Assessment of potentially toxic elements in vegetables cultivated in urban and peri-urban sites in 1 the Kurdistan region of Iraq and implications for human health 2

10 Vegetable fields in and around urban areas in the Kurdistan region of Iraq may have higher than 11 background concentrations of potentially toxic elements (PTEs) from contamination sources including 12 municipal waste disposal and waste water used for irrigation. The purpose of this study was to assess PTE 13 concentrations in soils and the edible parts of field-grown vegetables to quantify potential health risks to 14 the local population. In this survey, 174 soils and 26 different vegetable and fruit types were sampled from 15 15 areas around Sulaymaniyah and Halabja cities. Sampling was undertaken from fields in urban, peri- 16 urban and rural locations including sites close to areas of waste disposal. 17 The soils are calcareous (pH 7.67 - 8.21) and classified as silty loam, sandy or silty clay with organic matter 18 content between 6.62 and 11.4%. Concentrations of PTEs were typically higher in waste disposal areas 19 compared with urban, peri-urban and rural areas. Pollution load indices (PLI) suggested that agricultural 20 soils near waste disposal sites were contaminated with some trace elements. Potentially toxic element 21 concentrations in vegetables were highly variable. Higher total concentrations of PTEs were measured in 22 vegetables from the waste areas with decreasing concentrations in urban, peri-urban and rural areas. Risks 23 to human health were assessed using hazard quotients (HQ). Vegetable consumption poses no risk for 24 adults whereas children might be exposed to Ni, As and Cd. Although HQs suggest elevated risk for 25 children from consumption of some vegetables, these risks are likely to


Introduction
distilled water and shaking at 40 rpm for 30 minutes. Measurements were made using a Hanna pH-209 pH 79 meter with combined glass electrode (Ag/AgCl; PHE 1004) calibrated at pH 7.0 and 4.01, allowing 5 minutes 80 for the reading to stabilize. Loss on ignition (%LOI) was used to estimate the percentage of organic matter 81 in the samples. A known weight of <2 mm oven-dried soil in a pre-weighed ceramic crucible was placed in 82 a muffle furnace (Gallenkamp) overnight at 550 o C, to combust organic matter. The crucibles and 83 combusted soil were then placed in a desiccator to cool before weighing and calculation of %LOI. A 84 portion of <2 mm sieved and homogenised soil was finely ground using an agate ball mill (Retsch Model 85 PM400, Germany) before digestion with 70% hydrofluoric acid, nitric acid and perchloric acid (Trace 86 Element Grade (TEG), Fisher Scientific, UK) in a teflon-coated graphite block digester (Analysco, UK) using 87 PFA digestion vessels. The digested samples were diluted to 50 mL using MilliQ water and stored in PTFE 88 bottles (5% HNO3) pending elemental analysis. All digests were diluted 1:10 with MilliQ water immediately 89 prior to analysis. 90 The fresh weight (FW) of plant samples was recorded as soon as possible after sampling before 91 approximately half of each sample was washed with tap water and then thoroughly rinsed in distilled water 92 to remove surface soil contamination. The remaining plant material was left unwashed. Washed and 93 unwashed portions were oven dried at 70°C for 72 h and re-weighed to determine dry weight (DW). 94 Samples were finely ground in an ultra-centrifugal mill (Retsch Model ZM200, Germany) fitted with a 0.5 95 mm titanium screen. Ground material (200 mg) was digested in pressurised PFA vessels in 6.0 mL of 70% 96 Trace Analysis Grade (TAG, Thermo-Fisher Scientific, UK) HNO3 with microwave heating (Anton Paar 97 'Multiwave'). Digested samples were diluted to 20 mL using MilliQ water and stored pending elemental 98 analysis. Immediately before analysis, samples were diluted 1:10 with MilliQ water. 99 Elemental Analysis 100 Elemental analysis was undertaken using an ICP-MS (Model X-Series II, Thermo-Fisher Scientific) operated 101 in 'collision cell mode' (7% hydrogen in helium) to reduce polyatomic interferences. Samples were 102 introduced from an autosampler (Cetac ASX-520) through a concentric glass venturi nebuliser (flowrate c. 103 0.8 mL min -1 ). Internal standards were introduced to the sample stream via a T-piece and included Rh (20 104 ng mL -1 ) and Ir (10 ng mL -1 ) in 2% TAG HNO3. External multi-element calibration standards (Claritas-PPT 105 grade CLMS-2, Certiprep/Fisher) were used for calibration. Sample processing was undertaken using  (HVO4 2-), vanadate  ions are unlikely to be taken up by plants (Joy et al., 2015).

Risk assessment from PTEs in vegetables 132
Hazard quotients to assess risks to human health from consumption of locally grown produce were 133 calculated using the USEPA approach (USEPA, 2000): where HQM is the hazard quotient for a given element, RfD (mg kg -1 d -1 ) is the reference dose, defined as 136 the maximum tolerable daily ingestion of an element that has no adverse health impact. ADDM (mg kg -1 d -1 ) 137 is the average daily intake of the element: where DI represents the daily intake of fruit and vegetables (kg d -1 ), CFW is the concentration of an element 140 in the edible plant material (mg kg -1 FW) and WB is human body weight (kg). Assumed daily intakes (DI) 141 were 0.342 kg d -1 FW for adults and 0.232 kg d -1 FW for children with body weights of 70 and 16.2 kg, 142 respectively (Hamad et al., 2014). Reference doses were 0.003, 0.012, 0.04, 0.3, 0.0003, 0.00036 and 143 0.0035 mg kg -1 day -1 for Cr, Ni, Cu, Zn, As, Cd and Pb, respectively (WHO, 1982;US EPA Iris Database, 2009;144 Environment Agency, 2009a, Environment Agency, 2009b.

Statistical analysis 146
Descriptive statistics are presented as means, medians, standard errors, differences between means, 147 minima and maxima. A two way ANOVA was used to analyse concentration differences between sampling 148 sites. Pearson correlation coefficients and linear regressions were calculated to determine the relationships 149 between total trace elements in soil and edible parts of vegetables. The data were statistically analysed 150 using the statistical packages SPSS 17.0 and MINITAB 16.0. A probability (p) <0.05 was considered to be 151 statistically significant when testing null hypotheses.

Soil characterization 154
Soil properties are given in Table 2. Soil pH typically ranged from 7.67 to 8.21 confirming the calcareous 155 nature of the soils which have previously been classified as silty loams, sandy and silty clays (Rashid, 2010).

156
The highest pH values were recorded at sites irrigated with groundwater, springs and drilled wells in the 157 Anab area of Halabja. Long-term irrigation with waste water has been suggested to decrease soil pH and to 158 increase the organic matter of top soils (Rattan et al., 2005, Xu et al., 2010 although this pH difference 159 probably reflects the underlying geology and variation in soil type. The lowest soil organic matter (SOM) 160 (6.62 and 6.8% LOI) was recorded in fields irrigated with clean water and diluted waste water in the Kalar 161 river side and Anab area; the highest SOM (11.4% LOI) was in fields near the municipal waste disposal site 162 in Sulaymaniyah city, which is irrigated with diluted domestic waste water. Major element chemistry 163 reflects the calcareous geology of the area; calcium is the dominant cation present at significantly greater 164 concentrations (45900 -127000 mg kg -1 ) than other cations. Studies in areas of calcareous geology have 165 shown similar major element profiles (Iqbal andShah, 2011, Nazif et al., 2015).

166
No local environmental standards exist for trace elements in soils in Iraq therefore concentrations were 167 compared to UK Soil Guideline Values (SGV), European and WHO guidelines (EU/WHO) (

Trace element concentrations in fresh and waste waters 206
Physico-chemical properties of water samples are given in Table 3. All samples are within the safe pH range 207 recommended for irrigation waters (pH 6-9). The EC of the waters was typically in the range 208 500-600 µS cm -1 with the exception of waste water in the Halabja area (c. 950 µS cm -1 ). Trace element (e.g.

209
Co, Ni, Cu) concentrations are slightly greater in the waste waters compared to the fresh waters but are 210 well below the maximum concentrations recommended for irrigation water (Ayers, 1994) indicating that 211 the use of these waters for irrigation purposes is not likely to result in elevated toxic element 212 concentrations in the soils or plants.

Accumulation of trace elements in vegetables 214
The mean concentrations of PTEs in 26 types of unwashed and washed fruits and vegetables are presented 215 in Table 4. A significant variation between PTEs in all sample types was clear (p<0.05). The greatest 216 concentrations of Al, V, Cr, Fe, Ni, Se, Cd, Pb and U were 4653, 9.43, 14.7, 3392, 10.6, 0.196 was in the unwashed leafy vegetables (leek, celery, purslane, spring onion, tarragon and chard), Table 5.

269
For example mean proportions of Cr and Co derived from soil particles in washed samples were >40% for 270 purslane and tarragon, >60% in leek, celery and chard, and >80% in spring onion. Nickel concentrations 271 attributable to soil particles were typically >30% in washed vegetables. Washing was therefore insufficient 272 to remove all adhering soil particles for many vegetables but probably reflects concentrations a consumer 273 might be exposed to after washing vegetables in the home.