils with different P-buffer capacities and to evaluate whether the calibration can be enhanced through the knowledge of remaining P. The experimental design was completely randomized with four replications in a factorial arrangement involving five P concentrations and four lowland or seven upland soils. The remaining P for each soil was determined, P-buffer capacity was estimated, and the soils were grouped according to the results. Correlation coefficients showed that the remaining P is strongly dependent on clay and soil organic-matter content, and its determination was useful to the evaluation of the extractants. The classification and grouping of soils according to their P-buffer capacity improved the correlations between extracted P and plant response for Mehlich 1 and Bray 1 extractants. The Mehlich 3, Olsen, and resin methods presented better performances, independent of soil grouping. [ABSTRACT FROM AUTHOR]

Rare earth elements have been increasingly used in modern societies and soils are likely to be the final destination of several REE-containing (by)products. This study reports REE contents for topsoils (0–20 cm) of 175 locations in reference (n = 68) and cultivated (n = 107) areas in Brazil. Benchmark soil samples were selected accomplishing a variety of environmental conditions, aiming to: i) establishing natural background and anthropogenic concentrations for REE in soils; ii) assessing potential contamination of soils - via application of phosphate fertilizers - with REE; and, iii) predicting soil-REE contents using biomes, soil type, parent material, land use, sand content, and biomes-land use interaction as forecaster variables through generalized least squares multiple regression. Our hypotheses were that the variability of soil-REE contents is influenced by parent material, pedogenic processes, land use, and biomes, as well as that cultivated soils may have been potentially contaminated with REE via input of phosphate fertilizers. The semi-total concentrations of REE were assessed by inductively coupled plasma mass spectrometry (ICP-MS) succeeding a microwave-assisted aqua regia digestion. Analytical procedures followed a rigorous QA/QC protocol. Soil physicochemical composition and total oxides were also determined. Natural background and anthropogenic concentrations for REE were established statistically from the dataset by the median plus two median absolute deviations method. Contamination aspects were assessed by REE-normalized patterns, REE fractionation indices, and Ce and Eu anomalies ratios, as well as enrichment factors. The results indicate that differences in the amounts of REE in cultivated soils can be attributed to land use and agricultural sources (e.g., phosphate-fertilizer inputs), while those in reference soils can be attributed to parent materials, biomes, and pedogenic processes. The biomes, land use, and sand content helped to predict concentrations of light REE in Brazilian soils, with parent material being also of special relevance to predict heavy REE contents in particular. [Display omitted] • Semi-total REE concentrations were quantified in mineral surface soils in Brazil. • The geochemical background for REE was determined by the MAD method. • Anthropogenic sources of REE in cultivated soils were evidenced. • Soil-REE contents are mostly controlled by biomes, land use, and soil granulometry. • Heavy REE soil contents are highly dependent on soil parent material. [ABSTRACT FROM AUTHOR]

• Usefulness of varied proximal sensors to study tropical soil minerals was tested. • Representative soil classes and parent materials of tropical regions were studied. • Proximal sensor data help characterize tropical soil mineralogy. • Accurate predictions of contents of clay-sized minerals were achieved. • pXRF outperformed Vis-NIR for predictions of most clay-sized minerals. Proximal sensors combined with X-ray diffraction (XRD) have optimized soil characterization, but scarce studies have focused on predicting the contents of minerals under this scope. The objectives herein were to: a) use the portable X-ray fluorescence (pXRF) spectrometry, diffuse reflectance spectroscopy in the range of visible and near-infrared (Vis-NIR), magnetic susceptibility (χ) and XRD to characterize the mineralogy of soils derived from representative Brazilian soil parent materials, and b) create models to quantify the minerals obtained via XRD. Twenty-two soil profiles developed from gabbro, gneiss, quartzite, mineral and organic sediments were described with 53 soil horizons sampled. Each sample had the sand, silt and clay fractions separated and analyzed with XRD, pXRF, χ, and Vis-NIR. Models were created using the Random Forest algorithm permuting the following predictor variables (separately and combined): pXRF, parent material (PM), χ, soil texture (sand, silt, and clay content), and Vis-NIR. Models' accuracy was calculated using the leave-one-out cross-validation method. Si, Al, Fe, Ca, K, and Ti contents obtained by pXRF and the χ discriminated the soil particle size fractions according to the parent material. XRD analysis allowed the evaluation of the pedogenetic development of soils and their relation to the respective parent material. The best models for mineral contents were found for hematite (Hm) (1 0 4)+(Gt) (1 3 0) (R2 = 0.85), Hm (1 1 0) + Mh (1 3 1) (R2 = 0.76), kaolinite (Kt) (0 0 1) (R2 = 0.73), Kt (0 0 2) (R2 = 0.80), mica (Mc) (0 0 1) (R2 = 0.77), and Mc (0 2 0) + Kt (0 2 0) (R2 = 0.81). Clay mineralogy content was accurately modeled using only pXRF and parent material data. This approach can facilitate and speed up detailed soil mineralogy characterization. Further studies are encouraged to model the content of minerals found in the sand and silt fractions of soils with diverse mineralogy via proximal sensors and using larger data sets. [ABSTRACT FROM AUTHOR]

Although parent material (PM) is one of the five soil formation factors providing key information on soil variability, the complexity of PM distributions and the difficulty of reaching PM in deep soils prevent its detailed assessment. Proximal sensors such as portable X-ray fluorescence (pXRF) spectrometer and magnetic susceptibility (MS) may be helpful in predicting soil PM in a more practical and accessible way. This pilot study aimed to create spatial PM predictive models for three distinct PMs (charnockite, mudstone, and alluvial sediments) of an experimental farm (Brazil) through random forest (RF) algorithm based on soil samples analyzed via pXRF and MS. Soils were sampled in A and B horizons following a regular-grid design covering the whole study area. The RF algorithm was calibrated to predict PMs using samples from the B horizon of soils with known PM. The prediction model was applied to the area for mapping PM across the whole farm. For validation, PM was identified at 15 different sites and compared with the predicted PM shown on the maps via overall accuracy, Kappa coefficient, producer's and user's accuracies. Al, Fe, Si, Ti, and MS proximal sensor data discriminated well among soils derived from charnockite, mudstone, and alluvial sediments. The map built based on B horizon data showed greater accuracy (overall accuracy = 0.93, Kappa coefficient = 0.85, user's accuracy = 0.92, and producer's accuracy = 0.97) than the map built from the model using A horizon samples (0.73, 0.48, 0.48, and 0.58). These results could represent alternative methods for reducing costs and accelerating the assessment of soil PM spatial variability, supporting soil mapping, and optimized agronomic and environmental decision-making. • Proximal sensors were used for soil parent material (PM) prediction. • pXRF and MS differentiated soil types and their respective parent materials. • Different parent materials were predicted using random forest algorithm. • pXRF and MS combined with RF were successful for PM prediction. [ABSTRACT FROM AUTHOR]

Maghemite (Mh) is a ferrimagnetic mineral that governs magnetism in the clay fraction, especially in well weathered soils from tropical and subtropical regions. In this work, we assessed different methods for Mh quantification [1: area of the reflections obtained by X-ray diffraction (XRD) (standard method); 2: Rietveld refinement; 3: selective dissolution with H 2 SO 4 and; 4: magnetic susceptibility (χ)] by application to Brazilian soils with contrasting parent materials (itabirite, basalt, gabbro, tuffite, gneiss and amphibolite). The methods based on H 2 SO 4 dissolution were found to overestimate the Mh content with respect to the standard method based on X-ray diffraction. This was particularly so with samples from gneiss and amphibolite and samples with higher contents in poorly crystalline Fe oxides. Mh contents based on magnetic susceptibility measurements or the XRD-Rietveld refinement were closer to those provided by the standard method. Mh was found to accumulate in well weathered soils and soils with a high Fe 2 O 3 content. Exception were soils from tuffite, which had high Fe 2 O 3 contents (~330 g kg −1 ), exhibited low Mh levels (~5.5 g kg −1 ). Irrespective of the parent material, Mh particles typically fall in a single domain magnetic behavior; however, they can also acquire multidomain characteristics, probably as a result of isomorphic substitution and aggregation with hematite. [ABSTRACT FROM AUTHOR]

Erosion is a major cause of soil degradation. Understanding the sources of sediment and the processes that exacerbate this problem is crucial for soil conservation. Remote and proximal sensing technologies have contributed to advance our understanding of soil erosion processes and support soil conservation research. The objective of this study is to use two proximal sensors, a portable X-ray fluorescence spectrometer (pXRF) and a magnetic susceptibilimeter, to identify the origin of sediments deposited in the lower portion of a drainage subbasin by using a fingerprinting approach. Five sediment tracing elements were selected: Fe, Si, Al, Ti, and Zr. The sediment source area is composed of gabbro, granite, and felsic orthogneiss, upon which two soil classes were formed: Rhodic Ferralsol (Dystric) (Latossolo Vermelho Distroférrico típico, LV) and Haplic Acrisol (Argissolo Vermelho-Amarelo Distrófico típico, PVA). In the source area, 35 samples were collected from the topsoil layer (0–2 cm of depth) using a regular grid of 50 × 50 m2. In the deposition area, 24 samples were collected in a regular grid of 7 × 7 m2. The source of the sediments was determined using cluster analysis with data provided by the proximal sensors and topographic attributes. Results indicated that the Haplic Acrisol, which is more susceptible to erosion, contributed most to sediment deposition. Analysis of pXRF data showed its potential to identify and differentiate sediments and their sources. Cluster models based on the abundance of Fe, Si and Al, as well as the Ti/Zr ratio, demonstrated the best results with a global accuracy of 88% and a Kappa coefficient of 0.74. The use of the susceptibilimeter data alone did not yield satisfactory results. Results indicate the usefulness of pXRF in identifying the source of sediments in drainage basins, making it a valuable tool in soil conservation and erosion assessments. • pXRF was capable of identifying the origin of sediments. • Fe, Si, Al and the Ti/Zr ratio were efficient in identifying sediment sources. • The Chromic Acrisol contributed most with sediment deposition. • Magnetic susceptibility contributed little to identify sediment sources. [ABSTRACT FROM AUTHOR]

Phosphorus (P) retention capacity of soils directly affects the management of phosphate fertilization, and has economic and environmental importance. Brazil has expressive agricultural production and high edaphoclimatic variability; still, the detailed spatial distribution of P retention capacity is not known. Thus, machine learning models were created to estimate the P retention capacity of soils at superficial horizons using the variables: clay content, sand content, soil organic matter, pH, base saturation. The four best models were combined to create an ensemble (a combination of models), which was applied in a dataset (5524 samples) that comprises a sample data from the entire Brazilian territory. The ensemble was used to build a map of P retention capacity. Most of the country presented soils with medium P retention (40–60%), and large areas with very high retention capacity were found in the south of Brazil, a subtropical region mostly associated with clayey soils. High P retention was observed for Histosols, probably related to humic-Al(Fe) complexes. Nitisol, Ferralsol, Gleysol and Cambisol classes presented high P retention, especially for clayey soils with qualifiers related to bases/weathering (Dystric) and organic matter (Umbric and Humic). The correlation between Fe oxide minerals and P retention varied with soil color and drainage conditions mainly associated to redoximorphic reactions; the same was not observed for Al oxide minerals. This pioneer study revealed the distribution of P retention capacity across different soil classes in Brazil, which can help to define land use suitability and soil P management strategies that ensure profitable agricultural exploration. [Display omitted] • In 34% of the country, soils had P retention capacity framed as Low and 22% as High. • The Histosol class presenting soils with elevated P retention capacity (>60%). • For mineral soils, the highlighted classes were Nitisol, Ferralsol, Gleysol and Cambisol. • Soil color and drainage conditions had little effect on soil P retention capacity. [ABSTRACT FROM AUTHOR]

Diverse projects are being carried out worldwide focusing on development of more accurate soil maps and one of the most valuable sources of data are the existing soil maps. This work aimed to (i) compare two data mining tools, KnowledgeMiner and decision trees, to retrieve legacy soil data from a detailed soil map, (ii) to create and validate the predicted soil maps in the field with the objective to identify the best method for modeling and refining soil maps, (iii) extrapolating soils information to the surrounding similar areas and (iv) to assess the accuracy of this soil map. The study was carried out in Minas Gerais state, Southeastern Brazil. From a detailed soil map, information of 12 terrain attributes was retrieved from the entire polygon of each mapping unit of the map (MUP) and from a circular buffer around the sampled points (CBP). KnowledgeMiner and decision trees were employed to retrieve information per soil class and soil maps were created per method. A field validation of 20 samples was chosen by a cost-constrained conditioned Latin hypercube sampling scheme and the accuracy of all maps was assessed using a global index, Kappa index, and errors of omission and commission. The KnowledgeMiner MUP map had a greater accuracy than the other methods, being even more accurate than the original map, accounting for 80% of global index and a Kappa index of 0.6524. The information extracted by KnowledgeMiner provided rules for mapping the watershed surroundings with 70.97% of global index and a kappa index of 0.5586. Legacy soil data extracted by KnowledgeMiner from a detailed soil map and used to model soil class distribution outperformed decision trees, promoted improvements on the existing soil map, and allows for the creation of a low cost soil map for the surroundings of the study area. [ABSTRACT FROM AUTHOR]

• Does combining proximal sensor always bring synergy for predicting soil texture? • Vis-NIR, NixPro TM , and pXRF data were tested for soil texture predictions. • PXRF provided key information to create accurate prediction models. • Smoothed Vis-NIR data preprocessing and dry NixPro TM condition positively influenced the results. • Soil texture and textural classes were accurately predicted by at least one approach. Soil texture is a primary variable influencing many soil chemical-physical-biological processes, providing important information for decision-making regarding sustainable soil management. The standard traditional methods for determining soil texture, however, are performed manually and are time-consuming, costly, and generate chemical wastes. As an alternative, portable X-ray fluorescence (pXRF) spectrometry and visible near-infrared spectroscopy (Vis-NIR) have been increasingly used worldwide to predict soil attributes. Other sensors (e.g., NixPro TM color sensor) are also promising, but less evaluated to date. Thus, investigations towards proximal sensor data fusion for prediction of soil textural separates (clay, silt, and total, coarse, and fine sand contents) and soil textural classes (loam, loamy sand, etc) in tropical soils are rare. Therefore, this study aimed to evaluate proximal sensor data for predicting soil particle size fractions and soil textural classes (both Family particle size classes and USDA soil texture triangle) via random forest algorithm in tropical regions. A total of 464 soil samples were collected from A (n = 208) and B (n = 256) horizons in Brazil. Soil samples were submitted to laboratory analyses for soil texture and proximal sensor (pXRF, Vis-NIR, and NixPro TM) scanning. Samples were randomly split into 70% for modeling and 30% for validation. The best approach varied according to the predicted attribute; however, pXRF data were key information for soil texture prediction accuracy. The best results delivered highly accurate predictions via the aforementioned proximal sensors for rapid assessment of soil texture (total sand R2 = 0.84, RMSE = 7.60%; silt 0.83, 6.11%; clay 0.90, 5.64%; coarse sand 0.87, 6.30%; fine sand 0.82, 5.27%). Categorical prediction accuracy for soil textural classes (Family particle size classes, overall accuracy = 0.97, Kappa index = 0.95; USDA soil texture triangle, 0.83, 0.73) was enhanced when the predictions were made by soil order sub-datasets. Smoothed Vis-NIR preprocessing and dry NixPro TM color data positively influenced the results. The results reported here represent alternatives for reducing costs and time needed for evaluating soil texture, supporting agronomic and environmental strategies in Brazilian conditions. Further works should extend the results of this study to temperate regions to corroborate the conclusions presented herein regarding the fusion of these three proximal sensors. [ABSTRACT FROM AUTHOR]

• Characterization of 4 profiles in some of the oldest erosion surfaces in the world. • Detailed assessment of extremely deep and old soils, reaching depths of 26 m. • Major soil property changes found between R/C horizons, rather than C/B horizons. • Evidences of the influence of a pre-weathered mantle in the formation of soils. Brazil has some of the oldest soils of the world. Landscapes in southeastern Brazil are very old and stable. Some of the soils formed therein have suffered many pedogenetic cycles and hence are difficult to study thoroughly due to their depth and age. In the southeastern State of Minas Gerais, Brazil, we studied four soils developed from gneiss formed on these old surfaces (one sampled to a depth of 26 m) that comprised a chronosequence. For most soils of the world, the major change in properties is at the C/B horizon interface, but in these soils the major change is at the R/C interface. In each soil, a certain suite of clay minerals (kaolinite, gibbsite, goethite, and hematite) had weathered from the primary minerals of gneiss. The composition of the suite differed among the four soils, but in an individual soil it was relatively constant with depth, including through the C/B boundary. In the C horizon, many of the minerals were in the form of aggregates in the silt fraction. Across the C/B boundary there was little change in chemistry and mineralogy, but there the aggregates broke down to clay particles that were incorporated into soil structure. Overall, the chronosequence showed that soil properties changed greatly from the youngest, lowest part of the landscape, while between the older soils, higher in the landscape, the change was small. Soil properties appeared influenced by the different zones of the weathered mantle, hypothesized to be gibbsitic on the very top, kaolinitic in the most part, and bissialitic (2:1 clay minerals) closer to the fresh rock. [ABSTRACT FROM AUTHOR]

Oxisols are generally deep, weathered-leached soils that present morphologically homogeneous profiles. Despite their apparent homogeneity, different parent materials and pedogenetic processes grant these soils diverse chemical and mineralogical attributes. Proximal sensors might help to detect the variability of soil attributes along such old and deep profiles. The objective of this study was to investigate the physical, chemical and mineralogical variability of Oxisol profiles originated from different parent materials with aid of visible and near-infrared diffuse reflectance spectroscopy (Vis-NIR DRS) and portable X-ray fluorescence (pXRF) spectrometry, in order to evaluate if these proximal sensors can complement the analysis of these deep and weathered tropical soils. Two Oxisol profiles were selected: Anionic Acrudox (AA) and Typic Hapludox (TH). One hundred samples were collected in each profile down to 2 m of depth, following a grid of 10 × 10 cm, for pXRF, Vis-NIR and XRD analyses. Vis-NIR DRS spectra provided key mineralogical information in agreement with X-ray diffraction results, showing differential occurrence of kaolinite, hematite, maghemite and gibbsite/kaolinite ratio in these profiles. PXRF detected high chemical variability across the profiles and allowed to map the spatial chemical distribution of both soils. Proximal sensors provided an inexpensive and efficient way to help complement the identification of meaningful variability across Oxisol profiles that remained even after the long formation process of these very weathered tropical soils. • Very weathered and apparently homogeneous Oxisols were detailed investigated. • Anionic Acrudox (AA) and Typic Hapludox (TH) were analyzed via proximal sensing. • Proximal sensors detected detailed variations across horizons of the studied profiles. • AA showed higher contents of metals, while TH presented more Si and Al. • TH presented greater variation of properties between horizons than AA. [ABSTRACT FROM AUTHOR]

• Soil mineralogy plays a fundamental role in pXRF patterns. • Selective dissolution of SOM and iron forms can be detected by pXRF. • Low contents of SOM did not interfere in pXRF analyses. • Soil moisture must be considered for pXRF analyses in field. • Soil fertility and soil particle size distribution are related to pXRF data. Portable X-ray fluorescence (pXRF) has great potential for numerous applications in soil science. However, the basic knowledge about the effects of soil properties on pXRF spectra are still poorly studied, which may lead users to biased interpretations of mathematical models. The present study aimed to evaluate the outcomes of moisture, soil organic matter content (SOM), and iron forms on pXRF data. The work was conducted with seventeen soil samples from the central region of São Paulo state (Brazil). Three selective dissolution treatments were applied to remove: (i) soil organic matter (−SOM), ii) SOM and poorly crystalline iron forms (−o), iii) SOM and poorly crystalline plus well crystalline iron forms (−d). One additional treatment iv) including water addition (+W) was also carried out. The effects of treatments were evaluated for sandy and clayey samples. Soil particle size distribution and elemental content affected the bremsstrahlung and characteristic peaks counts. In +W, there was a generalized decrease in counts mainly for the light elements (magnesium, aluminum and silicon). Regarding the selective dissolution procedures, alterations were verified, reflecting mainly the removal power of reagents. Generally, the most pronounced alterations occurred for −d and moderate alterations for −SOM and −o. The pXRF data showed high correlation with particle size distribution and mineralogy attributes. The kaolinite, gibbsite, Fe 2 O 3 , Al 2 O 3 , SiO 2 , TiO 2 and MnO contents were quantified with satisfactory accuracy (0.61 < R2 < 0.97). The pXRF was able to detect changes caused by the selective dissolution treatments and soil particle size distribution. Sources of uncertainty, mainly soil moisture, must be considered. The understanding of the fundamentals of energy interaction with the sample matrix in the X-ray range is the starting point for characterizing the soil through pXRF. [ABSTRACT FROM AUTHOR]

Brazil presents high soil variability, however, the soil magnetic susceptibility (χ) distribution on national scale is yet to be investigated. Thus, soil data (218 samples from the literature and 36 original samples) were utilized to estimate soil low frequency χ (χ lf) across Brazil through machine learning algorithms (Generalized Additive Model Using Splines, Generalized Linear Model, Random Forest and Support Vector Machine) in order to assess the applicability of such methodology on a nationwide scale. Two variable groups were tested: texture and Fe 2 O 3 contents – Group 1; and texture, Fe 2 O 3 , Al 2 O 3 , SiO 2 contents – Group 2. The best performing model was applied to a database containing 3817 and 3973 georeferenced soil samples from superficial and subsuperficial horizons, respectively, obtained from soil surveys conducted by the Brazilian Agricultural Research Corporation. Polynomial Support Vector Machine using Group 2 variables presented the best performance (R2 = 0.91). Variable importance measurements showed that χ lf is chiefly related to Fe 2 O 3 and clay content. The spatial distribution of χ lf on national scale varied from very high (>150 × 10−6 m3 kg−1) to very low (≤5 × 10−6 m3 kg−1), considering the interpretation bands that we propose for Brazilian soils. Extensive areas of magnetic soils were associated with their formation upon mafic rocks or iron-rich formations. Despite few exceptions, the observed tendency shows that χ lf increases parallel to soil red color, drainage and weathering degree, and hence Rhodic Nitisol, Umbric Nitisol, Rhodic Ferralsol and Umbric Ferralsol presented the highest χ lf values. [Display omitted] • Rock type is a key factor to understand the soil magnetic susceptibility (χ lf) distribution. • χ lf increase in red soils with good drainage and higher weathering-leaching degree. • The distribution of χ lf varied largely according to soil classification. • Five value bands were proposed to aid in the interpretation of soil χ lf values. [ABSTRACT FROM AUTHOR]

• 1975 samples of varied soils were analyzed for fertility and elemental contents. • pXRF data was successfully used to rapidly predict soil fertility properties. • Random Forest outperformed other machine learning algorithms. • Prediction models allowed spatial distribution of soil fertility properties. • Inexpensive and environment-friendly soil fertility prediction was achieved by pXRF. Conventional soil fertility analyses are laborious, costly, time-consuming, and produce hazardous waste. The high demand of these laboratory-based analyses prompted us to investigate an environment-friendly, rapid, and inexpensive methodology for soil fertility assessment. Portable X-ray fluorescence (pXRF) spectrometry allows the determination of total elemental concentration in soils quickly, simply and without hazardous waste production. However, incipient usage of this technology for the prediction of soil fertility properties has been reported for tropical conditions. Soil samples were collected from seven Brazilian states (n = 1975) aiming to use pXRF data to predict contents of available or exchangeable Ca2+, Mg2+, Al3+, K+ and P by testing different algorithms using 70% of the samples for model training, and the remaining 30% for model validation. In addition to point data predictions, the best performing models were applied to data obtained from a farm within the studied regions with a known cropping history to create soil fertility maps and illustrate another applicability of this approach. The attested use of pXRF data and machine learning algorithms stepwise Generalized Linear Model (GLM) and Random Forest (RF) to predict the contents of relevant soil fertility properties exhibited great potential. Validation of the models corroborated that RF resulted in more accurate predictions than GLM. Validation R2 values ranged from 0.59 to 0.82. Maps created were coherent with expected distributions of soil fertility attributes. This environment-friendly methodology may be used for the assessment of soil fertility properties in a wide range of tropical and subtropical soils with minimum waste generation and reduced costs. [ABSTRACT FROM AUTHOR]

• Prediction of soil texture via pXRF and Vis-NIR DRS data was evaluated. • Combination of A and B horizons data resulted in predictions with R2 above 0.80. • In general, the best predictions were achieved using pXRF data. • RF algorithm outperformed other algorithms for soil texture prediction. • The best prediction models were obtained with pXRF + RF using B horizon data. Portable X-ray fluorescence (pXRF) spectrometry and visible near-infrared diffuse reflectance spectroscopy (Vis-NIR DRS), used separately or in tandem, have become important techniques for determination and prediction of soil attributes worldwide. However, there is little information available regarding the effectiveness of their combined use in tropical soils. This study aimed to predict soil texture using pXRF and Vis-NIR DRS, evaluating the efficiency of using these proximal sensors separately and in tandem. A total of 315 soil samples were collected from A and B horizons in Brazil. Soil samples were submitted to analyses of texture, pXRF and Vis-NIR DRS. Vis-NIR DRS spectral data pre-processing was evaluated by comparing results delivered by the derivative smoothing methods Savitzky-Golay (WT), Savitzky-Golay with Binning (WB), and data without the pre-processing treatment (WOT). Four algorithms were utilized for predictions: Gaussian Process (Gaussian), Support Vector Machine with linear (SVM-L) and radial (SVM-R) kernels, and Random Forest (RF). In general, models using only pXRF data slightly outperformed models using Vis-NIR DRS (WT, WB, WOT) data alone. Models combining data from both sensors achieved similar results to those obtained by pXRF alone. The best predictions of sand, silt, and clay contents were obtained via pXRF + RF using B horizon data, reaching R2 values of 0.91, 0.81, and 0.83, respectively. Although pXRF alone provided slightly better results, soil texture can be accurately predicted via both pXRF and Vis-NIR DRS data, separately and in tandem. These sensors can contribute to reduce costs and time required for tropical soil texture determination. [ABSTRACT FROM AUTHOR]

• First time pXRF and Vis-NIR data were combined to predict tropical soil types. • Vis-NIR spectral data preprocessing is relevant to increase its predictive efficiency. • SVM-L, SVM-R, C5.0, and RF algorithms provided efficient predictions. • RF model using pXRF and soil B horizon data outperformed other models. • Tropical soil type predictions using only pXRF data provided accurate results. Recently, portable X-ray fluorescence (pXRF) spectrometer and visible near-infrared (Vis-NIR) spectroscopy are increasingly being applied for soil types and attributes prediction, but a few works have used them combined in tropical regions. Thus, this work aimed at analyzing models' performance when predicting soil types at subgroup taxonomic level via pXRF and Vis-NIR separately and together. 315 soil samples were collected in both A and B horizons in three important Brazilian states. Samples undergone laboratorial analyses for soil classification and were submitted to pXRF and Vis-NIR (350–2500 nm) analyses. Vis-NIR spectral data preprocessing was evaluated utilizing Savitzky-Golay (WT) and Savitzky-Golay with Binning (WB) methods. Four classification algorithms were employed in modeling: Support Vector Machine with Linear (SVM-L) and Radial (SVM-R) kernel, C5.0, and Random Forest (RF). Predictions were made using only B horizon and using A + B horizon data. Overall accuracy and Cohen's Kappa index evaluated model quality. Both sensors displayed efficacy in soil types prediction. A + B horizons data combined using pXRF + Vis-NIR via SVM-R (WT and WB) delivered accurate predictions (89.32% overall accuracy and 0.75 Kappa index), but the best predictions were achieved using only B horizon data via pXRF with RF, pXRF + Vis-NIR (WT) with RF, pXRF + Vis-NIR (WB) with C5.0, and pXRF + Vis-NIR (WB) with RF (89.23% overall accuracy and 0.80 Kappa index). For tropical soils, soil subgroup prediction using only B horizon data obtained by pXRF in tandem with RF algorithm may be a viable alternative to assist in soil classification, especially when the acquisition of Vis-NIR is not possible. [ABSTRACT FROM AUTHOR]

Knowledge about parent material (PM) is crucial to understand the properties of overlying soils. Assessing PM of very deep soils, however, is not easy. Previous studies have predicted PM via proximal sensors and machine learning algorithms, but within small areas and with low soil variety. This study evaluates the efficiency of using portable X-ray fluorescence (pXRF) spectrometry together with machine learning algorithms in a large area populated with a wide variety of soil classes and land uses to build accurate PM distribution maps from soil data. Samples from A and B horizons were collected from 117 sites, totaling 234 samples, along with representative PM samples of the predominant PMs in the region. Elemental contents of PM and soil samples were obtained using pXRF. Random Forest (RF), Support Vector Machine (SVM) and Linear Discriminant Analysis (LDA) models built with and without Principal Component Analysis (PCA) were used to predict PMs from A and B horizon samples separately. For validation, PM was identified in 23 different spots and compared with the predicted PM via overall accuracy and Kappa coefficient. Maps built with models excluding PCA had an overall accuracy ranging from 0.87 to 0.96 and kappa coefficient ranging from 0.74 to 0.91. Maps generated with PCA based models reached 100% overall accuracy. Prediction of PM using pXRF with samples from either A or B horizon and machine learning algorithms offers solid results even when applied in large areas with high land use and soil class variability. • Soil parent material can be accurately predicted via pXRF analysis of soils. • Maps from B horizon samples were slightly superior to A horizon samples. • Overall map predictive accuracy was strong. [ABSTRACT FROM AUTHOR]