Fuzzy-Based Human Health Risk Assessment for Shallow Groundwater Well Users in Arid Regions

Abstract

The conventional point-estimate human health risk assessment (HHRA) primarily uses average concentrations of a limited number of samples due to the high monitoring costs of heavy metals in groundwater. The results can be erroneous when concentrations significantly deviate from the average across the collected samples in an investigation region. The present research developed a hierarchical fuzzy-based HHRA (F-HHRA) framework to handle variations in limited data sets and subjectively established a broader range of risks for various exposure groups. Groundwater samples from 80 to 120 m deep in shallow wells were collected from agricultural farms along Wadi Rumah in the Qassim Region of Saudi Arabia. Laboratory testing found total dissolved solids much higher than the promulgated drinking water quality standards. As the aftertaste issue eliminated the raw water potability, the study considered dermal exposure for HHRA. The collected samples were tested for thirteen potential heavy metals (HMs), including barium (Ba), boron (B), cadmium (Cd), chromium (Cr), copper (Cu), iron (Fe), lead (Pb), lithium (Li), manganese (Mn), silver (Ag), strontium (Sr), thallium (TI), and zinc (Zn). Cu, Fe, Pb, Ag, and TI were lower than the detectable limit of the inductively coupled plasma mass spectrometry device. Concentrations of the remaining HMs in wastewater outfalls that were much less than the groundwater eradicated the impact of anthropogenic activities and affirmed natural contamination. Apart from 10% of the samples for Mn and 90% of the samples for Sr, all the other HMs remained within the desired maximum allowable concentrations. Point-estimate and fuzzy-based approaches yielded ‘low’ dermal non-cancer risk and cancer risk for all groups other than adults, where dermal cancer risk of Cr remained in the ‘acceptable’ (1 × 10⁻⁶ and 1 × 10⁻⁵) risk zone. Although dermal risk does not require controls, scenario analysis established the rationality of F-HHRA for more contaminated samples. The proposed hierarchical F-HHRA framework will facilitate the decision-makers in concerned agencies to plan risk mitigation strategies (household level and decentralized systems) for shallow well consumers in Saudi Arabia and other arid regions.

1. Introduction

Raised levels of toxic and bioaccumulative heavy metals (HMs) in groundwater, posing cancer and non-cancer risks to consumers, have been frequently reported in arid regions [1,2]. Without creating an objectionable taste or odor to the source water, these metals are more deceptive than pollutants with acute problems. The potential hazards of HMs concern both the ecological environment and human health, owing to toxicity, limited biodegradability, and the ability to accumulate in living organisms [3]. The adverse effects of HMs on human metabolism are proven because they serve as catalytic and structural elements in enzymes and proteins. An overextended duration of exposure to hazardous materials can amass crucial bodily organs, including the liver, brain, kidneys, and bones. The degree of accumulation is contingent upon the particular chemical structure and its components [4,5]. Cutaneous and intestinal absorption are widely acknowledged as the predominant pathways of exposure [6,7].

Certain metallic elements can elicit adverse consequences even when their concentrations fall within the prescribed international guidelines [8]. As toxic water use through dermal, inhalation, and oral routes poses serious health concerns, the human health risk assessment (HHRA) assesses the concentration, distribution, and causes of pollution caused by trace elements and their potentially dangerous consequences on human health [9]. HHRA systematically identifies significant risks, proposes control measures to reduce exposure levels, and attains acceptable levels of risk [10].

Most of the studies assessed human health risks for surface water sources contaminated with anthropogenic activities in the USA [11,12] and Canada [13,14,15]. Other studies on deep aquifers of different regions evaluated heavy metal contamination risk in sub-soils and rocks, for instance, Iran [16,17], India [18,19], Bangladesh [20,21], and Pakistan [10,22]. Some studies in Africa and the Sahara also identified high levels of metals in confined aquifers, conflicting with human consumption [23,24,25]. Several studies appraised the impact of anthropogenic activities on groundwater [26,27,28,29,30], while some specific studies investigated the impact of industrial discharges on shallow groundwater in the river plain [31,32,33]. Deep aquifers primarily supply water to larger urban areas, while shallow wells are a common source of water supply in smaller rural settlements [34]. The past studies in arid regions have reported both cancer and non-cancer risks for raw water drawn from confined aquifers through deep wells [21,35,36]; however, there is no examination of HHRA in shallow wells.

In addition to water scarcity due to limited renewable resources in the Kingdom of Saudi Arabia (KSA), the deep groundwater does not comply with the KSA drinking water standards due to natural contamination from radionuclides, salt, and HMs [37,38,39]. In urban areas of most regions with groundwater reliance, membrane (reverse osmosis) treatment complies with the desired standards by simultaneous metals removal with salts [40]. Rural areas with agricultural lands and farms exist in almost all the provinces of KSA with shallow irrigation wells 80 to 120 m deep supplying water for domestic uses, e.g., washing, flushing, and bathing. As villagers, farmers, and workers do not use shallow well water for drinking, dermal risks from bathing do exist.

The availability of water quality data is the primary reason for HHRA studies for deep groundwater sources, as the water suppliers regularly monitor heavy metals in both raw and treated water for performance evaluation of their treatment facilities. Most of the above studies used USEPA and Health Canada cancer and non-cancer risk assessment guidelines to evaluate each exposure group’s point-estimate assessment approach using average contaminant levels [41,42]. The conventional point-estimate approach does not appropriately consider the extremely high or low concentrations in the area under study. One significantly higher value than the maximum allowable concentration level (MACL) may neutralize several extremely low values while averaging the concentrations, and contrariwise. To handle the issue of ‘central value,’ some studies used a probabilistic risk assessment approach [9,43,44], which needs large data sets and holds inherent complexities in the application. In addition, various uncertainties due to limited samples, sampling inaccuracies, and measurement errors also affect the final assessment results. Fuzzy-based methods have also been integrated with probabilistic methods for HHRA to aggregate aggregated water quality scores [9,41].

The present study developed an HHRA framework for smaller settlements in arid regions that use water from shallow wells for washing and bathing. The first study aimed to calculate the dermal risk for groups exposed to eight heavy metals monitored in samples of shallow wells, using USEPA and Health Canada human health risk assessment guidelines for point-estimate HHRA. The second objective was to develop a fuzzy-based hierarchical risk assessment approach to consider the entire range (lower, equal, or higher than the allowable concentrations) of estimated risk levels for different exposure groups (e.g., infant and teen) and risk type (dermal) at different levels of the hierarchy. The third objective was to compare the point-estimate human health risk assessment (P-HHRA) results with fuzzy-based human health risk assessment (F-HHRA) to evaluate the impact of allocating relative importance to the frequency and severity of the measured heavy metals. The proposed framework was applied to the rural areas with agricultural farms and orchids in the Qassim Regions of KSA.

2. Materials and Methods

2.1. Case Study Area

The study area encompassed four important cities, Buraydah, Unayzah, AlRass, and AlBukayriah, located along the Wadi Rumah (N25°52′04″, E43°20′57″ to N26°21′25″, E44°10′50″) in the Qassim Region of KSA (Figure 1). The total area spans approximately 80 km² at an elevation between 590 m to 660 m from mean sea level. As the region is known for extensive agricultural activities, several date farms and agricultural lands exist along the Wadi Rumah, primarily due to the easily accessible groundwater at shallow depths ranging between 80 and 120 m. Tertiary-level wastewater treatment facilities are also located along the Wadi, discharging their overflows (mostly in rainy periods) from the four cities after meeting agricultural requirements. Very low concentrations of HMs compared to raw groundwater found in wastewater samples obtained from wastewater outfalls eliminated the impact of anthropogenic activities and testified to the natural source of contamination.

The groundwater in the Qassim Region primarily exists in fossil aquifers, which are not replenished by surface water and are being depleted at an unsustainable rate due to over-extraction for irrigation and other uses [37,45]. The unconfined aquifer meets the agricultural water quality standards for conservative and non-conservative water quality parameters. Past studies reported the presence of HMs in the Saq aquifer, one of the primary confined groundwater sources in Saudi Arabia [46]. HMs in shallow wells aquifers cannot be overlooked as hundreds of villagers and farmers working in farms and agricultural fields are potentially exposed to these metals due to the direct use of groundwater for washing, swimming, and bathing (Figure 1). ArcGIS (Version 10.8) was used to generate all the risk maps of the study area.

2.2. Water Quality Sampling and Analysis

Multiple field trips were conducted to collect the samples from various farms spread across the study area shown in Figure 1; the sampling locations are not shown for confidentiality reasons. The sampling team was equipped with a handheld sampler to draw samples from raw water reservoirs next to shallow wells, which further distribute the pumped groundwater to date orchids and farms through drip or sprinkler irrigation systems. The farmers lift raw water to small tanks installed at the workers’ and owners’ residences for washing and bathing. The samples were collected from the wells’ direct discharge lines to the raw water reservoirs in 1-L glass bottles, stored in an ice box, and transported to the Central Laboratory of the National Water Company (NWC) in the Qassim region for laboratory examination within 4 h of collection.

Inductively coupled plasma mass spectrometry (ICP-MS) (@thermo 6000 ICP-OES) measured the HMs’ concentrations in the collected samples. ICP-MS’s ability to detect low concentrations of HMs is suitable for groundwater samples. Thirteen metals of concern were measured, including barium (Ba), boron (B), cadmium (Cd), chromium (Cr), copper (C), iron (Fe), manganese (Mn), lead (L), lithium (Li), silver (S), strontium (St), thallium (TI), and zinc (Zn). The copper, iron, lead, silver, and thallium levels were lower than the detectable range of ICP-MS and excluded for HHRA.

Table 1. Heavy metals of concern with non-cancer health risks in the study area.

No.	Heavy Metal	Symbol	Maximum Allowable Concentration (MAC) (ppb)
			KSA 1	USEPA 2	Health Canada 3
1	Boron	B	2.4	-	5.0
2	Barium	Ba	1.3	2.0	2.0
3	Cadmium	Cd	0.003	0.005	0.007
4	Chromium	Cr	0.05	0.1	0.05
5	Lithium	Li	-	-	-
6	Manganese	Mn	0.4	-	0.12
7	Strontium	Sr	-	-	7.0
8	Zinc	Zn	3.0	5.0	5.0

¹ MEWA (2022) [49], ² USEPA (2011) [50], ³ Health Canada (2012) [51].

lists the HMs of concern in the study area and the maximum allowable concentrations (MAC) established by the Ministry of Environment, Water, and Agriculture (MEWA) Health Canada and the United States Environmental Protection Agency (USEPA), along with the associated risk type. According to the toxicological profile prepared by the Agency for Toxic Substances and Disease Registry (ATSDR), dermal exposure to B, Ba, Li, and St can cause skin irritation and sensitization, while chronic exposure can lead to systemic toxicity, including reproductive and developmental effects and possibly cancer. Moreover, oral exposure to these heavy metals can cause gastrointestinal irritation, vomiting, and diarrhea, leading to kidney damage, reproductive and developmental effects, and possibly cancer (ATSDR, 2010, 2012, 2020). The presence of Li in drinking water may interfere with the thyroid. Although there are no set MAC, Li levels in drinking water should be less than 700 ppb as per USEPA [47]. Cd and Cr are established carcinogens and pose various health risks through dermal and oral exposures. Exposure to Cd can damage kidneys and lungs, increasing the risk of lung cancer, while exposure to Cr can cause skin irritation, lung cancer, and other respiratory problems. The International Agency for Research on Cancer (IARC) classified Cd as a Group 1 carcinogen, meaning it is known to be carcinogenic to humans, while Cr was classified as a Group 1, 2A, or 2B carcinogen depending on its chemical form [48]. Notably, due to high TDS levels, the residents do not drink groundwater; therefore, HHRA in the present study did not include oral exposure.

2.3. Human Health Risk Assessment

Figure 2 outlines the methodology developed in the present study for HHRA of the end-users exposed to naturally contaminated shallow wells with HMs in the study area. In addition to the P-HHRA approach, the present study developed an F-HHR approach to accommodate epistemic (subjective definitions of risk levels) and aleatory (randomness in data) uncertainties. The methodology compares the dermal risk levels by considering the point estimates of HMs’ risk levels and allocating the relative weights to the samples with varying concentrations. Finally, the process culminates in proposing possible risk controls by comparing the P-HHRA and F-HHRA approaches. The details of the methods used are given in the following subsections.

2.4. Point Estimate Human Health Risk Assessment

Human health risk assessment evaluates the potential risks from exposure to chemicals, pathogens, or other hazards. The present study performed the four steps of Health Canada’s Guidance on Human Health Detailed Quantitative Risk Assessment (HDQRA) for chemicals [52] to assess HM risks in the shallow wells in agricultural areas alongside the Wadi Rumah. Known as risk identification, the problem formulation associated with health risk is the first step, consisting of identifying chemicals of concern, expected exposure pathways, and exposed populations or receptors in the area under study. The chemicals of concern in the study area, including B, Ba, Cd, Cr, Li, Mn, Sr, and Zn, were identified through the review of past studies [2] in the region and personal communication with personnel from NWC, followed by their presence in groundwater through laboratory analysis. For instance, Haider [2] carried out a non-cancer risk assessment for deep wells in Buraydah City of Qassim Region and identified Mn, Cr, and Ar as the chemicals of concern. Excluding Cr and Cd, all the chemicals were identified as critical for non-cancer risk per the USEPA and Health Canada HHRA guidelines [51,53].

The receptors (exposed population), including workers, farm owners, and villagers living inside the study area, have not commonly used the drawn water for drinking anymore since realizing the health consequences of contaminated water almost two decades ago. As the use of pumped water for drinking is not observed, the present study considered dermal exposure pathways with the current washing and bathing uses.

Table 2. Recommended values of risk factors for different exposure groups [50,52].

Risk Factors	Units	Infants	Toddlers	Children	Teenagers	Adults
Average time (AT)	day	210	1610	2555	2920	10,950
Exposure duration (ED)	year	0.583	4.417	7	8	30
Exposure frequency (EF)	day/year	365	365	365	365	365
Mean daily water intake b	Liters	0.287	0.340	0.453	0.673	1.090
Time of dermal exposure (ET)	hours/day	0.58	0.58	0.58	0.58	0.58
Skin surface area (SA)	cm2	6030	6640	8660	12,000	18,000
Unit conversion factor (CF)	L/cm3	0.001	0.001	0.001	0.001	0.001
Mean body weight (kg) b	Kg	6.44	14.88	28.15	61.51	80.73
Age range	years	0 to <0.583	0.583 to <5	5 to <12	12 to <20	20 to 80
Fraction for lifetime cancer risk a	-	0.006	0.06	0.09	0.1	0.75

^a: (Health Canada, 2010) Guidance on Human Health Detailed Quantitative Risk Assessment for Chemicals (DQRACHEM), [52]; ^b: Calculated based on the Exposure Factors Handbook (USEPA, 2011) [50]. Integrated Risk Information System (IRIS), maintained by the US Environmental Protection Agency (USEPA).

describes mean body weights, water intake rates, skin areas for each age group (infants, toddlers, children, teenagers, and adults) categorized in HDQRA, and fractions of lifetime cancer risk [9,50,52].

In the second step, known as exposure assessment, the average daily dose (ADD), representing the average dose for dermal exposure over a specific period, was calculated as mg/kg/day, using Equation (1) [50].

$$ADD=\frac{\mathrm{C}\times \mathrm{IR}\times \mathrm{SA}\times \mathrm{Kp}\times \mathrm{ET}\times \mathrm{EF}\times \mathrm{ED}\times \mathrm{CF}}{BW\times AT}$$

(1)

where C is the average concentration of heavy metal in groundwater (mg/L), IR represents the drinking water intake rate (L/day), SA is the surface skin area (cm²), Kp is permeability coefficient (cm/hour), ET is the time of dermal exposure (hours/day), EF is exposure frequency (day/year), CF is unit conversion factor (L/cm³), ED denotes exposure duration (days), BW is body weight (kg), and AT is averaging time (days), which represents the time over which the dose is averaged. The AT value equaled the ED value as the exposure dose was averaged over the exposure duration.

For heavy metals with confirmed/possible human health effects, cancer risk (CR) and non-cancer risk (NCR) through dermal routes were assessed. As a hazard quotient (HQ), NCR was characterized as the ratio of ADD to RfD in Equation (2) [50].

$$HQ=\frac{\mathrm{ADD}}{\mathrm{RfD}}$$

(2)

The non-cancer risk for each exposure group was estimated as hazard index (HI) from the summation of HQs for individual HMs. Finally, an overall HI was calculated.

The cancer effects on humans from exposure to carcinogenic HMs were assessed regarding the incremental lifetime cancer risk (ILCR) [50]. CR across various stages of life was considered as the estimated ADD for each exposure group multiplied by the appropriate cancer risk slope factor (SF) using Equation (3) [50].

$$CR=\mathrm{ADD}\times SF$$

(3)

In the last part of the exposure assessment, Equation (4) aggregated CRs for different life stages using a simple weighting sum approach for estimating the ILCR over lifetime exposure [50].

$$\mathrm{ILCR}={\displaystyle \sum }_i^n{\mathrm{CR}}_{\mathrm{i}}\times {\mathrm{F}}_{\mathrm{i}}$$

(4)

CR_i in Equation (4) represents the estimated cancer risk for each group (i), and F_i is the corresponding fraction over 80 years of lifespan.

Table 2 presents the values of risk factors used for various exposure groups. The toxicity assessment step used reference doses and slope factors recommended by the USEPA Integrated Risk Information System and Health Canada Toxicological Reference Values for cancer and non-cancer risk estimation [50]. USEPA defined 1 in 1,000,000 (1 × 10⁻⁶) to 1 in 10,000 (1 × 10⁻⁴) as the acceptable risk range of cancer risk (ILCR) and HQ less than ‘1′ as an acceptable NCR [54].

Table 3. Slope factor and reference dose values for the selected heavy metals (Source: [50,55]).

Heavy Metals	Symbol	Reference Dose (RfD) (mg/kg/day)	Slope Factor (SF) (mg/kg/day)
Barium	Ba	0.2	-
Boron	B	0.2	-
Cadmium	Cd	0.0005	6.1
Chromium	Cr	0.003	0.5
Lithium	Li	0.002	-
Manganese	Mn	0.14	-
Strontium	St	0.6	-
Zinc	Zn	0.3	-

describes the RfDs and SFs, and

Table 4. Categorization of the CR, NCR, and the risk color coding scheme.

Cancer Risk 1–3		Non-Cancer Risk 1–2		Risk Level Color Coding Scheme
Range	Level	Range	Level
<10−6	Low	<0.2	Low
10−6–10−5	Acceptable	0.2–0.9	Acceptable
>10−5	High	>0.9	High

¹ Hu et al., 2020 [13]; ² Hu et al., 2019b [9]; ³ Sullivan et al., 2005 [54].

categorizes the risk levels used in the present study into subjective scales as low, acceptable, and high.

2.5. Fuzzy-Based Human Health Risk Assessment

F-HHRA addressed the problem related to averaging the spread in data, as illustrated in Figure 3. Figure 3a shows that a few high values can influence the average value in a given data set, while some extremely low values can alter the average value in Figure 3b. The estimated risk using the average values in P-HHRA could be misleading, as the actual risk might be lower than the calculated risk for the situation illustrated in Figure 3a and higher than the estimated risk for Figure 3b. P-HHRA can only yield reliable results if the samples collected from locations (wells) spatially distributed over the study area have consistent sub-soil strata with negligible human and measurement errors, resulting in concentrations not much varied from the mean, as shown in Figure 3c.

The fuzzy synthetic evaluation (FSE) technique combining fuzzy logic and synthetic evaluation was used to perform the F-HHRA. Instead of combining the HQs for NCR calculation and using ILCR for cancer risk assessment, FSE classified the estimated HQs and CRs for each heavy metal into three groups, i.e., low, medium, and high (level 1 in Figure 2). Subsequently, the values were aggregated for each exposure group at level 2 and the dermal route at level 3. Finally, level 4 estimated the exposed population’s NR and NCR risk levels. A five-scale subjective rating representing ‘1’ as ‘low’, 3 as ‘medium’, and 5 as ‘high’ attended the variation in data and uncertainties due to imprecise measurements and spatial variations. The step-by-step procedure developed for FSE, well-rounding the overall risk, is described in the following steps.

• Step 1: Calculate cancer (Rc) and non-cancer (Rnc) risks for each heavy metal

Cancer and non-cancer risks of each heavy metal were calculated and arranged in the set of risk levels E, for both Rc and Rnc, as e1= low; e2= acceptable; e3= high. In the evaluation matrix, rij denotes the degree to which the risk level ej satisfies the heavy metal ‘i’. For example, boron (B) results revealed that 50% of the samples showed low non-cancer risk for the infants’ age group, 50% medium, and 0% high as per the risk ranges given in Table 4. Equation (5) passes the risk matrix for a heavy metal:

$$\frac{0.5}{\mathrm{low}}+\frac{0.5}{\mathrm{acceptable}}+\frac{0}{\mathrm{high}}$$

(5a)

$$=\frac{0.5}{1}+\frac{0.5}{3}+\frac{0.00}{5}$$

(5b)

Equations (6a) and (6b) state the estimated CR and NCR matrix forms.

$${\left({\mathrm{R}}_{\mathrm{i}}^{\mathrm{c}}\right)}_{1\times 3}=\left({\mathrm{r}}_{\mathrm{i}1}^{\mathrm{c}},{\mathrm{r}}_{\mathrm{i}3}^{\mathrm{c}},{\mathrm{r}}_{\mathrm{i}5}^{\mathrm{c}}\right)$$

(6a)

$${\left({\mathrm{R}}_{\mathrm{i}}^{\mathrm{nc}}\right)}_{1\times 3}=\left({\mathrm{r}}_{\mathrm{i}1}^{\mathrm{nc}},{\mathrm{r}}_{\mathrm{i}3}^{\mathrm{nc}},{\mathrm{r}}_{\mathrm{i}5}^{\mathrm{nc}}\right)$$

(6b)

where ${\mathrm{r}}_{\mathrm{ij}}^{\mathrm{c}}$ are estimated ILCR and ${\mathrm{r}}_{\mathrm{ij}}^{\mathrm{nc}}$ are estimated HQ values distributed into “low”, “medium”, and “high”, as defined in Table 4.

Subsequently, Equations (7a) and (7b) calculated the CR and NCR of each heavy metal $\mathrm{i}$ (i = 1, 2, …, k).

$${\mathrm{Rc}}_{\mathrm{i}}={{\displaystyle \sum }}_{\mathrm{j}=1}^5\left({\mathrm{s}}_{\mathrm{j}}\times {\mathrm{r}}_{\mathrm{ij}}^{\mathrm{c}}\right)$$

(7a)

$${\mathrm{Rnc}}_{\mathrm{i}}={{\displaystyle \sum }}_{\mathrm{j}=1}^5\left({\mathrm{s}}_{\mathrm{j}}\times {\mathrm{r}}_{\mathrm{ij}}^{\mathrm{nc}}\right)$$

(7b)

where ${\mathrm{s}}_{\mathrm{j}}$ represents the rating given to factor $\mathrm{i}$, namely, ${\mathrm{s}}_{\mathrm{j}}=1,3,$5; ${\mathrm{Rc}}_{\mathrm{i}}$ represents the ILCR of each heavy metal for each exposure group (e.g., ILCR due to Cr for infants); and ${\mathrm{Rnc}}_{\mathrm{i}}$ is each heavy metal’s HQ for each exposure group (e.g., HQ due to B for teens). It is important to note that Equations (6a) and (7a) were used for Cr and Cd only.

• Step 2: Calculate cancer and non-cancer risk for each exposure group

To calculate the cancer and non-cancer risks of dermal exposure at level 2 of the FSE hierarchy shown in Figure 2, the relative weights of each heavy metal estimated at Level 1, for both CR and NCR, within each exposure group, $\mathrm{W}=\{{\mathrm{w}}_1$, ${\mathrm{w}}_2,\dots {\mathrm{w}}_{\mathrm{k}}\}$, were determined. With $‘\mathrm{k}’$ as the number of heavy metals within an exposure group, the weights assigned to cancer and non-cancer risk of chemical $\mathrm{i}$ were calculated as

$${\mathrm{w}}_{\mathrm{i}}^{\mathrm{c}}={\mathrm{R}}_{\mathrm{i}}^{\mathrm{c}}/{{{\displaystyle \sum }}^{ }}_{\mathrm{i}=1}^{\mathrm{k}}{\mathrm{R}}_{\mathrm{i}}^{\mathrm{c}}$$

(8a)

$${\mathrm{w}}_{\mathrm{i}}^{\mathrm{nc}}={\mathrm{R}}_{\mathrm{i}}^{\mathrm{nc}}/{{{\displaystyle \sum }}^{ }}_{\mathrm{i}=1}^{\mathrm{k}}{\mathrm{R}}_{\mathrm{i}}^{\mathrm{nc}}$$

(8b)

In FSE, the evaluation results were obtained by calculating the fuzzy composition of the weight vector $\mathrm{W}$ and the evaluation matrix $\mathrm{R}$, namely, $\mathrm{D}=\mathrm{W}\times \mathrm{R}$. Thus, the membership functions of exposure group (t) were calculated using Equations (9) and (10):

$${\mathrm{d}}_{\mathrm{tj}}^{\mathrm{c}}={{{\displaystyle \sum }}^{ }}_{\mathrm{i}=1}^{\mathrm{k}}{\mathrm{w}}_{\mathrm{i}}^{\mathrm{c}}\times {\mathrm{r}}_{\mathrm{ij}}^{\mathrm{c}}$$

(9a)

$${\mathrm{d}}_{\mathrm{tj}}^{\mathrm{nc}}={{\displaystyle \sum }}_{\mathrm{i}=1}^{\mathrm{k}}{\mathrm{w}}_{\mathrm{i}}^{\mathrm{nc}}\times {\mathrm{r}}_{\mathrm{ij}}^{\mathrm{c}}$$

(9b)

$${\left({\mathrm{D}}_{\mathrm{t}}^{\mathrm{c}}\right)}_{1\times 3}={\left({\mathrm{W}}_{\mathrm{i}}^{\mathrm{c}}\right)}_{1\times \mathrm{k}}\times {\left({\mathrm{R}}_{\mathrm{i}}^{\mathrm{c}}\right)}_{\mathrm{k}\times 3}=\left({\mathrm{d}}_{\mathrm{t}1}^{\mathrm{c}},{\mathrm{d}}_{\mathrm{t}3}^{\mathrm{c}},{\mathrm{d}}_{\mathrm{t}5}^{\mathrm{c}}\right)$$

(10a)

$${\left({\mathrm{D}}_{\mathrm{t}}^{\mathrm{nc}}\right)}_{1\times 3}={\left({\mathrm{W}}_{\mathrm{i}}^{\mathrm{nc}}\right)}_{1\times \mathrm{k}}\times {\left({\mathrm{R}}_{\mathrm{i}}^{\mathrm{nc}}\right)}_{\mathrm{k}\times 3}=\left({\mathrm{d}}_{\mathrm{t}1}^{\mathrm{nc}},{\mathrm{d}}_{\mathrm{t}3}^{\mathrm{nc}},{\mathrm{d}}_{\mathrm{t}5}^{\mathrm{nc}}\right)$$

(10b)

where k denotes the number of heavy metals, and k = 2 in Equation (10a) and k = 8 in Equation (10b).

With the membership functions, the cancer and non-cancer risks of each exposure group $‘\mathrm{t}’$ (t = 1, 2, …, l) were estimated using Equation (11):

$${\mathrm{Rc}}_{\mathrm{Gt}}={{\displaystyle \sum }}_{\mathrm{j}=1}^3\left({\mathrm{s}}_{\mathrm{j}}\times {\mathrm{d}}_{\mathrm{tj}}^{\mathrm{c}}\right)$$

(11a)

$${\mathrm{Rnc}}_{\mathrm{Gt}}={{\displaystyle \sum }}_{\mathrm{j}=1}^3\left({\mathrm{s}}_{\mathrm{j}}\times {\mathrm{r}}_{\mathrm{tj}}^{\mathrm{nc}}\right)$$

(11b)

• Step 3: Calculate cancer and non-cancer risk for dermal exposure route

Level 3 calculated the cancer and non-cancer risks for the dermal route. With $\mathrm{q}$ as the number of exposure routes (q = 1), the weights of the exposure group ‘t’, ${\mathrm{W}}_{\mathrm{t}}=\left\{{\mathrm{w}}_1,{\mathrm{w}}_2,\dots {\mathrm{w}}_{\mathrm{l}}\right\}$, were determined to estimate cancer and non-cancer risk levels for the dermal exposure route using Equation (12):

$${\mathrm{w}}_{\mathrm{Gt}}^{\mathrm{c}}=\frac{{\mathrm{Rc}}_{\mathrm{t}}}{{{\displaystyle \sum }}_{\mathrm{t}=1}^{\mathrm{l}}{\mathrm{Rc}}_{\mathrm{t}}}$$

(12a)

$${\mathrm{w}}_{\mathrm{Gt}}^{\mathrm{nc}}=\frac{{\mathrm{Rnc}}_{\mathrm{t}}}{{{\displaystyle \sum }}_{\mathrm{t}=1}^{\mathrm{l}}{\mathrm{Rnc}}_{\mathrm{t}}}$$

(12b)

where ${{{\displaystyle \sum }}^{ }}_{\mathrm{t}=1}^{\mathrm{l}}{\mathrm{Rc}}_{\mathrm{i}}$ denotes the sum of cancer risk, and ${{{\displaystyle \sum }}^{ }}_{\mathrm{t}=1}^{\mathrm{l}}{\mathrm{Rnc}}_{\mathrm{i}}$ denotes the sum of non-cancer risk for $‘\mathrm{l}’$ (l = 5) exposure groups. The membership functions of exposure route q’ were calculated using Equations (13) and (14):

$${\mathrm{d}}_{\mathrm{Elj}}^{\mathrm{c}}={{\displaystyle \sum }}_{\mathrm{t}=1}^{\mathrm{q}}{\mathrm{w}}_{\mathrm{Gt}}^{\mathrm{c}}\times {\mathrm{d}}_{\mathrm{tj}}^{\mathrm{c}}$$

(13a)

$${\mathrm{d}}_{\mathrm{Elj}}^{\mathrm{nc}}={{\displaystyle \sum }}_{\mathrm{t}=1}^{\mathrm{q}}{\mathrm{w}}_{\mathrm{Gt}}^{\mathrm{nc}}\times {\mathrm{d}}_{\mathrm{tj}}^{\mathrm{nc}}$$

(13b)

and

$${\left({\mathrm{D}}_{\mathrm{Elj}}^{\mathrm{c}}\right)}_{1\times 3}={\left({\mathrm{W}}_{\mathrm{G}}^{\mathrm{c}}\right)}_{1\times \mathrm{q}}\times {\left({\mathrm{D}}_{\mathrm{G}}^{\mathrm{c}}\right)}_{\mathrm{q}\times 3}=\left({\mathrm{d}}_{\mathrm{Elj}1}^{\mathrm{c}},{\mathrm{d}}_{\mathrm{Elj}2}^{\mathrm{c}},{\mathrm{d}}_{\mathrm{Elj}3}^{\mathrm{c}}\right)$$

(14a)

$${\left({\mathrm{D}}_{\mathrm{Elj}}^{\mathrm{nc}}\right)}_{1\times 3}={\left({\mathrm{W}}_{\mathrm{G}}^{\mathrm{nc}}\right)}_{1\times \mathrm{q}}\times {\left({\mathrm{D}}_{\mathrm{G}}^{\mathrm{nc}}\right)}_{\mathrm{q}\times 3}=\left({\mathrm{d}}_{\mathrm{Eljl}1}^{\mathrm{nc}},{\mathrm{d}}_{\mathrm{Elj}2}^{\mathrm{nc}},{\mathrm{d}}_{\mathrm{Elj}3}^{\mathrm{nc}}\right)$$

(14b)

where ${\left({\mathrm{D}}_{\mathrm{G}}^{\mathrm{c}}\right)}_{\mathrm{q}\times 3}$ and ${\left({\mathrm{D}}_{\mathrm{G}}^{\mathrm{nc}}\right)}_{\mathrm{q}\times 3}$ are $\mathrm{q}\times 3$ matrices that contain $\mathrm{q}$ matrices of ${\left({\mathrm{D}}_{\mathrm{t}}^{\mathrm{c}}\right)}_{1\times 3}$ and ${\left({\mathrm{D}}_{\mathrm{t}}^{\mathrm{nc}}\right)}_{1\times 3}$, respectively.

With the membership functions, the cancer and non-cancer risks of dermal exposure route $‘\mathrm{q}’$ were estimated using Equation (15):

$${\mathrm{Rc}}_{\mathrm{Eqj}}={{{\displaystyle \sum }}^{ }}_{\mathrm{j}=1}^5\left({\mathrm{s}}_{\mathrm{j}}\times {\mathrm{d}}_{\mathrm{Elj}}^{\mathrm{c}}\right)$$

(15a)

$${\mathrm{c}}_{\mathrm{Eqj}}={{{\displaystyle \sum }}^{ }}_{\mathrm{j}=1}^5\left({\mathrm{s}}_{\mathrm{j}}\times {\mathrm{d}}_{\mathrm{Elj}}^{\mathrm{nc}}\right),\mathrm{where}{\mathrm{s}}_{\mathrm{j}}=1,3,5.$$

(15b)

A risk score of fewer than 1 indicates ‘low’ human health risk, from 1 to ≤2 ‘acceptable’, a risk score from 2 to ≤3 corresponds to ‘moderately acceptable’ risk, 3 to ≤4 defines ‘moderately high’ risk, and 4 and higher represents ‘high’ risk.

3. Results

3.1. Water Quality Monitoring

Groundwater samples collected from the reservoirs, receiving raw water from shallow wells in the area under study, were analyzed for thirteen HMs. In addition, pH, temperature, dissolved oxygen (DO), total suspended solids (TSS), total dissolved solids (TDS), chlorine, and total coliforms were also analyzed to determine the overall nature of the groundwater quality. The pH in all the samples ranged between 6.7 and 8.2, within the KSA’s regularity drinking water quality standards range of 6.5–8.5 [56]. Depending on the velocity in the delivery line of the wells discharging groundwater in the raw water reservoir, DO varied from 2.6 mg/L to 8.5 mg/L. TSS remained less than 1 mg/L in all the samples, which resulted in turbidity lower than 1 NTU.

Conversely, the average TDS concentration of over 2000 mg/L violates the WHO guideline value of 500 mg/L and the KSA’s drinking water quality standards of 1000 mg/L [47]. As the raw water has a salty aftertaste, the hypothesis of not using raw water for drinking, as established in the methodology section, is established. Although chlorine stayed absent, the presence of total coliforms in one sample affirms the non-potability of the raw water.

Figure 4 presents the HMs monitoring results for eleven samples from shallow wells spread over the study area. The Figure compares the upper extreme, 90 percentile, median, 10 percentile, and lower extreme values with the MAC values prescribed by USEPA [50] with the help of whisker plots. The Figure also shows the data spread with the respective color shading for each box plot. The results show that Ba levels in all the samples were below the MAC value of 2000 ppb. Although more than 90% of the samples showed Cd levels less than 1.14 ppb, the maximum value of 3.68 ppb in one sample signifies the need for F-HHRA. Cr levels stayed within the MAC, with relatively fewer variations amongst the samples than Cd. Li ranged between 0 and 70 ppb, with a median value of 31.5 ppb showing no significant health impact.

Except for one sample exceeding the MAC value of 50 ppb, the concentration of Mn was well below MAC, with an average value of 12.3 in the remaining samples. The Zn levels (average ~75 ppb) were way lower than the MAC of 5000. Boron remained well below MAC with a maximum of 3878 ppb and a median of 850 ppb. Conversely, Sr, in approximately 90% of samples, exceeded the 4000 ppb MAC with a median concentration of 4163 ppb. The limits for each element represent established health and safety standards set by regulatory agencies in Table 1. Values exceeding or approaching the limits raise concerns about potential health and safety implications and demands for cancer and non-cancer HHRA.

3.2. Point Estimate_HHRA

Figure 5 illustrates the results of P-HHR for cancer risk through dermal exposure to Cr and Cd. Chromium and cadmium are recognized carcinogens, and their presence in the environment may increase the likelihood of developing cancer [50]. The population in the study area does not drink this raw water due to its unpalatable taste from total dissolved solids (TDS) higher than the national standard of 700 mg/L [56]. The results for cancer risk through dermal exposure in Figure 5a show that ILCR is low (<1 in 100,000) for all exposure groups, except in the case of Cr exposure to adults; ILCR slightly exceeds (1.34 × 10⁶) to enter the acceptable risk class. Figure 5b illustrates the P-HHR results for non-cancer risk through dermal routes, which are less hazardous than cancer risk. The Figure demonstrates that the dermal non-cancer calculations are reassuring, indicating that the risk is low for potential human exposure to all the contaminants.

3.3. Fuzzy-Based HHRA

• Step 1: Calculate cancer and non-cancer risk for different heavy metals at level 1

The FSE approach initiated the development of risk matrices for all the HMs using Equations (5) and (6). Consider the case of the dermal cancer risk matrix for chromium (Cr) exposure, where the ILCR for 20% of the samples remained less than 10⁻⁵ (low risk) and for 80% of the samples between 10⁻⁶ and 10⁻⁵ (acceptable risk) (Table 4). Equation (6a) generated the following risk matrix:

$${\left(R_i^c\right)}_{1\times 3}=\left(r_{i1}^c,r_{i3}^c,r_{i5}^c\right)=\left(0.20,0.80,0.00\right)$$

Then, Equation (7a) estimated the overall risk of Cr.

$$R{c}_i={{\displaystyle \sum }}_{j=1}^5\left(s_j\times r_{ij}^c\right)=1\times 0.20+3\times 0.80+5\times 0.00=2.6$$

Similarly, $R{c}_i$ for all the HMs were calculated at level 1 of the hierarchy shown in Figure 2.

Figure 6a,b present all the results of F-HHRA for non-cancer and cancer risk scores for each heavy metal. The F-HHRA results of dermal cancer risk are consistent with the P-HHRA (Figure 5a), where exposure to Cr in adults resulted in a ‘moderately acceptable’ risk and ‘acceptable’ for adults due to Cd exposure. The dermal NCR in Figure 6b remained ‘low’, similar to P-HHRA.

• Step 2: Calculate cancer and non-cancer risks for different exposure groups

To assess the risk for each exposure group at level 2 of the risk assessment hierarchy, Equation (8a) calculated the importance weight of the infant group as

$${\mathrm{w}}_{\mathrm{i}}^{\mathrm{c}}={\mathrm{R}}_{\mathrm{i}}^{\mathrm{c}}/{{\displaystyle \sum }}_{\mathrm{i}=1}^{\mathrm{k}}{\mathrm{R}}_{\mathrm{i}}^{\mathrm{c}}=2.60/\left(2.60+1.20\right)=2.60/3.80=0.68$$

The weight of Cd was 0.32. Subsequently, FSE estimated the membership functions for each exposure group. For example, the calculated membership functions for dermal cancer risk in adults using Equation (10a) are as follows:

$${\left(D_t^c\right)}_{1\times 3}={\left(W_i^c\right)}_{1\times k}\times {\left(R_i^c\right)}_{k\times 3}=\left(d_{t1}^c,d_{t3}^c,d_{t5}^c\right)$$

$$=\left[0.68 0.32\right]\times \left[\begin{array}{ccc}0.20& 0.80& 0.00\\ 0.90& 0.10& 0.00\end{array}\right]$$

$$D_t^c=\left[\begin{array}{ccc}0.424& 0.576& 0.000\end{array}\right]$$

Likewise, the membership functions for all the exposure groups were calculated.

Then, Equation (11a) evaluated the cancer risk for each exposure group $‘t’$ (t = 1, 2, …, l) with known membership functions.

$$R{c}_{Gt}={{\displaystyle \sum }}_{j=1}^3\left(s_j\times r_{tj}^c\right)=1\times 0.424+3\times 0.576+5\times 0.000=2.15$$

Figure 7 illustrates the calculated F-HHRA results by aggregating the risk scores for heavy metals for each exposure group. In the case of the dermal cancer risk assessment, only adults are exposed to a ‘moderately acceptable’ risk, while the risk level is ‘low’ for the remaining exposure groups. All the groups are exposed to ‘low’ non-cancer risk through dermal exposure.

• Step 3: Calculate overall cancer and non-cancer risks for dermal exposure

This step evaluated the overall cancer and non-cancer risks for oral and dermal routes (q) at level 3 of the risk assessment hierarchy shown in Figure 2. Equation (12a) determined the weights of exposure group ‘t’, $W_t=\left\{w_1,w_2,\dots w_l\right\}$, for each exposure route. For instance, the weight of the adult exposure group was found as follows:

$${\mathrm{w}}_{\mathrm{Gt}}^{\mathrm{c}}=\frac{{\mathrm{Rc}}_{\mathrm{t}}}{{{\displaystyle \sum }}_{\mathrm{t}=1}^{\mathrm{l}}{\mathrm{Rc}}_{\mathrm{t}}}=\frac{2.016}{1.00+1.00+1.00+1.00+2.016}=0.350$$

Similarly, the weights of all the exposure groups were calculated. Then, the weights of exposure groups assessed the overall risk score. Equation (14a) obtained the overall dermal cancer risk membership function.

$${\left(D_{\mathrm{mj}}^c\right)}_{1\times 3}={\left(W_E^c\right)}_{1\times q}\times {\left(D_E^c\right)}_{q\times 3}=\left(d_{\mathrm{mj}1}^c,d_{\mathrm{mj}3}^c,d_{\mathrm{mj}5}^c\right)$$

$$=\left[0.16 0.16 0.16 0.16 0.35\right]\times \left[\begin{array}{ccc}1.00& 0.00& 0.00\\ 1.00& 0.00& 0.00\\ 1.00& 0.00& 0.00\\ 1.00& 0.00& 0.00\\ 0.424& 0.576& 0.00\end{array}\right]$$

$$D_{mj}^c=\left[\begin{array}{ccc}0.788& 0.202& 0.000\end{array}\right]$$

Equation (15b) evaluated the final fuzzy-based non-cancer risk.

$$R{c}_{yj}={{{\displaystyle \sum }}^{ }}_{j=1}^5\left(s_j\times d_{mj}^{nc}\right)=1\times 0.788+3\times 0.202+5\times 0.00=1.39$$

Likewise, the dermal cancer risk was ‘acceptable’, and the dermal non-cancer risk was ‘low’. Figure 8a–d compare P-HHRA and F-HHRA results on spatial risk maps for the studied area. Figure 8a,b show an acceptable (more than 1 in 1,000,000) dermal cancer risk, while Figure 8a,b show that the non-cancer risk (Figure 8c,d) is low through dermal exposure.

3.4. Scenario Analysis

The study aimed to evaluate two approaches for HHRA. Figure 8 demonstrates no difference in the two HHRA approaches due to no significant variations in heavy metal levels exceeding the MAC. The scenario analysis appraised the usefulness of the F-HHRA approach for dealing with substantial variations in exposure to varying boron concentrations in infants, the most vulnerable group, in the study area.

Table 5. Scenario analysis evaluating non-cancer risk for infants with varying levels of boron levels in groundwater samples using P-HHRA and F-HHRA.

No.	Scenario	P-HHRA	F-HHRA
S1	In 90% of the samples, boron levels pose low and 10% high non-cancer health risk to infants.	Medium Risk (HQ = 0.2)	Acceptable (1.4)
S2	In 80% of the samples, boron levels pose medium, 10% low, and 10% high non-cancer risk to infants.	Medium risk * (HQ = 0.81)	Moderately acceptable * (3.0)
S3	In 60% of samples, boron levels pose high, 30% medium, and 10% low non-cancer risk to infants.	Medium Risk (HQ = 0.79)	High (4.0)

* As low as reasonably practical (ALARP).

presents the results of the scenarios with the highly variable conditions explained in Figure 3. The results show that the point-estimate HHRA approach resulted in medium risk in all three scenarios of varying boron levels in the collected groundwater samples. On the other hand, the results of the fuzzy-based HHRA approach reflect more rational risk assessment results by giving respective weights to different pollution levels of the chemical of concern in the samples.

4. Discussion

A sustainable groundwater source should be safe and remain available for a long period. Reducing threats to human health, including cancer, and conserving water resources are among the primary objectives of KSA’s Vision 2023 [58]. HHRA is a tool to manage the health threats potentially caused by exposure to various chemicals and other social hazards [59]. The present study used HHRA to appraise the safety of groundwater for dermal exposures. HHRA results provide information regarding health risk and insights about needs and planning of source protection and required treatment strategies. All of these findings affirm the long-term sustainability of the natural resources in arid regions.

The weathering of HMs in sub-soils and rocks results in their presence as trace elements (i.e., levels less than 1000 mg/kg) [60]. In most cases, the heavy metal contamination in the groundwater is natural in KSA [1]. The present study also found negligible levels of heavy metals in the samples collected from the wastewater outfalls (results not presented) and affirms natural contamination as the root cause of concerning levels of heavy metals in shallow groundwater wells.

The present study found no health risk from Ba and Zn because their values were much lower than the MACs in all the collected samples. The dermal cancer risk assessment revealed low values (i.e., less than acceptable) from all the chemicals across the entire range of age groups, excluding Cr risk to adults where an ILCR value of 1.34 × 10⁶ enters the ‘acceptable’ risk range for adults. A past study reported 1 in 100,000 oral non-cancer health risks from Mn and Cr in raw water drawn from deep wells in the Qassim Region; however, the water is treated through reverse osmosis before the municipal supply to ensure public health protection [2]. The findings reporting a ‘low’ non-cancer risk through dermal exposure agree with the present study. Ganiyu et al. [61] reported considerable oral cancer and non-cancer risks from Cd, lead, Zn, Fe, and Mn to infants, children, and adults from drinking shallow (up to 30 m deep) wells in the Southwest of Nigeria. The possibility of oral risk is higher in areas where TDS levels lie within the WHO guideline value of 500 mg/L [62]. The population in the study area is not exposed to oral risk due to the salty aftertaste in raw water.

A low human health risk means no risk reduction is required; generally, a calculated risk of up to 1 in 1000,000 poses a low risk. For an acceptable risk, as low as reasonably practical (ALARP), different organizations established their limits based on socioeconomic analysis. For instance, Health Canada set an acceptable ILCR up to 1 in 100,000 for the British Columbia, Alberta, and the Atlantic provinces [52]. USEPA defined the acceptable ILCR range between 1 × 10⁻⁶ and 1 × 10⁻⁴, and HQ < 1 [54]. These findings open a point of discussion about establishing the ALARP region in risk assessment. Applying fuzzy logic facilitates the decision-makers in the interpretation and subjective definition of HHRA results [9]. The present study used three discrete ranges for P-HHRA: ‘low’, ‘acceptable’, and ‘high. The five ranges used in F-HHRA more effectively cover the ALARP region. Excluding low (<1) and high (>4), the estimated crisp risk levels between 1 and 4 define the ALARP region, leaving a wider scale of risk mitigation planning for the decision-makers. For instance, a ‘moderately acceptable’ risk zone with risk scores between 1 and 2 (acceptable region) and between 2 and 3 (moderately acceptable region terminating ALARP region) directs toward a need for some level of water quality monitoring in the areas with pollution levels close to or higher than MAC.

As per the present findings, the water is safe for washing and bathing. After the ALARP region, scores above 3, moderately high risk, certainly need careful investigations and comprehensive risk reduction, as most of the wells’ sites were contaminated. The responsible agencies, such as municipalities, private water suppliers, Water Directorates, and the National Water Company, need to develop and implement a risk management plan for the affected area in the region. An effective risk management plan should include the following measures. Groundwater resources are concealed and relatively more persevered from anthropogenic activities, and changes in water quality take much more time than in the surface water resources [63]. Regular monitoring and surveillance of the groundwater wells can detect the spatiotemporal changes in the study area and specify the locations with high levels of HMs. Educational campaigns and awareness programs can inform residents in the affected areas about the potential health risks associated with consuming contaminated water [64]. Such programs should provide information on alternative water sources and the importance of using water filters or other treatment methods to reduce heavy metal exposure.

In addition, implementing strict regulations on the disposal of industrial and agricultural waste through regular inspections and penalizing the violators can eliminate the possibility of groundwater contamination from infiltrating polluted surface water [1]. Finally, implementing remediation strategies to reduce the concentration of HMs in the contaminated source, such as chemical precipitation, adsorption, and ion exchange, secure public health at the consumer end [65]. Promoting low-cost household-level treatment can also reduce HM exposure at the point of use. Low-cost water treatment technologies such as ceramic water filters, made of locally available clays and zeolites, can effectively remove Cr, Cd, and Ar at household levels [66].

The maximum allowable concentrations for non-bottled water specified by the MEWA in KSA [49] match the standard values recommended by USEPA [50] and Health Canada [51]. The presence of heavy metals in groundwater sources and the limited studies on HHRA for groundwater sources demand detailed investigations on the topic. The present study is the first effort on HHRA for shallow well users in the KSA. The findings provide valuable insights to concerned organizations, including municipalities, ministries, and research entities, for planning limited groundwater sources in KSA and other African and Arabian Gulf countries with similar water scarcity characteristics.

5. Conclusions

The costly monitoring of heavy metals in water samples generally leads to reliance on fewer samples for human health risk assessment, except for large-scale national and international funding agencies’ projects. The point-estimate HHRA using average concentrations for calculating HQ and ILCRs for different heavy metals may yield erroneous findings, particularly when samples largely deviate from the average. The fuzzy-based HHRA methodology developed in the present study effectively deals with such situations and gives more rational HHRA results with a broader spectrum of human health risks for various exposure groups in a given study area.

Shallow groundwater wells in the Qassim Region of KSA have high total dissolved solids; due to the aftertaste, the consumers use pumped water for non-potable uses, e.g., washing and bathing. The baseline screening of the agricultural farms around the Wadi Rumah catchment highlights the need for assessing cancer and non-cancer health risks through dermal exposure. The point-estimate and fuzzy-based HHRA approaches yielded ‘low’ non-cancer risks through dermal exposure in the study area, except the dermal cancer risk for Cr lay in the ‘acceptable’ risk zone for the adult group.

The present study used the typical values for risk factors, e.g., intake rate, body weight, and skin area, without site-specific data. Future studies can explore the impact of varying risk factors using local numbers through fuzzy or probabilistic-based approaches. The future research can also analyze the effectiveness of different household-level or decentralized treatment strategies in reducing human health risks through heavy metal exposure. Furthermore, studies should investigate the potential synergistic effects of exposure to multiple heavy metals and other environmental contaminants, which may exacerbate the associated health risks. The proposed hierarchical fuzzy-based model offers a robust and flexible framework for HHRA and designing targeted interventions to address heavy metal contamination in groundwater resources in arid regions.