Modeling soil moisture from in situ portable X-ray spectrometer measurements: a novel approach for correcting geochemical data across different environments and climatic conditions

Journal Pre-proof Modeling soil moisture from in situ portable X-ray spectrometer measurements: a novel approach for correcting geochemical data across different environments and climatic conditions Thomas Vincent Gloaguen, Amélia Paula Marinho Reis, Magali Philippe, Gaël Le Roux PII: S0883-2927(24)00171-9 DOI: https://doi.org/10.1016/j.apgeochem.2024.106066 Reference: AG 106066 To appear in: Applied Geochemistry Received Date: 21 July 2023 Revised Date: 21 May 2024 Accepted Date: 6 June 2024 Please cite this article as: Gloaguen, T.V., Marinho Reis, A.P., Philippe, M., Le Roux, G., Modeling soil moisture from in situ portable X-ray spectrometer measurements: a novel approach for correcting geochemical data across different environments and climatic conditions, Applied Geochemistry, https:// doi.org/10.1016/j.apgeochem.2024.106066. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2024 Published by Elsevier Ltd. Journal Pre-proof 1 Modeling soil moisture from in situ portable X-ray spectrometer 1 measurements: a novel approach for correcting geochemical data across 2 different environments and climatic conditions 3 4 Abstract 5 The portable X-ray fluorescence (pXRF) spectrometer is widely employed for in situ 6 analysis of both contaminated and uncontaminated soils. However, the accuracy of the 7 measurements can be significantly affected by soil moisture, resulting in unreliable soil 8 pollution monitoring. This effect has already been studied and quantified, but this is 9 ineffective if the soil moisture in the field is unknown. Given the considerable variability of 10 soil moisture conditions across time and space, significant bias during in situ investigations 11 remains a main issue. This study introduces a novel method to estimate soil moisture 12 directly from pXRF field measurements, enabling its reliable use in almost any field 13 condition. The study was conducted using soil samples and in situ pXRF soil surface 14 measurements in Estarreja (Portugal) and Vicdessos (France). In the first experiment, the 15 innovative approach involved modeling soil moisture directly from the raw XRF 16 measurement errors obtained in moist soils, using multiple regression. In the second 17 experiment, metal concentrations were modeled as an exponential function of the moisture 18 content. The final model integrates both approaches to correct field data from geochemical 19 mapping in diverse environments, including a coastal region in Portugal and a mountainous 20 region in France. Our findings demonstrate that this simple, efficient and cost-effective 21 method accurately predicts soil moisture (U) using pXRF, as shown by the equation Umeasured 22 = 1.0028 x Uestimated (r2 = 0.9715). The model effectively corrected up to 70% of moisture23 induced errors in metal concentrations in the wettest soils and produced more reliable soil 24 Fe, Pb, and Zn maps. Specifically, the accuracy improvement was at least 32% in drier soils 25 Journal Pre-proof 2 (Portugal) and at least 55% in wetter soils (France). This study offers a cost-effective, 26 efficient solution for employing pXRF in geochemical mapping across different climatic 27 conditions and soil environments. 28 29 Keywords 30 pXRF; soil pollution; multiple regression model; geochemical mapping 31 32 1. Introduction 33 Over the past decade, the portable X-ray fluorescence spectrometer (pXRF) has emerged as 34 an important instrument for assessing soil contamination (Borges et al., 2020; Caporale et al., 35 2018; Kallithrakas-Kontos et al., 2016; Parsons et al., 2013; Ravansari et al., 2020; Rouillon 36 and Taylor, 2016). Recently, it has been integrated with other techniques such as Vis-NIR or 37 gamma-ray spectroscopy to improve the prediction of soil attributes (Li et al., 2021; Nawar et 38 al., 2022; Qingya et al., 2022). Some of the advantages of pXRF include its efficiency (analysis 39 in seconds to minutes), reliability, and versatility in analyzing various materials (rocks, soils, 40 organics, metals). Apart from evaluating soil contamination, pXRF finds application in soil 41 geochemistry and mapping (Benedet et al., 2020; Lemière, 2018; O’Rourke et al., 2016a; 42 Stockmann et al., 2016a; Weindorf et al., 2014, 2012; Young et al., 2016). 43 Nevertheless, the primary challenge lies in the measurement uncertainty due to variable 44 field conditions. Some authors have advised against underestimating metal concentrations using 45 pXRF data and recommend additional soil sampling and geostatistical simulation (Horta et al., 46 2021; Qu et al., 2022), or other complementary spectroscopy analyses (Li et al., 2021; Shrestha 47 et al., 2022) for more accurate analysis. To mitigate the effects of field conditions during pXRF 48 measurements, several precautions have been reported in the literature, including slight 49 compaction of the soil, removal of organic matter from the surface, and control of soil moisture 50 Journal Pre-proof 3 (Sharma et al., 2014; Weindorf et al., 2012; Zhu et al., 2011). Compacting soil and removing 51 coarse material is relatively straightforward, while measuring soil moisture can be time52 consuming, depending on the method employed. This contrasts with the fundamental principle 53 of XRF, which is designed to quick and practical analyses. 54 In environmental studies, the impact of soil moisture on XRF measurements is a well55 documented concern (Bastos et al., 2012; Kalnicky and Singhvi, 2001; Laiho and Perämäki, 56 2005; Padilla et al., 2019; Schneider et al., 2016; Stockmann et al., 2016c). Water molecules 57 scatter and absorb primary X-rays, which reduces the signal intensity, particularly in clayey 58 soils with a high Fe content (Ge et al., 2005; Stockmann et al., 2016b). Although some studies 59 have addressed the inaccuracies in data resulting from soil moisture (Akopyan et al., 2018; 60 Argyraki et al., 1997; Bastos et al., 2012; De La Calle et al., 2013; Parsons et al., 2013), most 61 in situ geochemical maps do not include a correction for soil moisture. The USEPA Method 62 6200 suggests that soil moisture content should ideally be below 20% to mitigate the impact on 63 XRF measurements (US Environmental Protection Agency, 2007). However, achieving this 64 condition in the field can be challenging due to regional, climatic, and seasonal variations. 65 Moreover, local variations in moisture across different soil sampling sites can result in 66 unreliable field data. Some researchers have proposed correcting XRF geochemical data in 67 hydromorphic wetland soils by correlating them with laboratory wavelength dispersive X-ray 68 fluorescence (Borges et al., 2020). Alternatively, soil moisture can be measured in the 69 laboratory for post-processing correction, although these methods are time-intensive. 70 Instruments like neutron probes, electrical conductivity-based sensors, or specific moisture 71 probes (Argyraki et al., 1997) are ocasionnaly employed for measuring soil moisture, but they 72 increase study costs and complicate and slow down the in situ XRF analysis. 73 To address these challenges, this study introduces a novel approach for modeling soil 74 moisture directly from the raw field XRF measurement errors. The objective of this method is 75 Journal Pre-proof 4 to systematically correct measurements at each sampling site, regardless of soil moisture 76 content. This correction method was applied to in situ geochemical mapping in two different 77 environments: a mountainous region of the French Pyrenees in more humid conditions, and a 78 coastal region of Portugal in drier conditions. 79 80 2. Methodology 81 2.1. Description of the study area, soil sampling, and mapping 82 The methodology flow chart is depicted in Figure 1. A soil sampling campaign with in situ 83 XRF measurements was conducted at two sites of the French Centre National de la Recherche 84 Scientifique (CNRS), known as Observatoires Homme-Milieu (OHM). 85 The OHM of Estarreja is situated near the city of Aveiro, Portugal. This area includes the 86 Estarreja Eco Park, one of the Portugal´s largest industrial facilities. Given the presence of 87 numerous plastic factories, metal equipment factories, and chemical plants, this area is 88 significantly contaminated (Barradas et al., 1992; Costa and Jesus-Rydin, 2001; Inácio et al., 89 2014, 1998; Marinho‐reis et al., 2020; Plumejeaud et al., 2018). Geochemical mapping of the 90 entire municipality of Estarreja (108.17 km2) was performed with a ThermoFisher handheld 91 field X-ray fluorescence analyzer (Niton XL3t – details provided below), on a regular 750 × 92 750 m grid, with 140 sample sites (1.8 samples/km2). For each site, the surface soil was 93 analyzed at three sub-sites within a 5 m radius area (each value represents an average of three 94 values). As described in the USEPA Method 6200 (US Environmental Protection Agency, 95 2007), coarse materials such as leaves, grass, stones, roots, etc., were removed, and the soil was 96 slightly compacted to ensure an adequate contact between the soil and the instrument. The 97 analyzing duration for each sample was 120 seconds. 98 The second site is the OHM at Vicdessos, located in the French Pyrenees mountain range. 99 The site has a history of contamination due to centuries of lead, zinc, and arsenic ores 100 Journal Pre-proof 5 exploration (Hansson et al., 2019, 2017; Simonneau et al., 2013). The study focused on valley 101 soils to investigate how runoff, atmospheric deposition, and urban activities contribute to soil 102 contamination. Soil sampling was conducted at 48 sites within this area, resulting in a density 103 of 8 samples/km2. This higher density compared to the sampling density in the OHM in 104 Estarreja is attributed to the greater geodiversity of the region. Additionally, 47 soil samples 105 were collected in the two adjacent Suc-et-Sentenac and Auzat valleys. The sampling and 106 analysis procedures were similar to those used in the OHM of Estarreja. 107 Regarding the technical characteristics of the ThermoFisher Niton XL3t, its analytical 108 capacity encompasses elements ranging from S to U. It features a small 3 mm sample area, and 109 is equipped with a gold (Au) x-ray tube capable of reaching 50 kV. The system incorporates 110 advanced semiconductor detectors and weighs about 1.3 kilograms. 111 To ensure the analytical quality of all field and laboratory measurements, nine certified 112 reference materials (BCR141-R, BCR142R, BCR145-R, IAEA-SL1, LKSD-3, RTH912, 113 STSD-3, SUD-1, and WQB1) were used, for obtaining precision (reproducibility of 114 measurements), limits of detection, generating calibration curves, and correcting the dataset. 115 The performance of the instrument is detailed in Table S1 (Supplementary material). 116 Geochemical maps were generated from field XRF data by analyzing semivariograms for 117 autocorrelation, spatial dependence, and isotropy. The interpolation method used was ordinary 118 kriging, which is more suitable for environmental studies (Goovaerts, 1999). Statistical 119 description, geostatistical analysis, data manipulation, and map production were performed 120 using SAGA GIS 9.3 and QGIS 3.28. 121 122 2.2. Correction of XRF data from the modeled soil moisture 123 Journal Pre-proof 6 To develop a method for estimating soil moisture directly from XRF data, two laboratory 124 experiments were conducted. Subsequently, the method for moisture correction was applied to 125 field data. 126 127 2.2.1. Experiment 1: Modeling soil moisture from XRF measurement errors 128 Twenty percent of the 140 sites (28 samples) were collected during the geochemical 129 mapping survey in Estarreja. Each sample consisted of three sub-samples collected within a 130 radius of five meters. Before sampling a different site, the equipment was thoroughly cleaned 131 with Milli-Q water. Based on particle size analysis of the 28 soil samples in triplicates (Horiba 132 LA950-V2 laser particle analyzer, Table S2, Supplementary material), five soils were selected, 133 ranging from sand to silt texture: sand, sandy loam, silt loam and silt. Soil texture is an important 134 factor influencing infiltration and moisture retention. The five soils were chosen based on their 135 sand content: 10.6% (representing 0-20%), 33.3% (20-40%), 49.0% (40-60%), 70.6% (60136 80%), and 87.6% (80-100%). Meanwhile, the clay content in the studied region varied 137 minimally. A chemical analysis of the soil samples is provided in Table S3 (Supplementary 138 material). After air drying at 20°C to 30°C in an isolated room, the soil samples were quartered 139 and sieved (< 2 mm). Triplicate soil samples of defined mass were prepared in vials for XRF 140 analysis (soil height in vial = 7 mm). Milli-Q ultrapure water was meticulously added to the soil 141 until saturation was achieved, and the samples were sealed for overnight equilibration. The 142 following day, the process of soil drying started: after drying at 35 °C for 60 min, the samples 143 were analyzed with the pXRF spectrometer and weighed for calculation of soil moisture. The 144 process was repeated five times. In order to accelerate the evaporation process, the subsequent 145 seven measurements were taken after drying at 60°C for 30 min, and the final two measurements 146 were taken after drying at 105 °C for 10 min (see evolution of the soil moisture during the 147 experiment in Supplementary material, Fig. S1). 148 Journal Pre-proof 7 The innovative approach for modeling soil moisture from XRF data involved considering 149 that the measurement errors from the pXRF spectrometer are substantially affected by soil 150 moisture. Soil moisture was modeled with 70 measurement errors: 5 soil samples at 14 drying 151 stages. The chemical elements were selected based on the correlation between their 152 concentration and soil moisture, and the mathematical model employed was a multiple linear 153 regression (Equation 1). 154 155 𝑈𝑒𝑠𝑡𝑖𝑚𝑎𝑡𝑒𝑑 =∑𝛽𝑖. 𝑒𝑟𝑟𝑜𝑟𝑖 𝑛 𝑖= 1 Equation 1 156 Where 157 errori is the measurement error of the pXRF spectrometer for the chemical element i 158 βi is the regression coefficient for element i 159 160 A separated soil samples dataset was exclusively used to validate the accuracy of the model. 161 In conjunction with the XRF field mapping directly on the soil surface, as described previously 162 (section 2.1), soil samples were collected at over 50% of the sites in Estarreja, precisely at the 163 location of the XRF measurement of the soil surface at sub-site R1 (0.5 cm depth, 164 approximately 50 g, n=71). The objective of this secondary sampling was to obtain precise 165 measurements of soil moisture and XRF values in the laboratory. Samples were sealed in Falcon 166 tubes with Teflon tape to prevent moisture loss during storage and transport. After XRF 167 measurements in the laboratory (3 triplicates for each soil sample), the samples were promptly 168 dried at 105°C to measure the exact soil moisture. A theoretical estimated soil moisture was 169 calculated based on the XRF results using Equation 1, after which the measured and the XRF170 based estimated moistures were compared. 171 172 Journal Pre-proof 14 4. Conclusion 315 The limitations imposed by soil moisture on in situ XRF analysis can be overcome through 316 the utilization of a predictive model constructed in two steps, enabling the estimation of of soil 317 moisture directly from XRF measurements with a high accuracy of 98%. This innovative 318 method offers a reliable solution for successful in situ XRF analysis and represents a significant 319 advancement in environmental research. The two-step process involves first estimating soil 320 moisture from in situ XRF measurement errors, followed by the correction of XRF 321 measurements for moisture. Through this post-processing method, the accuracy of geochemical 322 field maps is considerably enhanced, with differences in metal concentration before and after 323 correction exceeding 50%. It facilitates the extensive utilization of portable XRF instruments 324 in various environments, delivering precise and comparable results across different ecosystems, 325 climatic conditions and collection times. In summary, our approach offers a cost-effective and 326 efficient solution to mitigate the impact of soil moisture on in situ XRF analysis, thereby holding 327 consirable potential for advancing environmental soil research. 328 329 Acknowledgments 330 The project has been funded by the CNRS TRAM Project (ANR-15-CE01-0008) and 331 Observatoire Homme-Milieu Pyrénées Haut Vicdessos - LABEX DRIIHM ANR-11332 LABX0010. The research was also funded by FCT (Fundação para a Ciência e a Tecnologia, 333 Portugal) through projects UIDB/04683/2020, UIDP/04683/2020 (Institute of Earth Sciences, 334 pole of University of Minho). We extend our sincere gratitude to the scientific teams at the 335 Laboratoire Ecologie Fonctionnelle et Environnement (ECOLAB - UMR 5245 CNRS-UT3336 INPT) and of the Laboratoire Geographie de l´Environnement (GEODE - UMR 5602 CNRS337 UT2J) in Toulouse, France, for their invaluable assistance, both analytically and financially. 338 Journal Pre-proof 15 We are also acknowledge the Universidade Federal do Recôncavo da Bahia, Brazil, for 339 providing salary support to the first author during one year for research. 340 341 References 342 343 Akopyan, K., Petrosyan, V., Grigoryan, R., Melkomian, D.M., 2018. Assessment of residential 344 soil contamination with arsenic and lead in mining and smelting towns of northern 345 Armenia. J. Geochemical Explor. 184, 97–109. 346 https://doi.org/10.1016/j.gexplo.2017.10.010 347 Argyraki, A., Ramsey, M.H., Potts, P.J., 1997. Evaluation of Portable X-ray Fluorescence 348 Instrumentation for in situ Measurements of Lead on Contaminated Land. Analyst 122, 349 743–749. https://doi.org/10.1039/A700746I 350 Barradas, J.M., Cardoso Fonseca, E., Ferreira da Silva, E., Garcia Pereira, H., 1992. 351 Identification and mapping of pollution indices using a multivariate statistical 352 methodology, Estarreja, central Portugal. Appl. Geochemistry 7, 563–572. 353 https://doi.org/10.1016/0883-2927(92)90071-A 354 Bastos, R.O., Melquiades, F.L., Biasi, G.E. V, 2012. Correction for the effect of soil moisture 355 on in situ XRF analysis using low-energy background. X-Ray Spectrom. 41, 304–307. 356 https://doi.org/10.1002/xrs.2397 357 Benedet, L., Faria, W.M., Silva, S.H.G., Mancini, M., Demattê, J.A.M., Guilherme, L.R.G., 358 Curi, N., 2020. Soil texture prediction using portable X-ray fluorescence spectrometry and 359 visible near-infrared diffuse reflectance spectroscopy. Geoderma 376. 360 https://doi.org/10.1016/j.geoderma.2020.114553 361 Borges, C.S., Weindorf, D.C., Nascimento, D.C., Curi, N., Guilherme, L.R.G., Carvalho, G.S., 362 Ribeiro, B.T., 2020. Comparison of portable X-ray fluorescence spectrometry and 363 laboratory-based methods to assess the soil elemental composition: Applications for 364 wetland soils. Environ. Technol. Innov. 19, 100826. 365 https://doi.org/10.1016/j.eti.2020.100826 366 Caporale, A.G., Adamo, P., Capozzi, F., Langella, G., Terribile, F., Vingiani, S., 2018. 367 Monitoring metal pollution in soils using portable-XRF and conventional laboratory-based 368 techniques: Evaluation of the performance and limitations according to metal properties 369 and sources. Sci. Total Environ. 643, 516–526. 370 https://doi.org/10.1016/j.scitotenv.2018.06.178 371 Costa, C., Jesus-Rydin, C., 2001. Site investigation on heavy metals contaminated ground in 372 Estarreja - Portugal. Eng. Geol. 60, 39–47. https://doi.org/10.1016/S0013-7952(00)00087373 9 374 De La Calle, I., Cabaleiro, N., Romero, V., Lavilla, I., Bendicho, C., 2013. Sample pretreatment 375 strategies for total reflection X-ray fluorescence analysis: A tutorial review. Spectrochim. 376 Acta Part B At. Spectrosc. 90, 23–54. https://doi.org/10.1016/J.SAB.2013.10.001 377 Ge, L., Lai, W., Lin, Y., 2005. Influence of and correction for moisture in rocks, soils and 378 sediments on in situ XRF analysis. X-Ray Spectrom. 34, 28–34. 379 Journal Pre-proof 16 https://doi.org/10.1002/xrs.782 380 Goovaerts, P., 1999. Geostatistics in soil science: State-of-the-art and perspectives. Geoderma 381 89, 1–45. https://doi.org/10.1016/S0016-7061(98)00078-0 382 Hansson, S. V, Claustres, A., Probst, A., De Vleeschouwer, F., Baron, S., Galop, D., Mazier, 383 F., Le Roux, G., 2017. Atmospheric and terrigenous metal accumulation over 3000 years 384 in a French mountain catchment: Local vs distal influences. Anthropocene 19, 45–54. 385 https://doi.org/10.1016/j.ancene.2017.09.002 386 Hansson, S. V, Grusson, Y., Chimienti, M., Claustres, A., Jean, S., Le Roux, G., 2019. Legacy 387 Pb pollution in the contemporary environment and its potential bioavailability in three 388 mountain catchments. Sci. Total Environ. 671, 1227–1236. 389 https://doi.org/https://doi.org/10.1016/j.scitotenv.2019.03.403 390 Horta, A., Azevedo, L., Neves, J., Soares, A., Pozza, L., 2021. Integrating portable X-ray 391 fluorescence (pXRF) measurement uncertainty for accurate soil contamination mapping. 392 Geoderma 382, 114712. https://doi.org/10.1016/j.geoderma.2020.114712 393 Hubbell, J.H., Seltzer, S.M., 2009. X-Ray Mass Attenuation Coefficients | NIST [WWW 394 Document]. URL https://www.nist.gov/pml/x-ray-mass-attenuation-coefficients 395 (accessed 9.22.18). 396 Inácio, M., Neves, O., Pereira, V., Ferreira da Silva, E., 2014. Levels of selected potential 397 harmful elements (PHEs) in soils and vegetables used in diet of the population living in 398 the surroundings of the Estarreja Chemical Complex (Portugal). Appl. Geochemistry 44, 399 38–44. https://doi.org/10.1016/j.apgeochem.2013.07.017 400 Inácio, M.M., Pereira, V., Pinto, M.S., 1998. Mercury contamination in sandy soils surrounding 401 an industrial emission source (Estarreja, Portugal). Geoderma 85, 325–339. 402 https://doi.org/10.1016/S0016-7061(98)00027-5 403 Kallithrakas-Kontos, N., Foteinis, S., Paigniotaki, K., Papadogiannakis, M., 2016. A robust X404 ray fluorescence technique for multielemental analysis of solid samples. Environ. Monit. 405 Assess. 188, 1–10. https://doi.org/10.1007/s10661-016-5127-4 406 Kalnicky, D.J., Singhvi, R., 2001. Field portable XRF analysis of environmental samples. J. 407 Hazard. Mater. 83, 93–122. https://doi.org/10.1016/S0304-3894(00)00330-7 408 Laiho, J.V.-P., Perämäki, P., 2005. Evaluation of portable X-ray fluorescence (PXRF) - Sample 409 preparation methods. Geol. Surv. Finl. 38, 73–82. 410 Lemière, B., 2018. A review of pXRF ( fi eld portable X-ray fl uorescence ) applications for 411 applied geochemistry. J. Geochemical Explor. 188, 350–363. 412 https://doi.org/10.1016/j.gexplo.2018.02.006 413 Li, F., Xu, L., You, T., Lu, A., 2021. Measurement of potentially toxic elements in the soil 414 through NIR, MIR, and XRF spectral data fusion. Comput. Electron. Agric. 187, 106257. 415 https://doi.org/https://doi.org/10.1016/j.compag.2021.106257 416 Marinho‐reis, A.P., Costa, C., Rocha, F., Cave, M., Wragg, J., Valente, T., Sequeira‐braga, A., 417 Noack, Y., 2020. Biogeochemistry of household dust samples collected from private 418 homes of a portuguese industrial city. Geosci. 10, 1–20. 419 https://doi.org/10.3390/geosciences10100392 420 Mejía-Piña, K.G., Huerta-Diaz, M.A., González-Yajimovich, O., 2016. Calibration of handheld 421 Journal Pre-proof 17 X-ray fluorescence (XRF) equipment for optimum determination of elemental 422 concentrations in sediment samples. Talanta 161, 359–367. 423 https://doi.org/10.1016/j.talanta.2016.08.066 424 Nawar, S., Richard, F., Kassim, A.M., Tekin, Y., Mouazen, A.M., 2022. Fusion of Gamma425 rays and portable X-ray fluorescence spectral data to measure extractable potassium in 426 soils. Soil Tillage Res. 223, 105472. 427 https://doi.org/https://doi.org/10.1016/j.still.2022.105472 428 O’Rourke, S.M., Stockmann, U., Holden, N.M., McBratney, A.B., Minasny, B., 2016a. An 429 assessment of model averaging to improve predictive power of portable vis-NIR and XRF 430 for the determination of agronomic soil properties. Geoderma 279, 31–44. 431 https://doi.org/10.1016/J.GEODERMA.2016.05.005 432 O’Rourke, S.M., Stockmann, U., Holden, N.M., McBratney, A.B., Minasny, B., 2016b. An 433 assessment of model averaging to improve predictive power of portable vis-NIR and XRF 434 for the determination of agronomic soil properties. Geoderma 279, 31–44. 435 https://doi.org/10.1016/j.geoderma.2016.05.005 436 Padilla, J.T., Hormes, J., Magdi Selim, H., 2019. Use of portable XRF: Effect of thickness and 437 antecedent moisture of soils on measured concentration of trace elements. Geoderma 337, 438 143–149. https://doi.org/10.1016/j.geoderma.2018.09.022 439 Parsons, C., Grabulosa, E.M., Pili, E., Floor, G.H., Roman-Ross, G., Charlet, L., 2013. 440 Quantification of trace arsenic in soils by field-portable X-ray fluorescence spectrometry: 441 Considerations for sample preparation and measurement conditions. J. Hazard. Mater. 262, 442 1213–1222. https://doi.org/10.1016/j.jhazmat.2012.07.001 443 Plumejeaud, S., Reis, A.P., Tassistro, V., Patinha, C., Noack, Y., Orsière, T., 2018. Potentially 444 harmful elements in house dust from Estarreja, Portugal: characterization and genotoxicity 445 of the bioaccessible fraction. Environ. Geochem. Health 40, 127–144. 446 https://doi.org/10.1007/s10653-016-9888-z 447 Potts, P.J., West, M., 2008. Portable x-ray fluorescence spectrometry : capabilities for in situ 448 analysis, Royal Soci. ed. Cambridge. 449 Qingya, W., Li, F., Jiang, X., Hao, J., Zhao, Y., Wu, S., Cai, Y., Huang, W., 2022. Quantitative 450 analysis of soil cadmium content based on the fusion of XRF and Vis-NIR data. Chemom. 451 Intell. Lab. Syst. 226, 104578. 452 https://doi.org/https://doi.org/10.1016/j.chemolab.2022.104578 453 Qu, M., Guang, X., Liu, H., Zhao, Y., Huang, B., 2022. Additional sampling using in-situ 454 portable X-ray fluorescence (PXRF) for rapid and high-precision investigation of soil 455 heavy metals at a regional scale. Environ. Pollut. 292, 118324. 456 https://doi.org/https://doi.org/10.1016/j.envpol.2021.118324 457 Ravansari, R., Lemke, L.D., 2018. Portable X-ray fluorescence trace metal measurement in 458 organic rich soils: pXRF response as a function of organic matter fraction. Geoderma 319, 459 175–184. https://doi.org/10.1016/j.geoderma.2018.01.011 460 Ravansari, R., Wilson, S.C., Tighe, M., 2020. Portable X-ray fluorescence for environmental 461 assessment of soils: Not just a point and shoot method. Environ. Int. 134, 105250. 462 https://doi.org/10.1016/j.envint.2019.105250 463 Rouillon, M., Taylor, M.P., 2016. Can field portable X-ray fluorescence (pXRF) produce high 464 Journal Pre-proof 18 quality data for application in environmental contamination research? Environ. Pollut. 214, 465 255–264. https://doi.org/10.1016/J.ENVPOL.2016.03.055 466 Schneider, A.R., Cancès, B., Breton, C., Ponthieu, M., Morvan, X., Conreux, A., Marin, B., 467 2016. Comparison of field portable XRF and aqua regia/ICPAES soil analysis and 468 evaluation of soil moisture influence on FPXRF results. J. Soils Sediments. 469 https://doi.org/10.1007/s11368-015-1252-x 470 Shand, C.A., Wendler, R., 2014. Portable X-ray fluorescence analysis of mineral and organic 471 soils and the influence of organic matter. J. Geochemical Explor. 143, 31–42. 472 https://doi.org/10.1016/j.gexplo.2014.03.005 473 Sharma, A., Weindorf, D.C., Man, T., Abdalsatar, A., Aldabaa, A., Chakraborty, S., 2014. 474 Characterizing soils via portable X-ray fluorescence spectrometer: 3. Soil reaction (pH). 475 Geoderma 232–234, 141–147. https://doi.org/10.1016/j.geoderma.2014.05.005 476 Sharma, A., Weindorf, D.C., Wang, D., Chakraborty, S., 2015. Characterizing soils via portable 477 X-ray fluorescence spectrometer: 4. Cation exchange capacity (CEC). Geoderma 239–240, 478 130–134. https://doi.org/10.1016/j.geoderma.2014.10.001 479 Shrestha, G., Calvelo-Pereira, R., Roudier, P., Martin, A.P., Turnbull, R.E., Kereszturi, G., 480 Jeyakumar, P., Anderson, C.W.N., 2022. Quantification of multiple soil trace elements by 481 combining portable X-ray fluorescence and reflectance spectroscopy. Geoderma 409, 482 115649. https://doi.org/https://doi.org/10.1016/j.geoderma.2021.115649 483 Simonneau, A., Chapron, E., Courp, T., Tachikawa, K., Le Roux, G., Baron, S., Galop, D., 484 Garcia, M., Di Giovanni, C., Motellica-Heino, M., Mazier, F., Foucher, A., Houet, T., 485 Desmet, M., Bard, E., 2013. Recent climatic and anthropogenic imprints on lacustrine 486 systems in the Pyrenean Mountains inferred from minerogenic and organic clastic supply 487 (Vicdessos valley, Pyrenees, France). Holocene 23, 1764–1777. 488 https://doi.org/10.1177/0959683613505340 489 Stockmann, U., Cattle, S.R., Minasny, B., McBratney, A.B., 2016a. Utilizing portable X-ray 490 fluorescence spectrometry for in-field investigation of pedogenesis. Catena 139, 220–231. 491 https://doi.org/10.1016/j.catena.2016.01.007 492 Stockmann, U., Jang, H.H., Minasny, B., McBratney, A.B., 2016b. The Effect of Soil Moisture 493 and Texture on Fe Concentration Using Portable X-Ray Fluorescence Spectrometers, in: 494 Digital Soil Morphometrics. pp. 63–71. https://doi.org/10.1007/978-3-319-28295-4 495 Stockmann, U., Jang, H.J., Minasny, B., McBratney, A.B., 2016c. The effect of soil moisture 496 and texture on Fe concentration using portable X-Ray fluorescence spectrometers, in: 497 Digital Soil Morphometrics. pp. 63–71. 498 US Environmental Protection Agency, 2007. Method 6200 : Field portable X-ray fluorescence 499 spectrometryfor the determination of elemental concentrations in soil and sediment, Test 500 Methods For Evaluating Solid Waste, US Environmental Protection Agency. 501 https://doi.org/10.1017/CBO9781107415324.004 502 Weindorf, D.C., Bakr, N., Zhu, Y., 2014. Advances in portable X-ray fluorescence (PXRF) for 503 environmental, pedological, and agronomic applications. Adv. Agron. 128. 504 https://doi.org/10.1016/B978-0-12-802139-2.00001-9 505 Weindorf, D.C., Zhu, Y., Mcdaniel, P., Valerio, M., Lynn, L., Michaelson, G., Clark, M., Ping, 506 C.L., 2012. Characterizing soils via portable x-ray fluorescence spectrometer: 2. Spodic 507 Journal Pre-proof 19 and Albic horizons. Geoderma 189–190, 268–277. 508 https://doi.org/10.1016/j.geoderma.2012.06.034 509 Young, K.E., Evans, C.A., Hodges, K. V., Bleacher, J.E., Graff, T.G., 2016. A review of the 510 handheld X-ray fluorescence spectrometer as a tool for field geologic investigations on 511 Earth and in planetary surface exploration. Appl. Geochemistry 72, 77–87. 512 https://doi.org/10.1016/j.apgeochem.2016.07.003 513 Zhu, Y., Weindorf, D.C., Zhang, W., 2011. Characterizing soils using a portable X-ray 514 fluorescence spectrometer: 1. Soil texture. Geoderma 167–177. 515 https://doi.org/10.1016/j.geoderma.2011.08.010 516 Journal Pre-proof Table 1 Statistical summary of the multiple regression model for estimating soil moisture from measurement errors obtained during X-ray fluorescence (XRF) spectrometry analysis. Analysis of variance Degree of freedom Sum Squares Mean Squares F Critical value F Regression 15 3.3317 0.2220 159.37 5.94E-50 Residus 74 0.1034 0.0013 Total 89 3.4342 Regression statistics Multiple coefficient of determination 0.9848 Coefficient of determination 0.9699 Error 0.0373 Observations 75 Coefficient of the regression Constant 0.0939 Cr Error 0.0935 As Error 0.1291 Zn Error 0.0392 V Error 0.0439 S Error 0.0005 Ti Error -0.0002 K Error -0.0055 Co Error -0.0294 Fe Error 0.0099 Sc Error 0.0084 Ca Error -0.0053 Pb Error -0.1392 Rb Error 0.1435 Zr Error 0.0780 Sr Error -0.3694 Table 2 Mean concentrations of Fe, Zn and Pb in soil samples before and after soil moisture correction, in the in situ entire XRF datasets at the OHM in Estarreja (Portugal) and the OHM in Vicdessos (France). OHM Estarreja OHM Vicdessos Number of samples 140 95 Mean moisture (± SD) 27 % (± 20%) 53 % (± 18%) Mean [Fe] before correction (mg kg-1) 11,200 20,600 Mean [Fe] after correction (mg kg-1) 14,800 32,100 Difference due to moisture effect (%) 32.1 55.8 Mean [Zn] before correction (mg kg-1) 61 92 Mean [Zn] after correction (mg kg-1) 79 142 Difference due to moisture effect (%) 29.5 54.3 Mean [Pb] before correction (mg kg-1) 28 18 Mean [Pb] after correction (mg kg-1) 34 26 Difference due to moisture effect (%) 21.4 44.4 Journal Pre-proof Journal Pre-proof Fig. 1. Methodology flow chart for correcting geochemical XRF field data using soil moisture modeling. Journal Pre-proof Fig. 2. Evolution of XRF spectra during the drying of soil sample S2 from 84% to 0% gravimetric moisture in Experiment 1. Journal Pre-proof