Introduction

Concentrations of heavy metals (HMs) in confined and unconfined aquifer groundwater threaten customers with cancer and non-cancer risks through ingestion, dermal, and inhalation routes. Users and regulatory agencies are concerned about the presence of toxic and bioaccumulative HMs in groundwater sources in arid regions1. Heavy metals pose hazards to both ecological life and human health due to their toxicity, slow biodegradability, and accumulation in living organisms2. With their roles as catalytic and structural elements in various proteins and enzymes, metals pose significant human health risks through cutaneous and intestinal absorption3. The interference of overexposure to HMs with various metabolic processes in the human body results in a range of toxicological impacts on bodily organs, including the bones, brain, and kidneys. Nevertheless, the accumulation is a function of the specific chemical structure of a metal and the coexistence of other components (e.g., ionic radius and electronegativity) that affect its bioavailability, uptake, and retention in the body4.

The meager presence of exceeding maximum allowable concentrations (MAC) established by regulatory agencies does not guarantee complete health safety, as even minute amounts within MAC pose health risks to vulnerable population groups, such as infants, children, and older adults5,6. Although low concentrations do not significantly affect cell viability, sub-lethal detrimental biological effects of low levels of HMs can suppress immune responses, leading to long-term health consequences7,8. The human health risk assessment (HHRA) is a systematic process that evaluates significant health risks associated with exposure to various heavy metals and minimizes these exposures through mitigation actions to achieve acceptable levels of risk9.

HHRA is a crucial for mitigating the health risks from contaminated groundwater, which is a vital source of municipal water supplies worldwide. Due to the primary reliance on surface water in the United States and Canada, the availability of demographic data, and the existence of well-structured official mechanisms, most studies have evaluated the impact of anthropogenic activities on surface water quality10,11. Several studies on groundwater quality investigated confined aquifers in Asian countries, including Bangladesh12, Iran13,14, India15,16, Pakistan17, and regions in Africa and the Sahara18,19. Confined (or deep) aquifers are generally the source for large urban settings, while small agricultural settings, such as farms and residents on crop fields, rely on shallow aquifers20. HHRA studies on shallow wells are currently limited, primarily due to the limited availability of data and the insufficient consideration given to smaller exposure groups.

Studies have accused natural and anthropogenic activities of contaminating groundwater with HMs in the Kingdom of Saudi Arabia (KSA). For instance, studies by Mallick et al.1 and Alshehri et al.21 highlighted that domestic and industrial wastewater discharges, as well as the use of fertilizers and pesticides in agricultural applications, contribute to high levels of heavy metals in KSA. Thabit et al.22 identified the natural weathering of rocks as the primary cause of elevated HMs in the Qassim Region. Tasleem et al.23 draw attention to the cumulative impact of natural and anthropogenic activities on the increased levels of lead (Pb) and cadmium (Cd) in groundwater in the Madinah Region. Considering the promulgation of regulatory measures by the KSA’s government for limiting HMs before discharging industrial effluent to the municipal sewers and the efficient performance of tertiary-level wastewater treatment facilities, natural contamination can be identified as the primary root cause of HMs’ existence in groundwater1,24,25.

Shallow alluvial aquifers near major cities, often settled above the major deep aquifers, are often plagued by groundwater pollution. Studies reported the following HMs in KSA groundwater: aluminum, arsenic, barium, boron, cadmium, chromium, cobalt, copper, iron, mercury, molybdenum, manganese, nickel, lithium, selenium, strontium, lead, vanadium, and zinc22,26. Alfaifi et al.26 identified concentrations of As, Mn, Cr, Ni, Se, and Zn exceeding the permissible drinking water quality limits in the groundwater of southern Saudi Arabia. They identified agricultural activities and the dissolution of halite and gypsum in Sabkha deposits along the Red Sea Coast as the primary cause of HMs in groundwater. Additionally, their monitoring results revealed that Fe, Li, B, Cu, Mo, Sr, and V were all within the permissible limits.

The concerned authorities in KSA face a twofold challenge: the low availability and natural contamination of source water27. HMs without unpleasant taste or smell are less noticeable than contaminants with acute and evident problems. Water directorates and municipalities in urban centers with groundwater reliance supply highly treated water using reverse osmosis (RO) membranes, which remove taste issues (due to salts) and health concerns (due to heavy metals, HMs). Rural and agricultural settings in KSA are spread across almost all regions, where untreated shallow wells meet both agricultural and residential demands. Fortuitously, total dissolved solids (TDS) levels exceeding 600 mg/L in most groundwater in KSA prompt consumers to seek alternative desalinated options, such as bottled water or household-level purifiers, thereby eliminating the possibility of human health risks through ingestion. The use of untreated shallow wells (80–120 m deep) for washing and bathing demands an efficient dermal risk assessment methodology for agricultural settings in KSA.

Most past studies28,29 in KSA have evaluated point-estimate health risks associated with contaminated water using average or maximum values of various risk factors (surface area, body weight, and concentrations) in the application of the USEPA risk assessment model30. Groundwater pollution assessment remains challenging due to uncertainties associated with the limited number of sampling points, which may be unable to fully capture the variations in hidden subsoil conditions and geological formations. The HHRA process itself involves various factors that contribute to health risks, such as body weights, exposure durations, and skin surface areas of different population groups, all of which exhibit significant variations. The probabilistic health risk assessment approach utilizes inclusive distributions of data (contaminant levels and other parameters) that encompass the range of exposure states, resulting in a more accurate health risk assessment31.

The present study developed a probabilistic human health risk assessment approach for dermal exposure to contaminated groundwater in agricultural settings in Saudi Arabia. The main objectives of the study are to (i) monitor groundwater used at farm wells and determine exposure pathways and routes for heavy metals of concern, (ii) conduct probabilistic non-cancer and cancer risk assessment for various population groups, and (iii) evaluate the health risk severity and establish guidelines for sustainable use of groundwater resources in agricultural settings in Saudi Arabia. To evaluate the pragmatism, the proposed approach was applied to farms and date orchards in an agricultural setting of the Qassim Region in KSA.

Methods

Study region

The study region encompasses four major cities in Qassim Province, Saudi Arabia: Buraydah, Unayzah, ArRass, and AlBukayriah. Figure 1 illustrates a map showing the study region’s boundaries developed using QGIS version 3.42.0. The area is characterized by extensive agricultural activities due to the availability of suitable groundwater in the Saq aquifer, which is primarily located at depths of 80 m to 120 m. The selected region, alongside Wadi Rumah (44°05′01″N, 57°18′26″E and 43°31′15″N, 03°27′25″E), is located at an elevation of 590 m to 660 m above mean sea level. With over 39,000 hectares of cultivated land and an annual yield of 190,000 tonnes of dates, the Qassim Region has gained worldwide recognition for its date production32. Wheat, barley, and vegetables are other products that thrive in the region’s fertile soil, utilizing drip and sprinkler irrigation techniques. Because the groundwater in Qassim primarily exists in non-replenishable fossil aquifers, it is being depleted due to agricultural extractions along with residential and industrial uses27,33. Being in the arid environmental zone, the National Water Company (NWC) and municipalities in the Kingdom of Saudi Arabia (KSA) acknowledge the challenges of water scarcity. To this effect, tertiary-level wastewater treatment facilities that meet KSA’s restricted and unrestricted irrigation standards contribute to the region’s agricultural production34.

The unconfined aquifer meets the promulgated agricultural water quality standards for physical (turbidity), chemical (salts and nitrates), and biological (coliforms) water quality parameters. The presence of heavy metals in the Saq aquifer has been reported in past studies21,22. The present study assesses the health risk associated with exposure to these heavy metals.

Groundwater monitoring and testing

Groundwater samples were collected from twenty different farms spread across the municipal boundaries of Buraydah, Unayzah, AlBukayriyah, and ArRass cities in the study region outlined in Fig. 1, with sampling locations not shown for confidentiality. The farms were selected based on the known depth of the wells and the primary uses of groundwater. The information was obtained from the residents, who were mostly farm workers. The use of raw water for non-potable purposes and dermal exposure routes was also established from their lifestyle. The raw groundwater is stored in ground storage tanks constructed for irrigation through overland, drip, or sprinkler systems. In most cases, these tanks are located near farm wells, and the collected groundwater is pumped to the rooftops of residential buildings for purposes such as washing, cleaning, flushing, and bathing. The samples were stored in one-liter sterilized glass bottles and transported in an ice box to the Environmental Engineering Laboratory at Qassim University’s College of Engineering and NWC’s Central Laboratory within 4 h after collection.

Fig. 1
Fig. 1
Full size image

Location of the Qassim region in Saudi Arabia and boundaries of the study region, which covers agricultural farms in the vicinity of four major cities in Qassim. The map was prepared in the present study using QGIS version 3.42.0 (https://www.qgis.org).

Heavy metal concentrations, including Cu, Fe, Pb, Ag, Tl, Ba, B, Cd, Cr, Li, Mn, Sr, and Zn, were quantified using two distinct analytical techniques, selected based on the analyte and its expected concentration range. Preliminary colorimetric analysis was conducted using a DR 3900 Spectrophotometer (Hach, Colorado, USA) for Ba and B, which are amenable to spectrophotometric detection using Hach’s reagent-based methods. The remaining eleven metals were analyzed using Inductively Coupled Plasma–Optical Emission Spectrometry (ICP-OES; Thermo Scientific iCAP 6000 Series) for higher sensitivity and multi-element capability. These thirteen heavy metals were selected for analysis based on historical data35 and information provided by the National Water Commission (NWC). However, Cu, Fe, Pb, Ag, and Tl were found to be below the method detection limits (MDLs) and were therefore excluded from the subsequent human health risk assessment (HHRA). The remaining eight metals, including Ba, B, Cd, Cr, Li, Mn, Sr, and Zn, were retained for toxicokinetic evaluation and dermal exposure assessment.

Standard reference materials were used for instrument calibration and method validation. All measurements were performed in triplicate, and procedural blanks were included in each analytical batch. Ultrapure deionized water (18.2 MΩ·cm) was used as the blank control. Field samples were acidified immediately after collection using 0.1 M nitric acid (HNO₃, pH < 2) to preserve metal speciation and minimize adsorption to container surfaces. Method detection limits (MDLs) for the selected analytes were established following USEPA protocols, using seven replicate measurements of low-concentration standards. Calibration curves were constructed using multi-element standard solutions across six concentration levels, with correlation coefficients (R²) of ≥ 0.998. The calibration ranges and MDLs for each metal were as follows: Ba: 10–1000 µg/L (MDL = 10 µg/L), B: 50–1000 µg/L (MDL = 50 µg/L), Cd: 0.5–100 µg/L (MDL = 0.2 µg/L), Cr: 1–500 µg/L (MDL = 0.3 µg/L), Li: 3–500 µg/L (MDL = 1.0 µg/L), Mn: 1–1000 µg/L (MDL = 0.4 µg/L), Sr: 1–1000 µg/L (MDL = 0.5 µg/L), and Zn: 2–1000 µg/L (MDL = 0.7 µg/L).

Toxicological review

The present study adopted the guidelines established by the United States Environmental Protection Agency (USEPA) for HHRA regarding dermal exposure to consumers in the study region. Table 1 presents the maximum allowable concentrations (MACs) established by the Ministry of Environment, Water, and Agriculture (MEWA) of the Kingdom of Saudi Arabia (KSA) and the US Environmental Protection Agency (USEPA), along with the potential exposure routes.

According to a toxicological review of barium and compounds published by USEPA36, data on the dermal absorption of barium compounds are not available. Based on the possibility of raised blood pressure, USEPA37 established a MAC of 2 mg/L for Ba in drinking water. Health Canada38 also established the MAC level of 2 mg/L for Ba in drinking water. Later, the USEPA Integrated Risk Information System (USEPA-IRIS) derived a daily reference dose (RfD) of 200 µg/kg for nephrotoxicity from a 2-year mouse study and updated the drinking water equivalent level (DWEL) for humans to 7 mg/L for Ba39. Absorption of Ba during showering and bathing through the dermal route is negligible due to a low dermal permeability coefficient (Kp) of 1 × 10−3 cm/h40. However, a study by Thang et al.41 on dermal exposure to low concentrations (687 µg/L) of Ba in groundwater in Vietnam reported that both long-term and short-term exposures promote invasion of keratinocytes (the primary cell type in the outermost layer of the human skin, which constitutes 90% of epidermal human skin cells), interfering with normal skin functions and enhancing malignant characteristics (i.e., potential to promote neoplastic changes). According to the USEPA Guidelines for Dermal Risk Assessment, 7% is specified as the fraction of chemicals absorbed in the gastrointestinal tract (ABSGI) for Ba in water40.

Table 1 Maximum allowable concentrations (MAC) for heavy metals of concern.
Full size table

As per the Health Effects Support Document for Boron published by USEPA51, Boron can not be absorbed through intact human or animal skin except severely damaged skin. Health Canada43 recommends a MAC level of 5 mg/L for Boron in drinking water. The USEPA44 and Health Canada45 established MAC levels of 0.005 mg/L and 0.007 mg/L, while MEWA42 recommended 0.003 mg/L for Cd in drinking water. Long-term exposure to high Cd levels through ingestion can affect the kidneys or bones45. The guideline is based on adverse effects on the kidney, as they occur at low exposure levels and are well characterized. Although both cancer and non-cancer health effects of Cd in drinking water have been linked to ingestion and inhalation routes, a low permeability constant of 1.0E-03 cm/h suggests negligible absorption through the dermal route during bathing, showering, and washing52. Past studies have reported health repercussions, including renal dysfunction and osteoporosis, in humans resulting from long-term dermal exposure to soil and dust contaminated with cadmium (Cd)53. The USEPA Guidelines for Dermal Risk Assessment report a 5% ABSGI for cadmium in water40.

The International Agency for Research on Cancer (IARC) classified Cd and Cr as Group 1, carcinogenic to humans54. According to USEPA-IIS, Cr (VI) can cause developmental, immune system, lower respiratory, male reproductivity, hematological, liver, and gastrointestinal tract toxicities in humans. It is potentially carcinogenic to the human gastrointestinal tract. The IRIS recommends an overall chronic reference dose (RfD) of 9 × 10−4 mg/kg/day47. The USEPA Guidelines for Dermal Risk Assessment report an ABSGI of 2.5% and a Kp value of 2.0E-03 cm/h for Cr in water40.

The carcinogenicity of Li has not been established in proven literature. However, based on a provisional oral reference dose (p-RfD) of 2 µg/kg/day, the USEPA provides a non-regulatory Health Reference Level (HRL) of 10 µg/L for drinking water supplies55. The same screening level of 10 µg/L has been recommended by the United States Geological Survey56. Assuming drinking water is the only source of lithium in daily intake, the USGS published 60 µg/L for assessing Li in groundwater57. However, literature does not prove lithium’s dermal health risk.

The likelihood of Mn exposure through dermal and inhalation routes from drinking water is almost negligible58. Although a few studies have associated Mn in drinking water with neurological effects in children, there is no experimental evidence of dermal absorption of Mn48. The USEPA Guidelines for Dermal Risk Assessment report a 6% ABSGI and a Kp value of 1.0E-03 cm/h for Mn in water40. According to Health Canada49, limited studies on humans have reported adverse effects of Sr on bones, while adverse impacts of high doses have been proven in animal studies. Infants are particularly sensitive, with the highest sensitivity to adverse effects on bones during the first year of life. Dermal absorption of Sr through bathing and showering is unlikely to contribute significantly to overall exposure. Zn is an essential nutritional element for human health; however, excessive intake of Zn can impair the human immune response. The MAC of 5 mg/L, suggested by USEPA and Health Canada, is based on taste criteria for drinking water50.

Human health risk assessment

The present study followed four steps prescribed by the USEPA Health Risk Assessment Guidelines for HHRA: hazard identification, dose-response assessment, exposure assessment, and risk characterization59.

Based on the toxicological review presented in the previous section, the present study identified Ba, Cd, Cr, Mn, and Zn as noncancerous and Cd and Cr as cancerous hazards. The primary criteria for inclusion were the values of ABSGI and Kp, as reported by USEPA40, Risk Assessment Guidance. Figure 2 presents the exposure routes in the study region. The farmers, villagers, and agricultural workers do not drink raw groundwater drawn for farm irrigation due to average total dissolved solids (TDS) levels higher than 1000 mg/L, nor do they comply with KSA’s drinking water quality standards (KSA-DWQS), nor the World Health Organization (WHO) aesthetic drinking water quality guideline of 600 mg/L. Thousands of workers and residents in hundreds of these farms use raw groundwater for swimming, washing, and bathing, which can be potential receptors for heavy metals of concern through the dermal route. The study assumed no inadvertent water ingestion during washing and bathing activities for all population groups.

The agricultural farms host the families of farm owners and accommodate farm workers and their families. Therefore, the receptors in the study region comprise all five age groups: infants (0–6 months), toddlers (7 months to 4 years), children (5–7 years), teens (12–19 years), and adults (20–80 years). The average life expectancy at birth in KSA is ~ 77.6 years60.

Fig. 2
Fig. 2
Full size image

Exposure pathways and routes for probabilistic human health risk assessment of dermal exposure to heavy metals among consumers of shallow farm wells.

For dermal HHRA, the Average Daily Dose via dermal routes (\(\:{ADD}_{dermal}\)) was estimated using the following Equation40(USEPA 2004).

$$\:{ADD}_{dermal}=\frac{{C}_{w}\times\:SA\times\:{K}_{p}\times\:ET\times\:EF\times\:ED\times\:CF}{BW\times\:AT}$$
(1)

where \(\:{C}_{w}\) represents the concentration of heavy metal in water (mg/L), EF is the exposure frequency (365 days per year) considering the dermal exposure of most residents through bathing and ablution in the study region, ED denotes exposure duration for different age groups (years), Kp is the dermal permeability coefficient (cm/h), CF is the unit conversion factor (0.001 L/cm3), and AT represents average time equals ED x EF (days).

A past study by Khalid and Ali61 in KSA reported variations in body weights of men (65.1 ± 13.1 kg) and women (61.9 ± 13.1 kg), as well as in their heights: men (162.1 ± 6.7 cm) and women (149.7 ± 5.5 cm). Although the study is old, it provides sufficient evidence of significant variation in body weight (BW) and skin surface area (SA) among both males and females between the ages of 17 and 72 years. El Edelbi et al.62 reported skin area (SA) for newborns to one-year-olds ranging between 2100 cm² and 3700 cm², and for 1–18-year-old children and teens, between 6500 cm² and 13,500 cm². In these studies, upper-bound values for SA and BW match the USEPA63 recommended values in Table 2. Nevertheless, these variations suggest a probabilistic HHRA for more realistic estimates. Without detailed information on local risk factors (BW, SA, and ET) for probabilistic HHRA for KSA, the present study applied data published in the Exposure Factors Handbook by the USEPA, based on the analysis of the NHANES64. Table 2 summarizes the mean values for all the risk factors in Eq. (1). @Risk Software (Version 8.10.0 – Industrial Edition) generated probability distributions for SA, ET, and BW to account for the uncertainties associated with the limited data for these variables in KSA.

Table 2 Risk factors used in the present study for different exposure groups.
Full size table

The Hazard Quotient (HQdermal) was estimated to assess the potential non-cancer health risk of heavy metals in the raw water via the dermal pathway using Eq. (2).

$$\:{HQ}_{dermal=\frac{{ADD}_{dermal}}{{RfD}_{dermal}}}$$
(2)

Where RfDdermal represents the reference dose of heavy metals. Equation (3) derived RfDdermal from the ABSGI of each heavy metal.

$$\:{RfD}_{dermal}={RfD}_{oral}\times\:{ABS}_{GI}$$
(3)

Equation (4) estimated the Hazard Index (HIdermal) to evaluate heavy metals combined non-cancer health effects in farm wells via dermal exposure.

$$\:{HI}_{dermal}=\sum\:{HQ}_{dermal}$$
(4)

The HQdermal estimates the toxicity potential of heavy metals through the dermal route; a value of HQ < 1 corresponds to an acceptable risk, and HQ > 1 identifies a significant potential health risk. HI assesses the overall potential of non-cancer health risks due to exposure to more than one metal.

Equation (5) estimated the probability of cancer development in an individual due to dermal exposure to Cd and Cr.

$$\:{CR}_{dermal}={ADD}_{dermal}\times\:{SF}_{dermal}$$
(5)

where CRdermal represents the carcinogenic risk via dermal exposure, and SFdermal is the dermal slope factor of heavy metals derived from Eq. (6)40.

$$\:{SF}_{dermal}=\frac{{SF}_{ingestion}}{{ABS}_{GI}}$$
(6)

Finally, Eq. (7) estimated the incremental lifetime cancer risk (ILCR) by summing the CRdermal for different life stages over the lifetime exposure.

$$\:ILCR=\:\sum\nolimits_{i}^{n}{CR}_{i}\times\:{F}_{i}$$
(7)

where CRi denotes the cancer risk estimated from Eq. (5) for the dermal exposure of each group (i), and Fi represents the fraction of the exposure period over an 80-year lifetime. Table 3 lists the values of RfDdermal, Kp, ABSGI, and SF for all the heavy metals considered in the present study.

Table 3 Reference doses, slope factor, and reference dose values for the selected heavy metals.
Full size table

Both point-estimate and probabilistic HHR assessed cancer and non-cancer risks of dermal exposure to the residents of farms in the study area. The point-estimate HHRA used median values for body weight and surface area, as given in Table 2, to illustrate the central tendency of the variability in risk estimation. Meanwhile, probabilistic HHRA describes the variations in concentrations of heavy metals and uncertainties in body weights and skin surface area using probability distributions. Appropriate probability density functions (PDFs) were developed for the given data in the Exposure Factors Handbook (USEPA 2011). The 10,000 Monte Carlo permutations using @Risk Version 8.10.0 (Palisade Corp., USA) generated PDFs for HQs, HI, and ILCRs to identify the critical parameters, such as concentrations, body weight, and intake rates. HI < 1 represents an insignificant non-cancer health risk.

The present study employed reference doses and slope factors recommended by the USEPA’s Integrated Risk Information System and Health Canada’s Toxicological Reference Values for cancer and non-cancer risk estimation, as listed in Table 3, for toxicity assessment. The USEPA recommends the acceptable health risk range between 1 in 1,000,000 (1E-6) and 1 in 10,000 (1E-4)64(Sullivan et al. 2005). In “Guidance on Human Health, Detailed Quantitative Risk Assessment for Chemicals (DQRAChem),” Health Canada65 recommends an ILCR of less than 1E-5 as insignificant. The present study adopted a benchmark value of 1E-4 as the acceptable threshold for lifetime cancer risk.

Water pollution indices

Heavy metal pollution index

The heavy metals pollution index (HPI) inclusively assesses the cumulative impact of different heavy metals in water samples on the overall quality of groundwater. Equation (8) calculates the HPI for the study region68.

$$\:HPI=\frac{\sum_{i-1}^{n}{W}_{i}{Q}_{i}}{\sum_{i=1}^{n}{W}_{i}}$$
(8)

Where Wi are the unit weights of heavy metals, calculated by inversely proportioning the MAC values of individual parameters and then normalizing them, and Qi represents the sub-index of each heavy metal, estimated using the following Eq. 

$$\:{Q}_{i}=\sum\nolimits_{i=1}^{n}\frac{\left|{M}_{i}-{l}_{i}\right|}{{S}_{i}-{l}_{i}}\times\:100$$
(9)

Where Mi denotes the measured concentration of the ith sample (mg/L), li represents the ideal concentration of the ith heavy metal (mg/L), and Si is the maximum permissible value of the heavy metal in drinking water (mg/L). Both the numerator and denominator in Eq. (9) are calculated as absolute numerical differences without regard to sign. MAC values suggested by MEWA42 for KSA in Table 1 were considered as maximum permissible values (Si) for heavy metals, and ideal concentrations (li) were considered “0” for calculating Qi.

HPI values less than 100 indicate non-contaminated water, while values close to 100 indicate the threshold for a heavy metal contamination hazard. An HPI value higher than 100 indicates significant heavy metal contamination in the groundwater68.

Heavy metals evaluation index

The Heavy Metals Evaluation Index (HEI) directly assesses water pollution by comparing the monitored values in samples with the permissible levels, as shown in the following equation68.

$$\:HEI=\sum\nolimits_{i=1}^{n}\frac{{Hc}_{i}}{{Hmax}_{i}}$$
(10)

Where, \(\:{Hc}_{i}\) represents the measured concentration of a heavy metal (mg/L) and \(\:{Hmax}_{i}\) is the maximum allowable concentration of the heavy metal of concern (mg/L) suggested for KSA in Table 1.

HEI values less than 40 correspond to a low pollution level, between 40 and 80 indicate a medium level, and HEI values higher than 80 indicate a high pollution level due to heavy metals, posing a significant risk to human health.

Results

Heavy metals monitoring of farm wells

In addition to heavy metals of concern, pH and TDS were tested for the samples collected from farm wells in the study region. The probability density function for TDS in Fig. 3a shows that 90% of values lie between 500 and 3000 mg/L. Raw water drawn from farm wells has an 86% chance of exceeding the aesthetic drinking water quality criteria of 600 mg/L TDS. These results support the study’s hypothesis that residents generally do not use raw water for drinking purposes, and the primary route of exposure is through dermal contact. Figure 3b illustrates that the probability of raw water pH meeting the KSA-DWQS (6.5–8.5) is 0.97. Turbidity was found to be less than 1 NTU in all samples.

Figure 4a-e explains the variations for Ba, Cd, Cr, Mn, and Zn in the study region. Figure 4a shows a 0.97 probability that Ba will not exceed the MAC of 1.3, as suggested by MEWA42(2022). The Figure also illustrates a 0.996 probability of not exceeding the MAC of 2.0, as established by Health Canada38, and a 0.999 probability of not exceeding the MAC of 7, as established by USEPA39. The presence of Ba in the groundwater of the study region, compared to the local MAC value in KSA, underscores the need for a probabilistic non-cancer health risk assessment.

Fig. 3
Fig. 3
Full size image

Cumulative probability distributions for total dissolved solids and pH in the study region.

Figure 4b shows a 0.9 and 0.99 probability of B under the MACs of 2.4 mg/L by MEWA42 and 5.0 mg/L by Health Canada43. The toxicological review did not identify B as a potential hazard through dermal routes and was thereby excluded from the risk assessment. Figure 4c, which explains the results for Cd in raw water in the study region, shows a 0.97 probability of Cd not exceeding the MAC of 0.003 by MEWA42 and a value greater than 0.99 for the MAC values of 0.005 mg/L recommended by USEPA44 and 0.007 mg/L by Health Canada45.

Figure 4d describes cadmium levels in farm wells in the study region. Although the probability of Cr found in the study region not exceeding the MAC value of 0.1 mg/L recommended by USEAP59 is more than 0.99, the results show a 0.15 probability of Cr exceeding the MAC of 0.05 recommended by MEWA42(2022) and Health Canada46, suggesting the need for dermal HHRA. Figure 4e shows a 0.88 probability that Li will not exceed the non-regulatory screening level of 0.06 mg/L set by the USGS56. Due to the absence of established proof of health risk from ingestion or dermal exposure to lithium in water, the present study excluded Li from dermal HHRA.

Figure 4f illustrates a 0.01 exceedance probability of manganese relative to the MAC of 0.4 mg/L, as reported by MEWA42, and the MAC of 0.1 mg/L, as determined by Health Canada48. The present study included Mn in non-cancer dermal HHRA due to its presence in the water from farm wells in the study region, as well as the availability of Kp and ABSGI in Table 3, indicating potential dermal risk. Figure 4g illustrates a 0.7 probability that Sr will not exceed the MAC of 7.0 mg/L, as set by Health Canada49. The present study did not consider Sr in non-cancer HHRA due to the absence of established proof of potential hazard from dermal exposure.

Figure 4h, which describes the results for Zn in farm wells installed in the study region, shows a 0.01 exceedance of Zn over the MAC of 0.3, as determined by MEWA42, and a probability of more than 0.001 of the MAC value of 0.5 mg/L, as recommended by USEPA and Health Canada50. The present study included Zn in non-cancer health risk assessment due to its presence in the study region’s groundwater and the availability of Kp and ABSGI in Table 3, indicating a potential dermal risk.

Fig. 4
Fig. 4
Full size image

Heavy metals in the water from shallow farm wells in the study region. (a) Barium, (b) Boron, (c) Cadmium, (d) Chromium, (e) Lithium, (f) Manganese, (g) Strontium, and (h) Zinc. [HC represents Health Canada MAC values, and KSA represents MEWA42(2022) standards given in Table 1].

The Pearson correlation analysis in Table 4 revealed that most water quality parameters exhibited weak to moderate correlations, which were statistically non-significant (p > 0.05). Among the relationships tested, a statistically significant negative correlation was observed between barium (Ba) and manganese (Mn) (r = − 0.677, p = 0.032). These results indicate that higher Ba concentrations are associated with lower Mn concentrations, suggesting possible geochemical competition or differing source contributions of these metals in the aquatic environment. A moderate negative correlation was also found between Ba and chromium (Cr) (r = − 0.569, p = 0.086), while cadmium (Cd) and zinc (Zn) showed a moderate positive relationship (r = 0.567, p = 0.087). Although these latter associations were not statistically significant, they may reflect similar geochemical behavior or shared anthropogenic sources. pH demonstrated generally weak correlations with other parameters, with the highest being a slight positive correlation with Zn (r = 0.304, 0.393) and Cr (r = 0.261, p = 0.717), suggesting that within the studied pH range, heavy metal distribution in groundwater was not significantly influenced by pH variations.

Table 4 Pearson correlation matrix showing correlation coefficients between water quality parameters.
Full size table

Point-estimate human health risk assessment

Point-estimate HHRA was performed for the samples collected from the farm wells in the study region. The hazard index for dermal exposure (HIdermal) represents the cumulative non-cancer health risk from dermal exposure to the chemicals of concern (Ba, Cd, Cr, Mn, and Zn). The values of HQ > 1, as indicated by Eq. (1) to (4), suggest a likelihood of non-cancer health risk. The hazard index in Fig. 5a aggregates HQs for the heavy metals, indicating an insignificant (HI < 1) non-cancer risk from dermal exposure to the heavy metals for all population groups in the study region.

Figure 5b shows the box and whisker plots of the point estimate ILCR for cadmium and chromium exposure, using Eq. (7). The calculated ILCRdermal values based on MCL values given in Table 1 and risk factors presented in Table 2 for Cd exposure are as follows: 4.73E-0442, 7.88E-0444, and 1.10E-0345. The calculated ILCRdermal values for Cr exposure are as follows: 1.68E-0442, 3.36E-0459, and 1.68E-0446. Figure 5b demonstrates that the median ILCRdermal (1.33E-04) for Cd is less than the MEWA42 and USEPA44 values and slightly higher than Health Canada45, while the 75th percentile is slightly over 2.0E-04. In the case of Cr exposure, the median ILCRdermal (8.33E-05) is lower than all the ILCRdermal estimates based on the regulatory values, with the 75th percentile (1.56E-04) slightly exceeding 1.0E-04.

Fig. 5
Fig. 5
Full size image

Point-estimate dermal health risk assessment results: (a) Hazard index (HI), (b) Incremental lifetime cancer risk (ILCR). The boxes in the whisker plots correspond to the 25th, median, and 75th percentiles.

Probabilistic human health risk assessment

Figure 6 illustrates the probabilistic non-cancer HHRA results for dermal exposure. In Fig. 6a, the median (50th percentile) probability of HIdermal for all exposure groups stays between 0.1 and 0.2, while the values lie between 0.3 and 0.6 with 0.9 probability. Whisker and box plots in Fig. 6b show 5th, 25th, median, and 75th percentiles for HIdermal less than 0.4, suggesting no significant non-cancer health risk to any exposure group through dermal exposure to the chemicals of concern, including Ba, Cd, Cr, Mn, and Zn.

Probabilistic cancer risk assessment for dermal exposure to the exposed population was performed using Eqs. (5–7). Reference doses and slope factors, as listed in Table 3, were used to estimate cancer risk for each exposure group. Probability distributions were also plotted for the ILCR distributions for the MAC values of 0.003 mg/L for Cd and 0.05 mg/L for Cr, as recommended by MEWA42 for KSA, 0.005 mg/L for Cd and 0.1 mg/L for Cr, as per USEPA30,44, and 0.007 mg/L for Cd and 0.05 mg/L for Cr, as suggested by HC45,46 in Table 2.

Figure 7a shows that, although the dermal exposure to Cd shows the estimated ILCR value of 6E-05 for the farm well users at the median probability, which falls below the acceptable cancer risk of 1.0E-4, as well as the estimated ILCR of 4.0E-4, which is the MAC recommended by MEWA42. Although the ILCR increases to 3.2E-04 at 0.9 probability, it remains lower than the corresponding ILCR value of 9.0E-4 estimated at MAC suggested by MEWA42. Furthermore, ILCR from dermal exposure to cadmium remains lower than the ILCRs estimated for the MACs recommended by USEPA44 and Health Canada45.

Fig. 6
Fig. 6
Full size image

Probabilistic non-cancer dermal health risk assessment results showing Hazard Index (HIdermal) for all exposure groups: (a) cumulative probability distributions, and (b) box plots.

Figure 7b presents the cancer risk assessment results for Cr exposure through dermal routes for the exposed population in the study region. The probability distributions indicate that the calculated ILCR value of 8E-05 at the median probability is less than 1.2E-4, which is the estimated ILCR for MAC recommended by MEWA42 and HC46, as well as 3E-4 for USEPA30. These results indicate that the estimated ILCR, based on median values of risk factors, falls below the acceptable cancer risk of 1.0E-4 in the study region. Although the ILCR increases to 2.2E-04 at a 0.9 probability, it is still less than the corresponding ILCR value of 3.1E-4 estimated at a MAC of 0.05 mg/L, recommended by MEWA42 and HC46, and 6E-4 observed for USEPA30.

@Risk Software (Version 8.10.0 – Industrial Edition) Monte Carlo simulations using the Latin Hypercube Sampling scheme, which enhances distribution coverage by dividing each input distribution into equal-probability intervals and sampling once from each interval. The outputs were converged at 100,000 simulations with a convergence tolerance of 1% and a confidence level of 95%, and the estimated risk percentiles were found to be stable and reproducible.

Fig. 7
Fig. 7
Full size image

Cumulative probabilistic distributions of incremental lifetime cancer risk (ILCR) for dermal exposure in the study region, (a) ILCR of cadmium, the dotted lines represent the ILCR distributions for the maximum allowable concentration (MAC) values recommended by MEWA42, USEPA44, and Health Canada45, and (b) ILCR of chromium, the dotted lines represent the ILCR distributions for MAC values recommended by MEWA42, USEPA30, and Health Canada8 as shown in Table 1.

Figure 8 illustrates the sensitivity analysis results of probabilistic non-cancer risk assessment in the form of percentage contribution to variance for the most susceptible age groups (infants and adults). Figure 8a highlights Zn (59.5%) and Cr (18.1%) concentrations in farm wells as the top two positive contributors to non-cancer risk variation for infants. Monte Carlo simulations recognize exposure time with a 10.5% positive contribution and body weight with a 1.4% negative contribution to variation in non-cancer risk in infants as the third and fourth important risk factors. Conversely, Fig. 8b highlights Cr concentration and exposure time with 31% and 29.4% positive contributions to variation as the top two positive contributors to non-cancer risk variation for infants. Subsequently, Zn concentration with an 8.4% positive and body weight with a 6.5% negative contribution to variation in non-cancer risk in adults can be seen as the third and fourth significant risk factors. Skin surface area and Cd concentration in raw water also showed some positive contributions to the variation in non-cancer risk.

Fig. 8
Fig. 8
Full size image

Sensitivity analysis results show a contribution to variation in the Hazard Index (HIdermal), (a) infants and (b) adults.

Figure 9a and b also identified concentrations of heavy metals (60.1% Cd and 46.2% Cr) and exposure time (12.6% for Cd and 25.8% for Cr) as the highest positive contributors to cancer risk from dermal exposure. Further, the Monte Carlo simulations recognize adult body weight (2.6% for Cd and 5.6% for Cr) as the third major negative contributor to ILCR variance. The skin surface area also contributes positively to ILCR variation to some extent, i.e., 0.9% for Cd and 1.8% for Cr. These findings underscore the need for further site-specific investigations with local data to minimize uncertainties associated with exposure time, body weight, and skin surface area.

Heavy metals pollution indices

Figure 10 demonstrates the probability density functions for heavy metals pollution indices HEI and HPI. HEI is a direct tool for evaluating regulatory compliance, and was found to be 1.5 at the median and 17 at 0.999 probabilities, showing low (< 40) heavy metal pollution levels in the study region. HPI is a more elaborate method for assessing heavy metal pollution as it aggregates the performance scores of heavy metals, considering their relative weights as given in Eqs. (8) and (9). HPI of 17.2 at median, 84 at 0.95, and 100 at 97.5 probability in Fig. 10 indicate non-contaminated groundwater in the study region. Hence, both indices show a low level of heavy metal pollution.

Fig. 9
Fig. 9
Full size image

Sensitivity analysis results show the contributions of various risk factors to variation in the incremental lifetime cancer risk (ILCR) for dermal exposure, (a) cadmium and (b) chromium.

Fig. 10
Fig. 10
Full size image

Heavy metal pollution indices, heavy metal pollution index (HPI), and heavy metal evaluation index (HEI), in the study region.

Discussion

Water quality compliance

The Saq aquifer is the primary source of municipal and agricultural water supply in the Qassim Region69,70. Farm owners and farmers in this region draw groundwater from unconfined, shallow aquifers using wells installed at depths of 80 to 120 m. Heavy metals and minerals leaching from sub-surface soils and geology result in natural groundwater contamination in KSA1. Oral consumption of tap water is almost negligible due to the high TDS levels, with an exceedance probability of 0.86. This improvement in residents’ perception of the health effects of heavy metals in the study region has led to increased dermal exposure to heavy metals through activities such as bathing, showering, ablution, and swimming.

KSA Vision 203071 emphasizes the protection of human health and the conservation of natural resources for the long-term, sustainable use of available water resources. HHRA estimates potential threats to human health resulting from exposure to chemicals of concern in a given region72. Out of thirteen selected heavy metals, Cu, Fe, Pb, Ag, and Li were undetected using the Spectrostometric method. Out of the remaining eight, Ba, Cd, Cr, Mn, and Zn were selected for HHRA due to their potential adverse health effects through dermal exposure. These findings align with past studies on heavy metals of concern in KSA1,73. A low value of HPI in the study area (i.e., central KSA) differs from the findings of Alharbi et al.74 on heavy metal pollution along the Red Sea Coast in northwest KSA, which reported moderate to high HPI. The comparison highlights the importance of site-specific investigations in accurately identifying water quality issues and their corresponding solutions. The study aimed to assess the safety of the exposed population and provide valuable insights into point-of-use treatment technologies for potential oral use at the farm level.

Comparing monitored values of the heavy metals of concern in the study region with the promulgated MAC values suggested by MEWA42 in KSA, the following exceedance probabilities were found: Ba (0.03), Cd (0.003), Cr (0.15), Mn (0.01), and Zn (0.01). Nevertheless, considering the toxicity complexity of heavy metals like Cr, inherent uncertainties in exposure data, and potential to accumulate in human tissues and interact with other toxicants, recent studies suggested probabilistic HHRA for more accurate and practical conclusions75,76,77.

Health risk severity

The present study found a hazard index with a 0.99 probability of less than 1. Hu et al.9,10 state that a hazard index between 0.2 and 0.9 is acceptable, while an HI greater than 0.9 is considered high. Even considering this more conservative approach, hazard indices for all exposure groups were found to be below 0.9 with a 0.99 probability, suggesting an insignificant non-cancer risk through dermal exposure in the study region.

The cancer risk through dermal exposure to Cd was found to be less than the acceptable risk (1E-4) suggested by the USEPA64 at a median probability and 1.6E-4 at a 0.75 probability. At a 0.9 probability, ILCR increased to 3.2E-4 but remained lower than the corresponding estimated ILCR value of 6.9E-4 for a MAC of 0.003, as recommended by MEWA42 for KSA. Likewise, the ILCR estimated the dermal exposure of Cr in raw water drawn from farm wells to be less than 1E-4 at the median probability and 1.4E-4 at a 0.75 probability. Although it increased to 2.2E-04 at a 0.9 probability, it remained lower than the corresponding ILCR value of 2.4E-4 at a MAC of 0.05 mg/L, recommended by MEWA42 for Saudi Arabia. The present study applied the USEPA63 Exposure Factors Handbook for Human Health Risk Assessment. The upper bound values for SA and BW reported in local studies61,62 align with the median probabilities suggested by the USEPA63. As relatively less SA and BW result in higher health risk, ILCR values at the top quarter (1.6E-4 Cd and 1.4E-4 Cr at 0.75 probability) more rationally depict the average health risk in the study area.

Sensitivity analysis results revealed that Cr levels, exposure time, and infants’ body weight were the primary contributors to variations in non-cancer health risks from dermal exposure to the heavy metals of concern. In the context of cancer risk assessment, sensitivity analysis using Monte Carlo simulations identified concentrations of Cd and Cr, exposure time, body weight, and skin surface area as the most significant factors influencing ILCR in adults. These results align with a recent study by Abidola et al.78 on assessing the hazards of heavy metals in water resources.

Household-level treatment

The raw water in the study region can be used for oral consumption using any household-level treatment options, such as filtration, adsorption, membrane filtration, chemical treatment, or bioremediation. Different filtration techniques include granular activated carbon, sand filter (SF), and ceramic filter (CF). The SF can effectively remove Fe and Mn, which are commonly found in the study region, at the household level from contaminated groundwater79. Chemical adsorption, achieved by adding an iron coating, can enhance the efficiency of SF in removing multiple metals80. Although SF is a viable household-level treatment technology for agricultural settings, operational issues associated with regular maintenance (backwashing) and filter media replacement require further investigation supported by end-user education.

Membrane technologies are well-known for efficiently removing heavy metals by selectively separating contaminants considering their size and charge81. High removals of up to 99% have been reported for household-level reverse osmosis (RO) systems82. The simultaneous removal of TDS, pathogens, and heavy metals makes the RO system effective in producing high-quality drinking water. However, periodic maintenance, including membrane replacement and reject management, along with the associated costs, is one of the challenges to implementing RO systems at the household level. Nanofiltration (NF) membranes, which have larger pore sizes than reverse osmosis membranes, can effectively remove heavy metals, such as chromium and cadmium. In addition, NF can remove divalent and trivalent ions (e.g., sulfates) and partially some monovalent ions (e.g., sodium and chlorides), resulting in 80–90% TDS removal83. Operating at lower pressures with reduced energy requirements, lower initial and operating costs, and extended fouling periods are some of the advantages of NF over RO systems84,85.

Recent advancements in composite filters utilizing locally available, simple-to-operate, and low-cost materials such as clay, coconut coir, and activated charcoal have significantly enhanced the efficiency of heavy metal removal from water sources86,87. These composite filters utilize the natural adsorption capacity and porosity of these materials to effectively capture and remove heavy metal ions, such as lead (Pb), arsenic (As), cadmium (Cd), and chromium (Cr), which are commonly found in contaminated groundwater sources, particularly in rural and agricultural regions. Treatment technologies developed from these materials typically operate without the use of chemicals, making them safer and more sustainable88. In the specific context of KSA, where farming communities rely on naturally contaminated groundwater, the implementation of these low-cost, locally adapted treatment technologies can play a critical role in protecting public health and ensuring safe water access in these areas.

It is essential to note that groundwater in Saudi Arabia often contains substantial levels of total dissolved solids (TDS), which can cause taste problems and must be taken into account when selecting suitable water treatment methods. While raw groundwater in the study area is generally suitable for non-potable uses, such as showering, bathing, and swimming, farm owners and workers seek access to affordable and energy-efficient household-level solutions to ensure safe drinking water on their farms. Considering the specific water quality of the raw water in the study area, selecting a suitable treatment method at the household level involves several critical considerations. Farm owners may find the following four scenarios of raw water quality through a detailed physical and chemical examination of their wells:

  1. i)

    TDS < 600 mg/L and all heavy metals well below the MAC recommended by MEWA (2022),

  2. ii)

    TDS < 600 mg/L and some heavy metals of concern.

  3. iii)

    TDS between 600 and 1000 mg/L and heavy metals of concern, and.

  4. iv)

    TDS > 1000 mg/L and heavy metals of concern.

In the first case, farmers can use raw water for drinking after having a water quality analysis from a recognized laboratory. After laboratory examination, low-cost ceramic filtration is suitable for the second scenario, while nanofiltration is more suitable for the third scenario. RO membranes would be ideal for the last scenario, where groundwater is naturally contaminated. Table 5 presents a qualitative analysis of raw water quality scenarios with possible treatment types for ingestion in agricultural settings in KSA.

Table 5 Raw water quality scenarios in agricultural settings in KSA and possible treatment options for ingestion.
Full size table

The users can compare the market prices of these systems with those of bottled water and select the viable solution depending on several factors, such as the number of regular household members (possibly one or two workers in a small farm), the frequency of large and lavish gatherings at the large farmhouse, and other relevant considerations. Given these considerations, it is strongly recommended that further research and development be directed toward identifying and field-testing the most suitable water treatment methods for agricultural farms in Saudi Arabia, as well as optimizing the field conditions to enhance their effectiveness.

Limitations

The study has some limitations, including the use of the USEPA63 distribution for skin surface area, which may result in overestimated risk levels for all exposure groups. For instance, the range of SA for newborns to one-year-old Saudi babies reported by El Edelbi et al.62 varies between 2100 cm² and 3700 cm². In contrast, the USEPA’s Exposure Factors Handbook reports a range from 2400 cm² to 5400 cm² for the same age group. As the ADD (Eq. 1) decreases with an increase in body weight, an underestimated risk can be expected due to the use of higher body weights provided by the USEPA. For instance, the range of body weights for adults (men and women) reported by Khalid and Ali61 varies between 48.8 kg and 78.2 kg (mean ± standard deviation).

Nevertheless, body weight exceeding 110 kg has also been reported in studies on obesity in KSA89. In contrast, the USEPA (2011) recommends a range of 53.4 kg (5th percentile) to 117 kg (95th percentile), with a mean of ~ 80 kg, for individuals aged 20 to 80 years. These results indicate higher body weights in the USEPA63 population than in the Saudi Arabian population, likely due to the taller height of the typical person in the USA. However, as body weight has an inverse effect on ADD, considering higher body weights somehow normalizes the overestimated risk due to the larger skin area used. These limitations can only be minimized by developing SA and BW for infants (0–6 months), toddlers (7 months to 4 years), and all other population groups to ensure an accurate assessment of health risks in KSA. A limited number of samples may have also underestimated or overestimated the non-cancer or cancer risk levels in the study region. The present study utilized USEPA63 information for probabilistic risk assessment, as required details on surface area, body weight, and exposure time were not available. Future studies should develop local demographic data to facilitate more accurate risk assessment in KSA. Additional sampling points and frequent sampling to capture possible seasonal variations in raw water quality will enhance the reliability of the health risk modeling.

Conclusions

In agricultural settings in KSA, elevated salinity in unconfined aquifers limits residents’ use of groundwater for drinking, leaving dermal exposure through washing, bathing, and swimming as a pathway of potential risk. Probabilistic health risk assessment effectively addressed the uncertainties associated with limited groundwater monitoring and vague local information on exposure factors such as skin surface area and body weight. Monitoring results revealed negligible levels of Cu, Fe, Pb, Ag, and TI, while Ba, Cd, Cr, Mn, and Zn levels exceeded the MAC with probabilities of 0.03, 0.003, 0.15, 0.01, and 0.01, respectively.

HI remained below 1 with a 0.99 probability for all exposure groups, including infants, toddlers, children, teens, and adults. The ILCR for Cd exposure through the dermal route remained lower than 1E-4 at median probability, 2E-4 at a 0.75 probability, and 3.2E-4 at a 0.9 probability. Likewise, Cr exposure remained below 1E-4 at the median, 1.4E-4 at a 0.75 probability, and 2.2E-04 at a 0.9 probability. At median probability, ILCRs remained lower than the corresponding values of 4E-4 for Cd and 1.2E-4 for Cr estimated for the MAC values recommended for KSA. At all probabilities, cancer risk remained lower than the calculated ILCRs for the MAC values suggested by USEPA and Health Canada. Sensitivity analysis identified Zn and Cr concentrations, exposure time, and body weights of infants and adults as the primary contributors to variations in non-cancer risk, while Cd and Cr levels, exposure time, body weight, and skin surface were identified as the primary contributors to variations in cancer risk. Both the heavy metal pollution indices (HPI and HEI) indicate a low level of heavy metal pollution with 0.95 probability.

Although the study provides insights into the health risks associated with dermal exposure, future studies can further minimize uncertainties in the health risk assessment process by developing probability distributions for exposure factors that account for the Saudi Arabian population. Additionally, future research can investigate the impact of heavy metals on crops and soil, as well as related ecological risk. This study presents a detailed dermal risk assessment methodology, highlights uncertainties in exposure parameters, and establishes basic guidelines for treating raw water at the household level to ensure the provision of safe drinking water, thereby promoting long-term agricultural sustainability in Saudi Arabia.