Impact of the change in irrigation practices from untreated to treated wastewater on the mobility of potentially toxic elements (PTEs) in soil irrigated for decades

Long-term agricultural irrigation with untreated wastewater has resulted in metals and metalloids accumulation in soil. Little information is available on the consequences of a change to irrigation with treated water on the mobility of these potentially toxic elements (PTEs). The potential mobility of PTEs was assessed using sequential extractions performed on soil irrigated with untreated wastewater for a century in Mexico. The possible effects of change in irrigation practices on PTEs mobility was evaluated through batch experiments, simulating a decrease in pH, an increase in salinity, and in chlorine of the irrigation water. Geochemical modeling allowed predicting the speciation of mobilized PTEs. Soils irrigated with untreated water were mainly enriched with PTEs in surface horizons. Only Cd and As were found in the soluble or exchangeable fractions (< 20%). Cu and Pb were mainly associated with soil organic matter (OM), whereas As and Zn were bound to iron oxides, and Cr with refractory minerals. Batch experiments revealed that acidification resulted in the increased solubility of Cu, Zn, and Cd for surface samples, and As in deep horizons. In contrast, increased salinity only mobilized Zn, Cd, and Cr. Water chlorination mobilized higher amount of Zn, Pb, and Cd compared to the other experiments. As was not mobilized for these two experiments. A change in irrigation practices could increase the mobility of PTEs if water treatment is not adapted to the soil type. The mobilization of PTEs, especially As and Cd, could affect both crops and groundwater quality. It is essential to monitor this mobility to avoid future risks to human health.


Introduction
Metals and metalloids such as lead (Pb), cadmium (Cd), chromium (Cr), or arsenic (As) are potentially toxic elements (PTEs) which tend to accumulate in soils over time. Anthropogenic activities enhance fluxes of PTEs to soil primarily through irrigation and atmospheric deposition (Hu et al. 2018). PTEs are stored in soils mainly through chelation to soil organic matter (OM), or adsorption onto iron (Fe) and manganese (Mn) oxides, or clay minerals (Lair et al. 2007;Frierdich and Catalano 2012;Ugwu and Igbokwe 2019). Soils contaminated with PTEs can be a source of contamination for plants, surface, and groundwater, when geochemical conditions favor their release into the soil solution. They can then be absorbed by plants or leached into groundwater or surface Responsible editor: Fabio Scarciglia * Francisco Robert Alexander Ziegler Rivera ZieglerRivera90@gmail.com water bodies. Thereby, accessing the food chain would present great risks to human health (Edosa Otabor 2019). The reuse of wastewater for irrigation of agricultural field has been a common practice in many parts of the world (Asaad et al. 2022), as a source of plant nutrients (N and P) and OM, increasing soil fertility (Kiziloglu et al. 2008). However, wastewater can contain high loads of metals and metalloids (i.e., Pb, Cd, Cr, Hg, As), and organic contaminants depending on the industrial context of the region (Strady et al. 2016;Archundia et al. 2017a, b). In temperate regions, such as the surroundings of Berlin, London, and Paris, the flooding of soils with untreated wastewater has been used for several decades (Jaramillo and Restrepo 2017). These practices have however been stopped after a severe pollution by heavy metals (Abdelhafez et al. 2015). In semi-arid regions, irrigation with wastewater is still practiced because of the increasing demand for water resources resulting from the high population growth and lack of treatment facilities (Rezapour et al. 2011;Elgallal et al. 2016;Hussein et al. 2022). This often results in the spreading of infectious diseases and the pollution of water reservoirs (Leonel and Tonetti 2021;Mora et al. 2022). Water treatment prior to irrigation is thus promoted (Contreras et al. 2017) and consists of the reduction of suspended solids, OM, and nutrient load through primary and secondary treatments, and the elimination of potentially harmful microorganisms through tertiary treatment and disinfection such as chlorination or ozonation (Suslow 2000;Martínez et al. 2011).
An abrupt shift from untreated to treated wastewater (TWW) irrigation requires the adaptation of the management system as it can change abruptly the physical and chemical characteristics of soils (Klay et al. 2010;Morugán-Coronado et al. 2011). This management must consider the type of soil, the duration and volume of irrigation, the quality of the water, and the climate context (Tarchouna et al. 2010). Such a transition in irrigation practice might mainly affect soil (i) pH, (ii) salinity, and (iii) OM content and solubility (Jueschke et al. 2008;Morugán-Coronado et al. 2011;Gharaibeh et al. 2016;Shakir et al. 2017). Change in soil pH is critical in the mobility of trace metals and metalloids, since pH controls solubility, precipitation-dissolution reactions, and adsorption processes (Bourg and Loch 1995;Kabata-Pendias and Pendias 2001). In particular, soil acidification favors the mobility of metallic compounds, through desorption and competition for the absorption sites (Dijkstra et al. 2004;Sungur et al. 2014).
Changes in soil salinity can also increase the mobility of PTEs through the formation of soluble inorganic complexes, cation exchange, increased competition for available sites, and the dispersion of colloidal particles (Tarchouna et al. 2010). In addition, increased anions content such as chlorine (Cl − ) in the soil solution may also increase the metals solubility (e.g., Cd) through the formation of soluble complexes (Acosta et al. 2011;Caporale and Violante 2016). Finally, a change in soil pH and salinity can also affect OM solubility and favor the formation of soluble organo-metallic complexes (Greenland and Hayes 1981;Weng et al. 2002;Tan 2010). Such complexes could result in the increase in available PTEs for plants (Antoniadis and Alloway 2002;Haydee and Dalma 2017).
The Mezquital Valley is considered the most extensive site in the world (90,000 ha) irrigated with untreated wastewater (Guédron et al. 2014;Siebe et al. 2016). The continuous and prolonged use of wastewater (about 100 years) has caused the accumulation of potentially toxic elements (PTEs) in soils, the increase in soil organic matter. and salinity indicated by electrical conductivities of saturation extracts of up to 4 dS/m (Siebe 1998;Friedel et al. 2000). Soil PTEs (e.g., Zn, Cr, Cu, Pb, and Cd) are 3 to 6 times higher than regional background contents (Siebe 1994;Cajuste et al. 2001;Herre et al. 2004;Guédron et al. 2014;Siebe et al. 2016). Additionally, the release of protons by enhanced nitrification processes (Hernández et al. 2016) initiated a slight acidification (by about 1 pH unit) in these circumneutral to slightly alkaline soils (Chapela-Lara 2011; Siebe et al. 2016). The increased incidence of gastrointestinal diseases in the Mezquital Valley (Blumenthal et al. 2001) has motivated the construction of a wastewater treatment plant, which treats about 60% of the wastewater released in the valley since December 2017 (Drechsel et al. 2018). Our hypotheses are that acidification, the increase in salinity in the soil solution, and chlorination of the treated water will cause an increase in mobility of PTEs.
The objective of this study is to assess the effect of the change of untreated to treated wastewater on the mobility of PTEs in the Mezquital valley. The novelty of this work is related to the fact that mobility has not been studied in any agricultural site previously loaded with PTEs in complete soil profiles, where there is a change in water quality, and where increase in acidity and salinity has been reported. We here tested the effect of acidification, increased salinity, and chlorine content of the treated water, on the mobility of As, Cd, Cu, Cr, Pb, and Zn currently retained in the soils. This work was carried out using two complementary approaches to (i) determine the potential mobility of PTEs by sequential extractions in different soil profile samples, and (ii) assess the effect of different environmental scenarios on PTEs solubility using batch experiments with soil samples from an agricultural field irrigated for more than 90 years. In the batch experiments, the pH, ionic strength, and sodium hypochlorite doses of the supernatants were changed to assess their effect on PTEs mobility.

Study site
The Mezquital Valley is located in the southern part of the state of Hidalgo, 80 km north of Mexico City. The climate is semi-arid, with a mean annual temperature of 16-18 °C and mean annual rainfall of 400-600 mm, falling dominantly during summer (British Geological Survey 1998). Soils have developed from Quaternary alluvial and colluvial deposits, which cover late Tertiary volcanic tuff deposits. The 3 main types of soil in this area are Leptosols, Phaeozems, and Vertisols (Siebe 1994;Siebe et al. 2016). The valley is mainly used for agriculture (corn and alfalfa), and irrigation is performed via flooding Salazar et al. 2018). Since 1912, more than 90,000 ha have been irrigated with untreated wastewater (De la Cruz-Campa 1965;Siebe et al. 2016). In 2018, a large wastewater treatment plant was built, and most of the agricultural plots are irrigated with treated wastewater.
Water treatment is performed through biological activated sludge processing almost all year long, and physical-chemical processing, only after large rainstorms, when the wastewater is diluted. The treatment is designed to conserve plant macronutrients such as soluble nitrogen (N), phosphorus (P), and potassium (K), and to eliminate potential pathogens and meet the sanitary requirements set by the World Health Organization, namely less than 10 3 fecal coliforms and ≤ 1 helminth egg/liter of water (WHO 1989;CONAGUA 2010). The latter is achieved by chlorination.

Soil sampling and physico-chemical characterization
Soil sampling (vertic Phaeozem) was performed on a 2.25 ha plot located in Tlahuelilpan (Hidalgo), irrigated with untreated wastewater for 90 years with around 13 irrigation events per year (Hernández et al. 2016;Salazar et al. 2018). Surface soil samples (0-20 cm) were collected along a transect from the upstream (close to the wastewater inflow) to downstream part (close to the wastewater out, Fig. 1). Two soil profiles, located close to the water inlet (profile 1) and outlet (profile 2), were sampled by horizon (Fig. 1). A control sample was also collected in a soil only irrigated with rainwater (rainfed soil).
All samples were dried at room temperature for 72 h and sieved through a 2-mm mesh. The pH and electrical conductivity were measured in sieved samples with a Beckman potentiometer and a Lamotte conductometer after being shaken for 18 h with distilled water in a 1:2.5 ratio. Soil texture was determined in sieved samples using the Bouyoucos hydrometer method, where the samples were dispersed with sodium hexametaphosphate after organic matter destruction with H 2 O 2 .

Elemental analysis
All samples for elemental analysis were sieved (2 mm mesh) and ground (agate mortar) to obtain a fine powder. Total nitrogen (N) and total and organic carbon were determined using a CNHS/O Perkin Elmer 2400 elemental analyzer. The contents of chromium (Cr), copper (Cu), lead (Pb), and zinc (Zn) was determined by X-Ray Fluorescence (portable NITON XL3t Ultra). Samples were analyzed in triplicate, with a measurement time of 105 s per sample following the method 6200 (US-EPA 2007a).
Likewise, arsenic (As) and cadmium (Cd) contents (not determined by XRF in the soil samples due to its detection limits) were determined after acid digestion (HNO 3 and HCl, 3:1 v/v) using an Anton-PaarMultiwave 3000 microwave oven (USEPA 3051 A, US-EPA 2007b) and analyzed by inductively coupled plasma optical emission spectroscopy (ICP-OES) with a Perkin Elmer Optima 8300 DV equipment.

Sequential extractions
To assess the reactivity and soil carrier phases of PTEs, sequential extractions were performed in both soil profiles using the modified method of Zeien and Brümmer (1989). Eight extractions were performed to extract the water soluble (F1), mobile (F2), and exchangeable (F3) PTEs, and PTEs bound to Mn Oxides (F4), organic matter (F5), amorphous Fe oxides (F6), crystalline Fe oxides (F7), and the residual refractory minerals (F8). Details for the extraction protocol are given in supplementary material (Table SM1). Briefly, sequential extractions were performed at a 1:10 solid-to-liquid ratio, by adding 25 mL extractants to 2.5 g of soil sample in 50-mL polypropylene tubes. After each extraction step, samples were centrifuged (2500 rpm for 20 min) and the supernatant was filtered (0.45 µm) in 15-mL polypropylene tubes and stored at 4 °C until analysis. After each extraction step, tubes containing the soil extract were weighed for remaining moisture correction. The concentration of PTEs in solution in each of the fractions was determined by inductively coupled plasma optical emission spectrometry (ICP-OES, Perkin Elmer Optima 8300 DV with S10 Perkin Elmer autosampler).
To assess the environmental risk of PTEs, the mobility factor was calculated with the following equation (Narwal et al. 1999): where % MF is the mobility factor, F1 + F2 + F3 is the sum of the mass of the element associated with the fraction soluble in water (fraction 1), the mobile fraction (fraction 2: 1 M NH 4 NO 3 extraction), and the exchangeable fraction (fraction 3: 1 M NH 4 -acetate extraction).

Batch tests
To simulate the transition in irrigation practice from untreated to treated wastewater and asses its effect on the mobility of PTEs, batch experiments were conducted using specific soil horizons by changing the soil (i) pH, (ii) salinity, and (iii) chlorine content. Rainfed soils were used as a control for each tested scenario. In each 50 mL batch, 4 g of sieved and ground soil was placed in 40 mL solution (see below). All batches were duplicated and shaken at 120 rpm for 48 h on an orbital shaker (Shaker SK71) at room temperature. First, to assess the effect of change in pH, three pH scenarios were tested, starting from of 6.5 (i.e., the minimum value determined in both soils and irrigation water (Chapela-Lara 2011; Guédron et al. 2014), and then decreasing to 6.0 and 5.5. Soil pH was adjusted with diluted HCl (10%) addition. pH was measured four times in each batch (at 4, 12, 24, and 36 h) and HCl was added to maintain the pH. Second, the change in salinity was simulated by increasing soil EC from 1.5 dS/m (i.e., the maximum EC determined by Chapela-Lara (2011) in soil-water extracts of soils irrigated for 80 years) to 2.25 and 3.0 dS/m. The salinity increased was performed by Ca(NO 3 ) 2 additions to the soil solution. Finally, the effect of the chlorination of the waters was assessed using the same procedure as above using addition of NaClO at 1.0, 5.0, and 10 mg/L. It is worth mentioning that no chlorine concentration in treated water was available at the time of the experiment; hence, recommended doses applied in various wastewater treatment plants were used (Momba et al. 2008;Collivignarelli et al. 2018;Islami et al. 2019).
At the end of the experiment, all batches were centrifuged for 20 min at 2500 rpm in a HERMLE Z 513 centrifuge. The supernatant was filtered using 0.45 µm filters and then analyzed for PTEs by ICP-MS (Thermo Scientific iCAP Qc), dissolved organic carbon, and major ions (HCO 3 − , Cl − , NO 3 − , PO 4 3− , SO 4 2− , Ca 2+ , K + , Mg 2+ , and Na + ) by liquid chromatography, using a WATERS HPLC instrument.
Regarding sequential extraction, QA/QC was ensured by duplicating 20% of the samples. In order to verify the accuracy and validate the data of the sequential extraction method, we used a reference material (NIST 2710 A) under the same experimental conditions as the soil samples. The recovery rates of PTEs are shown in table SM3.
Finally, QA/QC of batch experiments was ensured by duplicating 20% of the samples and blank determinations.

Statistical analysis and geochemical modeling
Statistical analysis was performed using the R software. Prior to each analysis, the normal distribution of data was tested using Shapiro-Wilk's test. Pearson's correlation analysis was used to determine relationships between PTEs and the different parameters under investigation. Furthermore, a principal component analysis (PCA) was performed in order to distinguish groups between PTEs on all simulated scenarios. Analysis of variance (ANOVA test) was conducted to analyze the difference between the means of more than two groups. Tests were considered significant for a p value below 0.05.
Finally, to determine the theoretical elemental speciation in solution for each element, a geochemical modeling was carried out using the software Visual Minteq 3.1 (Gustafsson 2014). Input data used in the model were (i) concentration of PTEs in solution (mg/L), (ii) pH and EC, (iii) DOC (mg/L), and (iv) major cations and anions (mg/L).

Characterization of soil profiles and control sample
Four soil horizons were distinguished in profile 1 (P1), and five in profile 2 (P2) (Fig. 1). Granulometry was dominated by silt and clay in P1, and by clay in P2. Both profiles were alkaline with pH ranging between 7.5 and 8.3 (Table 1), typical of vertic Phaeozem. In both profiles, pH increased with depth. The electrical conductivity (EC) was higher in P1 (average = 0.55 ± 0.10 dS/m) without any obvious relation with depth, than in P2 (0.34 ± 0.04 dS/m) where EC decreased with depth.
Consistently, total carbon content decreased with depth in both profiles, and was mainly composed of organic carbon with very small amounts of inorganic carbon. Similarly, total nitrogen decreases with depth in both profiles, from 0.21 to 0.14% in P1 and from 0.26 to 0.14% in P2. The TC/TN ratio was higher in P1 than in P2 and decreases with depth in both profiles. TC/TN ranged between 6.4 and 11.6 for both soils underlying rapid organic matter (OM) mineralization and release of N.
The control soil presented similar pH value (alkaline), higher sand content, and lower EC, total carbon, and total nitrogen content compared to the most superficial samples of both profiles.

Content and distribution of PTEs in soils irrigated with untreated wastewater
The content of PTEs in superficial soil samples decreased with increasing distance from the water inlet ( Fig. 2A). This is consistent with observation of Siebe (1994), who reported higher accumulation of trace metals in soils located close to the irrigation water inflow in plots of the Mezquital Valley.
Consistently, higher contents in PTEs were found in profile 1 compared to profile 2 (Fig. 2B). In both profiles, PTEs contents decreased with depth for all elements except for As, which showed an increased in the Ah3 horizon (i.e., between 40 and 70 cm depth) in both profiles. Such in-depth increase in As between 40 and 70 cm coincides with the horizon with the best soil structure and higher permeability, namely medium to fine angular blocky, where As is probably displaced and retained during irrigation by percolation and subsurface lateral water flow. The decrease of trace metals with depth is consistent with previous studies in the Mezquital Valley which indicated larger metal accumulation in the till layer (0-30 cm) (Siebe 1994;Guédron et al. 2014). Regression analysis between organic carbon (OC) and the concentration of trace metals in each profile showed positive correlations for all elements (profile 1: OC-Cd (R 2 = 0.93), OC-Cr (R 2 = 0.95), OC-Cu (R 2 = 0.95), OC-Pb (R 2 = 0.97), OC-Zn (R 2 = 0.99); profile 2: OC-Cd (R 2 = 0.80), OC-Cr (R 2 = 0.94), OC-Cu (R 2 = 0.74), OC-Pb (R 2 = 0.95), OC-Zn (R 2 = 0.95)). In contrast, As was not correlated with OC. In the deepest horizons (horizons Ah2 and Ah3), Pb contents were similar to the regional background values reported in Siebe (1994), whereas those of Cd, Cr, Cu, and Zn were respectively 3.0to 4.7-, 7.7-to 8.6-, 3.2-to 4.0-, and 1.9-to 2.3-fold higher than regional background contents in Phaeozems (i.e., Cd: 0.3-1.0 mg/kg; Cr: 15-20 mg/kg; Cu: 7-14 mg/kg; Pb: 2-17 mg/kg; Zn: 27-60 mg/kg). Hence, this supports their enrichment and translocation, even in deep soil horizons.

PTEs carrier phases and potential mobility
Sequential extractions allowed the assessment of PTEs potential mobility and the identification of their carrier phases in the solid phase. Trace metals (Cd, Cr, Cu, Pb, and Zn) were found mainly associated with organic matter (OM), amorphous iron (Fe) oxides, and the residual fraction (Fig. 3). Their distribution onto these 3 main carrier phases differed between elements and depth in the profiles. In the till layer (i.e., top to ~ 30 cm deep), OM was the main carrier phase for Pb and a secondary carrier for Cu, Zn, and Cd, whereas no Cr was found associated with OM. In contrast, amorphous Fe oxides were the main carrier phases for Cu and Zn, and secondary carriers for Pb, Cd, and Cr. Finally, the residual fraction was the main carrier for Cr but a secondary one for Zn, Cu, Cd, and Pb.
In both profiles, the proportion of trace metals associated with the residual fraction gradually increased with depth below the till layer at the expense of the two other main fractions. It is worth mentioning that this increase mirrors the decrease in metal concentrations with depth. Although unidentified mineralogically, the residual fraction is defined chemically as the least reactive phase where metals are incorporated in the crystal structures (Qu et al. 2018). Therefore, elements present in this fraction, such as Cr, can be considered of geogenic origin, likely inherited from the parental materials (Yutong et al. 2016;Edosa Otabor 2019;Wang et al. 2019). Consistently, although the metal distribution onto the 3 main carrier phases is the same between the two profiles, the higher proportion of metals associated with the residual fraction in profile 2 likely results from the lower accumulation of metals at the downstream end of the plot (see Sect. 3.2).
Among the main carrier phases, Fe oxides (also Mn) are sensitive to redox changes and can undergo dissolution in reducing environments (Cornell and Schwertmann 2006). A proportion of Cu, Zn, Cd, and Pb is therefore susceptible to be mobilize under reducing conditions (Mahanta and Bhattacharyya 2011;Yutong et al. 2016;Suda and Makino 2016). In the same soils of the Mezquital valley, González-Méndez et al. (2017) have reported prevailing aerobic conditions with Eh fluctuating between 400 and 790 mV most of the time. Decreases in redox potential were reported during few days after irrigation, but Eh was found controlled mainly by nitrate reduction in soils irrigated with untreated wastewater. Because the reduction of Mn oxides occurs partly in the same redox range as the one of nitrate (i.e., Eh values between 400 and 200 mV), Cd and Pb are the only potentially mobilizable elements among the studied metals (Mn oxides fraction = 7-20% for Cd and 1-7% for Pb). In contrast, Fe oxides are reduced at lower redox potentials (300-100 mV). Therefore, Cu, Zn, and Pb associated with Fe oxides can be considered little mobile in these soils.  (Sofianska and Michailidis 2015;Yutong et al. 2016).
In contrast to trace metals, As was only found associated to the residual fraction (> 55%) and to amorphous iron oxides (< 45%) in both profiles (Fig. 3). Below the till layer, As associated to the residual fraction decreased to the benefit of amorphous Fe oxides, plus a minor proportion in the water-soluble fraction (14 to 19%) and in association with Mn oxides (0 to 7%). The increased solubility of As in deep soil horizons probably results from the (a)biotic reduction of arsenate (As V) which products arsenite (As III) uncharged ions (Abbas and Meharg 2008). Under oxidizing conditions, As (V) is sorbed or coprecipitated with oxyhydroxides, whereas under reducing conditions, As (III) can be released in soil porewater (Beauchemin and Kwon 2006;Fendorf et al. 2010). The presence of a relatively compacted tuff layer at the bottom of the profile (between 60 and 90 cm) likely favors waterlogging and reducing conditions during irrigation events (González-Méndez et al. 2017). Hence, such oscillations in redox condition might drive As speciation and distribution between the solid phase (i.e., amorphous Fe oxides) and the soil solution. This corroborates a previous study who reported As contents in soil groundwater above the maximum permissible limit for drinking water (Guédron et al. 2014).
Calculated mobility factors (MF, Table 2) confirm that the mobility of Cu > Zn > Pb and > Cr (MF < 4%) was low in both profiles, whereas Cd and As were relatively mobile with MF up to 17%. Cd was the most mobile element in superficial samples (MF up to 15%), consistently with the high percentage found in the exchangeable fraction (up to 15% in P1 and 6% in P2). This also highlights that a significant proportion of Cd might be available for plant uptake (Kabata-Pendias 2010; Yutong et al. 2016;Vollprecht et al. 2020). In contrast, As was the most mobile element in deeper samples (MF up to 17%) which highlights a potential for groundwater contamination.

Effect of increased acidity
In the control soil, all the elements analyzed presented values below the quantification limits (As = 0.028, Cd = 0.001, Cr = 0.002, Cu = 0.032, Pb = 0.001, Zn = 0.004, mg/L for all elements) for all pH scenarii.
In all batch experiments, As, Zn, and Cu were the elements the most released into solution with concentrations in the supernatant almost an order of magnitude higher than those of Cd, Cr, and Pb (Fig. 4). For all the studied elements, their solubility was lower in profile 2 compared with profile 1.
Higher releases of Zn (up to 0.34 mg/L) and Cu (up to 0.26 mg/L) to the solution were found for surface than deep soil samples, suggesting a higher reactivity or dissolution of OM and amorphous Fe oxides than the one of the residual fraction which increases with depth (see Sect. 3.3). Zn solubility increased with decreasing pH, whereas Cu did not show significant difference with changes in pH. Dissolved organic carbon (DOC) concentrations were similar in all cases, being slightly higher at pH 6.5 (Table SM4), supporting a higher solubility of OM and its associated elements under circumneutral conditions (Antoniadis and Alloway 2002;Ashworth and Alloway 2006). This is confirmed by the principal component analysis performed on profile 1 (Fig. SM1), where Cu in solution was found associated with DOC. In contrast, Zn was only found associated to Cd and Cu in this PCA. Like Zn, Cd exhibited the highest release to solution at pH 5.5 for surface samples, but almost no Cd was released for deeper soil samples. This likely supports that Cd is bound to a very refractory carrier phase in deeper horizon which is little reactive to changes in pH (see Sect. 3.3). Unlike the other elements, the concentration of Cr in the solution was almost similar for both soil profiles, although a higher one was found in the deep samples of profile 2 at pH 5.5. Cr concentrations were however low (< 0.06 mg/L) for all pH values. This suggests that Cr has a similar carrier phase for both soil profiles which is little affected by changes in pH. Similarly, Pb concentrations in the solution were always low (< 0.005 mg/L), for both soil profiles. The modeling of metals speciation released in the solution predicted similar behavior between Cd-Zn and Cu-Pb. Cd and Zn were mainly predicted under free species (Cd 2+ and Zn 2+ ) in the most acidic scenarios (5.5 and 6.0). In contrast, Cu and Pb were predicted to be complexed with dissolved organic matter in percentages greater than 90% in both cases for all pH values (Fig. SM2). Consistently Cu was significantly correlated with DOC (R 2 = 0.89). In comparison to metals, As solubility increased with depth, and was lower in profile 2 compared with profile 1. In both profiles, As solubility increased with decreasing pH, but the amount of As released in solution was proportional to the initial As load of the soil. This corroborates observation made for total contents and sequential extractions, where As concentration and solubility increased with depth. Arsenic showed significant correlations with bicarbonates (R 2 = 0.85) and Ca 2+ (R 2 = 0.63) (Table SM5), which suggests the dissolution of carbonate phases, where As is bound, with decreasing pH. The model predicted the presence of inorganic As V in solution, dominantly (> 70%) found under dihydrogen arsenate species (H 2 AsO 4 − ), and in a lesser extent as hydrogenarsenate (HAsO 4 2− ). This latter species tended to decrease with decreasing pH for both profiles, consistently with its expected speciation change with pH (Cordeiro Silva et al. 2012;Saha and Sarkar 2015). Dihydrogen arsenate species (H 2 AsO 4 − ) is considered toxic to plants because of its structural similarity with dihydrogen phosphate (H 2 PO 4 − ), an essential element uptaken by plants (Jost 2021).
The results of the analysis of variance are shown in Table 3 (profile 1) and 4 (profile 2). For profile 1, statistically significant differences (p value < 0.05) were observed between As, Cd, and Zn with the variation of pH. This confirms that these elements are the most released in solution with decreasing pH from 6.5 to 5.5.
Hence, a decrease in 1 unit of pH will only result in potential release in Zn, Cu, and Cd in surface soil layer. In contrast, such change significantly affects As mobility mainly in deep soil horizon, likely through the dissolution of carbonates.

Effect of increased salinity
As for the change in the pH scenario, all elements analyzed in solution of increased salinity experiments performed with the control soil were below the quantification limits, except for Cd, which presented a concentration of 0.003 mg/L when the EC was increased to 3 dS/m.
In both P1 and P2 profiles, the increase in salinity only resulted in the release of Zn, Cd, and Cr in solution, whereas Cu, Pb, and As concentrations remained below the detection limit in all cases (Fig. 5). Zn, Cd, and Cr solubility tended to decrease with depth for both soil profiles. The released in solution of Zn (up to 1.2 mg/L) and Cr (up to 0.8 mg/L) was, however, almost tenfold higher than for Cd (up to 0.06 mg/L). The increase in solubility of Zn for an EC of 3 dS/m was higher in samples from profile P1 compared to P2. In contrast, the increase in Cr solubility was similar or higher in surface samples of profile P2 compared The PCA (Fig. SM3) analysis and Pearson's correlation matrix (Table SM7) performed for both profiles only showed evidence of a similar behavior of Cd with NO 3 − , Ca 2+ and Mg 2+ , and Zn within the vector of HCO 3 − , suggesting a cation exchange of both elements onto charged surfaces likely with excess Ca in solution. Consistently, significant correlations between the desorbed elements and the analyzed components of the solutions were found between Cd-EC (R 2 = 0.74,), Cd-NO 3 − (R 2 = 0.74), Cd-Mg 2+ (R 2 = 0.74), and Cd-Ca 2+ (R 2 = 0.79) (Table SM7). In contrast, Cr did not show evidence of any significant correlation. The absence of (or the weak) correlation with DOC suggest a little impact of such increased in salinity on the reactivity of the OM (Table SM6).
Cr speciation was modeled only for the salinity scenario, where Cr solubility was the highest. For all horizons of profile 1, the model predicted that > 95% of the Cr was present in the form of chromium hydroxide (CrOH + ), with few percent of free Cr 3+ in the most superficial horizon of P1 and in all horizons of P2 (Fig. SM4). Under such reduced form, Cr is less toxic and least mobile than other Cr 6+ species (Oliveira 2012). For Zn and Cd, the model predicted a dominance of (i) free Cd 2+ and Zn 2+ and (ii) DOC bound Cd and Zn, two types of species potentially available for plants (Sposito et al. 1982;Grant et al. 1998;Sadeghzadeh 2013).
In comparison with the increased acidity scenarios, Zn, Cd, and Cr solubility were increased for the highest salinity in agreement with reported increased mobilization of trace metals with increasing salinity (Acosta et al. 2011). An increase in salinity of the soils would be especially favored, if irrigation with treated water leads to a change in the irrigation practice towards drip irrigation, rather than field overflow, since the lixiviation of excessive salts out of the rooting zone would be much less.

Effect of water chlorination
In control soil, all the elements analyzed presented values below the quantification limits for all NaClO doses.
In both P1 and P2 profiles, all trace metals showed rising release into solution with increasing NaClO doses, whereas As remained below quantification limits for all chlorination scenarios (Fig. 6). Highest metal contents in the supernatant were found for surface soil samples of P1 compared to P2. Zn, Pb, and Cu showed the highest solubility for P1 at NaClO of 10 mg/L reaching respectively 2.7, 1.4, and 0.9 mg/L in the supernatant. Cd showed a similar pattern with lower concentration in the supernatant in accordance with its lower total concentrations in soil. Both the PCA analysis (Fig. SM5) and Pearson's correlation matrix (Table SM9) for profiles P1 and P2 highlighted the associations of Zn, Pb, Cu, and Cd with DOC and Cl − , suggesting a significant release of these metals from the OM into the solution as complexes with DOC or with Cl − ions. As for the increased salinity scenario, Cu and Pb were predicted    (Table SM8) and OM solubility (Siregar et al. 2004;Mikutta et al. 2005) which has likely favored the release of these metals complexed with DOC. In contrast to the other metals, Cr exhibited a rising concentration in the supernatant with depth, and with almost similar concentrations for both soil profiles. In the supernatant, Cr was found associated with both Fe and DOC. According to selective extractions, this suggests that the fraction of Cr associated with the reactive amorphous Fe oxides has been preferentially released in the solution, followed by a complexation with DOC. Water chlorination shifts the speciation towards organically complexed species of Cd and Zn, and to a lesser extent to free ion species and inorganic Cl − species (Fig. SM6). Similar results were reported by Herre (Herre et al. 2004) who simulated water treatment in a column experiment. In addition, it has been reported that in soils, the mobility of Cd increases in the presence of Cl − due to the formation of the soluble complex CdCl + (Acosta et al. 2011;Tahervand and Jalali 2016). Studies have reported that the presence of Cl − favors the uptake of inorganic Cd complex (i.e., CdCl + ) by plants (McLaughlin and Singh 1999;Cheng et al. 2019). The model predicted that Cu and Pb were almost entirely complexed with dissolved organic matter (i.e., 100% for Cu and between 95 and 100% for Pb), consistently with other studies (Herre et al. 2004;Gu and Bai 2018;Seo et al. 2019), highlighting that organic acids present in the DOC act as chelating agents and promote the mobilization of PTEs (Weng et al. 2002). It has however been shown that organic complexes are less toxic than free ionic forms of trace metals (Inaba and Takenaka 2005;Ashworth and Alloway 2007).
The variation in the dose of NaClO presented the greatest significant variation with the means of 4 of the 6 elements analyzed (Cr, Cu, Pb, and Zn) in both profiles (Tables 3 and  4). Of all the sources of variation analyzed in this work in order to determine their effect on the mobility of PTEs, it was observed that water chlorination is the one that explains the greatest variation in the concentration of PTEs in both profiles analyzed, which shows the relevance of the dose of hypochlorite (or other chlorinating agent) in the concentration of heavy metals.
Arsenic was the only element that was not released in any scenarii of the chlorination simulation, highlighting that, as in the salinity scenarii, the simulated conditions did not cause the desorption of As present in the soil, mainly associated with Fe oxides, indicating that these mineral phases do Considering all metals, the chlorination scenario provides the highest solubility for Zn, Pb, and Cu compared to both pH and salinity experiments.

Environmental and human health implications
Among the main activities that cause pollution problems and adverse effects on the environment and human health is the use of wastewater (Balkhair and Ashraf 2016). In the present study, it was observed that the mobility and speciation of PTEs in soils irrigated with untreated wastewater for more than 90 years is affected by changes in acidity, salinity, and chlorination of the water. The availability of PTEs, especially Cd, could increase with increasing salinity and by simulating chlorination of the waters. Cd is one of the metals with the highest exchangeable capacity, being in this work the element with the highest percentage associated with this fraction. Considering this, Cd is easily soluble in soils, which makes it a bioavailable and cumulative element in the edible parts of plants (Luo et al. 2011). The most active fractions of heavy metals in soils, e.g., water soluble and exchangeable, are the most available to plants (Chavéz et al. 2016). In this sense, it was observed that Cd, as has also been reported in other studies (Siebe 1994;Ponce-Lira et al. 2019), is the most available metal for plants in the soils studied here, with the potential entry into the food chain. The food chain is the most important pathway of exposure to human beings (Balkhair and Ashraf 2016). When metals are incorporated into the food chain, due to their toxic and bioaccumulative nature, there is risk of causing adverse effects on human health (Othman et al. 2021). There are many studies that report various environmental and human health effects caused by the accumulation of heavy metals in soils and plants (Mapanda et al. 2005;Tahri et al. 2005;Elgallal et al. 2016;Gupta et al. 2021;Othman et al. 2021). These metals can affect humans mainly by two pathways, inhalation and ingestion (Balkhair and Ashraf 2016). Elements such as Cd and Pb are known to produce various health effects, mainly associated with kidney, lung, liver, and nervous system problems (Ponce-Lira et al. 2019). It is well documented that continuous wastewater irrigation accumulates heavy metals in food crops (Othman et al. 2021). Accumulation of heavy metals in crops depends on several factors, including plant species. Once present in the soil solution, metals can be transported to the root of the plant and to its different structures, causing a risk to human health when these crops are consumed (Balkhair and

Conclusions
The effect of the change in the soil-water management system on the mobility of PTEs was tested by the combination of sequential extractions of soil profile samples, and batch experiments simulating a decrease in pH, an increase in salinity and chlorine.
Sequential extractions showed that most of trace metals (Pb, Cd, Cr, Zn, and Cu) were mainly associated with organic matter, amorphous iron (Fe) oxides, and an unidentified residual fraction. Organic matter and amorphous Fe oxides were the main carrier phases in surface layer and decreased with depth with the rising of the residual fraction. Only Cr and As showed a different distribution with Fe oxides and the residual fraction as the main carrier phases. The assessment of their mobility through the calculation of mobility factor showed that Cd and As were the most mobile elements. For the other less mobile elements, their mobility decreased in the following order: Cu > Zn > Pb and > Cr.
Among tested experiments, the chlorination scenarii enhanced the most the solubility of Zn, Pb, and Cu compared to both pH and salinity experiments for the studied Phaeozem. In contrast, the increase in salinity favored the release of Cd and Cr. Finally, the decrease in pH greatly enhanced the release of As in the soil solution. These results show that the chlorination dosage in the treatment for irrigation water needs to be chosen as low as possible, to prevent risks of high Zn, Pb, and Cu release into groundwater. On the other hand, soil acidification could cause an increase in the mobility of PTEs, mainly As and Zn.
Soils of the Mezquital Valley are known to be efficient sinks for PTEs. However, a change from wastewater to treated water for irrigation could mobilize these elements to the soil solution with implication for the contamination of crops and groundwater. Our findings show that the mobility of heavy metals, mainly Cd and Zn, can increase due to the change in the quality of the irrigation water; in addition, it was observed that the mobility of As increases with depth and is affected by pH changes, which could explain the presence of this element in the groundwater of the site, which has been reported in other works. This increase in the mobility of PTEs could cause a greater risk in the near future to the human health of consumers of crops and water in the Mezquital Valley. Therefore, further studies will need to follow the long-term effect of such changes in irrigation practices, including the pH of the soil-water system, which might greatly affect the mobility of PTEs. The results of this work show the need to carefully monitor the change in water quality and its possible effects on the mobility of PTEs, in addition to developing different strategies to prevent the accumulation of PTEs in food crops, to minimize the chronic risk for the health of the exposed population in the Mezquital Valley. In order to avoid future risks to the human health of the consumers of the crops, it is advisable   to carry out investigations that use various risk assessment techniques, such as the uptake/transfer factor (UF) in maize and alfalfa (the most economically important crops in the valley), in addition to health risk index (HRI), enrichment factor (EF), contamination factor (CF), pollution load index (PLI), hazard quotient (HQ), and the morbidity status (ST).