Effect of cultivar type and soil properties on cadmium concentrations in potatoes

ABSTRACT

Potatoes are a staple food in New Zealand and a contributing source of dietary cadmium (Cd), although there is little information nationally on the soil and plant factors that affect Cd concentrations in the crop. We measured Cd concentrations in 10 commercial potato cultivars grown in three field sites across New Zealand and assessed the soil factors that affect Cd concentrations. Cadmium concentrations in potato tubers ranged from 0.004 to 0.574 mg kg⁻¹ DW (0.001–0.113 mg kg⁻¹ FW), although overall Cd concentrations were lower than the maximum limit of 0.1 mg kg⁻¹ FW for human consumption. There were significant differences (2.8-fold) in Cd concentrations between potato cultivars, although these varied by site. No significant relationships were found between soil properties and Cd concentrations in potatoes. Management of Cd in potatoes is likely to be most effective by avoiding growing high-Cd-accumulating cultivars in soils with elevated Cd. Further studies could determine if other management factors such as irrigation or type and rates of fertiliser affect Cd concentrations in potatoes.

KEYWORDS:

• Cadmium
• potato
• cultivar
• guidelines
• New Zealand

Introduction

Cadmium (Cd) is a potentially toxic heavy metal and impurity in many phosphorus (P) fertilisers (Williams and David 1973). As a result of extensive P fertiliser use, Cd has accumulated in some agricultural soils worldwide (McGrath 1986; Johnston et al. 1987; Gray et al. 1999). There is concern regarding the accumulation of Cd in soils because it can readily be taken up by food crops (Schroeder and Balassa 1963; Chaney and Hornick 1978; Zhao et al. 2015; Baldantoni et al. 2016), and then has the potential to enter the human food chain through the crops and crop based products we consume in our diet (Clemens et al. 2013; Rietra et al. 2017).

Along with rice (Oryza saliva L.) and wheat (Triticum aestivum L.), potatoes (Solanum tuberosum L.) are an important staple food in many parts of the world, but also a source of Cd in the human diet (Fan et al. 2009; Rietra et al. 2017). Food standards provide for a safe level of intake over a lifetime. Therefore, while an occasional exceedance presents a low health risk, it could present a trade or reputational risk that ultimately affects the viability of potato production systems in New Zealand. It is therefore important that Cd concentrations in potatoes are maintained below maximum acceptable standards set by regulatory bodies. In response, research has been undertaken to identify which soil and plant factors influence Cd concentrations in potatoes to help manage Cd uptake in this crop (McLaughlin et al. 1994; Maier et al. 2002; Fan et al. 2009). In a survey of Cd concentrations in potatoes across 359 sites across Australia, McLaughlin et al. (1997) reported topsoil chloride concentration were best related to Cd concentrations in potatoes with soil pH, total zinc (Zn) and cultivar types only having a small influence on Cd concentrations. In Sweden, Oborn et al. (1995) found that a combination of organic matter content and increasing soil pH were negatively correlated with Cd concentrations in potatoes across 69 commercial field sites. Al Mamun et al. (2017) also showed the importance of increased soil pH and organic matter on reduction of Cd concentrations in potatoes. In contrast, Tack (2014) reported no relationship between potato Cd concentrations and topsoil pH and organic matter across 21 sites in Belgium. As well as soil factors, significant differences in Cd concentrations between potato cultivars has sometimes been reported (McLaughlin et al. 1994, 1997; Dunbar et al. 2003; Ashrafzadeh et al. 2017). Ozturk et al. (2011) for instance showed Cd concentrations in potato tubers ranged between 0.08 and 0.32 mg kg⁻¹ DW in 16 cultivars from Turkey.

In line with many Western countries and South America, potatoes are a staple food in New Zealand and an important contributory source of dietary Cd (Pearson et al. 2018). Furthermore, potatoes are also a key export crop, worth an estimated $113 M in 2017 (Plant and Food Research 2017). Few studies have measured Cd concentrations in potatoes in New Zealand (e.g. Roberts et al. 1995; Kim 2005; Ashrafzadeh et al. 2017), and there is little information on the effects of soil properties and cultivar differences on Cd concentrations in potatoes across the main commercial growing areas (Pukekohe, Manawatū, Hawkes Bay and Canterbury). This information is important to determine if Cd concentrations in potatoes are currently compliant with national maximum limits (ML) for Cd of 0.1 mg kg⁻¹ FW outlined in the Food Standards of Australia and New Zealand (FSANZ) standard 1.4.1.

In addition, to address the concerns of Cd accumulation in New Zealand soils, a Cadmium Management Strategy (CMS) based on an assessment of the risk Cd poses to agricultural systems has been released (MAF 2011). The objective of the strategy is to ensure that ‘Cd in rural production poses minimal risks to health, trade, land use flexibility and the environment over the next 100 years’. The strategy outlines a risk-based framework for managing Cd accumulation in agricultural soils that focuses on a Tiered Fertiliser Management System (TFMS). The TFMS contains five tiers and four trigger concentrations, whereby as soil Cd concentrations increase, increasingly stringent fertiliser management practices are imposed. The trigger values are however considered only interim (Warne 2011), because there was little New Zealand-specific Cd data available at the time to inform their development. As a result, some overseas soil guideline values were used. It is uncertain however how robust these overseas trigger values are for protecting New Zealand agricultural systems. Total soil Cd concentrations tend to be lower than many other countries (Roberts et al. 1994). However, there have still been occasional exceedances of our food standards for Cd (Gray et al. 2001; Kim 2005). Therefore, for some regions of New Zealand, the trigger values currently in the TFMS may not be sufficiently consistent with the CMS’s objectives outlined above.

The development of New Zealand-specific risk-based trigger values firstly requires New Zealand-specific Cd data in a range of geographic areas, including plant concentrations in important food crops such as potatoes, along with information on the soil and plant factors which control Cd concentrations. The objectives of this study were (i) to determine Cd concentrations in 10 economically important potato cultivars, and (ii) to determine Cd concentrations in two potato cultivars grown across the main commercial potato cropping regions of New Zealand to assess the relationships between soil factors and Cd concentrations in potatoes.

Materials and methods

Sites and sampling

Soil cores and potato samples were collected from a combination of research trials and commercial field sites. In 2015, Cd concentrations were determined in 10 potato cultivars at Lincoln, eight at Pukekohe and two at the Waikato site (Table 1). In 2016 and 2017, Cd concentrations were determined in two commonly used potato cultivars (Moonlight and Innovator) grown across 35 commercial and research sites (Table 1). The sites covered seven of the 15 soil orders in New Zealand (Hewitt 2010) (data not shown).

Table 1. Summary of sites and details of soil and plant sampling.

Year	Sites	Location	No. of cultivars	Sampling details
Potato cultivar study
2015	2 research trial sites	Pukekohe Waikato	8 2	Potato cultivars were grown in 3 replicate plots (double-row, 1.8 × 1.5 m), arranged in a randomised design. Soil and plant sampling were undertaken at harvest. For each plot 5 soil cores (150 mm depth) were collected and combined to form a single composite sample and a single tuber was collected from 5 plants within each plot. In addition a composite soil sample (25 × 150 mm) was taken along a 100 m transect.
	1 research trial sites	Lincoln	10	Potato cultivars were grown in 4 replicate plots (single row, 3.5 × 0.76 m), arranged in a randomised design. Soil and plant sampling were undertaken at harvest. For each plot, 5 soil cores (150 mm depth) were collected and combined to form a single composite soil sample and a single medium sized tuber collected from 5 plants within each plot. In addition a composite soil sample (150 mm depth) was taken along a 100 m transect.
Potato survey
2016 and 2017	35 research trial and commercial sites	Canterbury Pukekohe Waikato Manawatū	2	At each commercial site 3 plots were established randomly across the field. These plots were the width of the planting row, by a length of ∼1.5 m sufficient to encompass a minimum of 5 plants. Within each plot, 5 tubers and 5 soil cores (150 mm depth) were collected At research sites, potatoes were grown and sampled as described above for the 2015 research trials at Pukekohe and Waikato.

Analysis

Soil and plant analysis

Soils were oven-dried (35°C) and sieved (<2 mm) before analysis. Soil pH was determined in a 1:2 soil:water solution by potentiometric analysis (Blakemore et al. 1987). Exchangeable potassium (K), calcium (Ca), magnesium (Mg) and sodium (Na) were measured in a 1 M ammonium acetate extracted at pH 7 (Blakemore et al. 1987) and analysed by Inductively Coupled Plasma Optical Emission Spectrometry (ICP-OES). Cation exchange capacity (CEC) was calculated by summing concentrations of extractable cations and extractable acidity. Bioavailable P (Olsen P) was determined by bicarbonate extraction (Olsen et al. 1954). Total carbon (C) and nitrogen (N) were determined by combustion using an Elementar Vario-Max CN Elemental analyser. Soil chloride concentrations were measured in a filtered 1:5 soil:water extract by ion chromatography. Pseudo total concentrations of Cd, Zn, P, aluminium (Al) and iron (Fe) were determined by microwave digestion using nitric acid and hydrogen peroxide as reported by Simmler et al. (2013); these will be referred to forthwith as total concentrations for simplicity. The digests were analysed by Inductively-Coupled Plasma Mass Spectrometry (ICP-MS) for Cd and ICP-OES for Zn, P, Al and Fe. Particle size analysis was measured using the pipette method (Claydon 1989).

Potato tubers were washed, peeled, roughly chopped, weighed, dried at 60°C to constant weight, and weighed again to enable reporting of data as fresh weight (FW) and dry weight (DW). Dried tubers were ground before digestion using the nitric acid and hydrogen peroxide microwave digestion method described by Cindrić et al. (2015). The digests were analysed for Cd by ICP-MS. In a selection of samples, peel was separated from the potatoes and the peel and flesh analysed separately for Cd and titanium (an indicator of soil contamination) by ICP-MS.

Quality control/assurance

Soil and plant samples were digested in batches of up to 40. Each batch included at least one digestion blank. One out of every 20 samples were analysed in duplicate to confirm repeatability of the analysis. Cadmium concentrations in procedural blanks were less than the detection limit and duplicate analysis of samples were within 5% of each other. The accuracy of soil and plant analysis was assessed using several internal and external certified reference materials including for soil (NIST Montana 2711; Interlab internal WEPAL soil 921; Interlab internal WEPAL soil 981), and plant (NIST 1573a, tomato leaves; ASPAC internal clover; ASPAC internal beetroot). Analytical results were within 5% of the certified values. The method detection limit for soil Cd was 0.020 mg kg⁻¹ and 0.005 mg kg⁻¹ in plant material.

Statistics

Statistical analyses were conducted using tuber Cd concentrations expressed on a dry weight basis. Comparison of Cd concentrations in cultivars within a site and between sites was undertaken separately using one-way ANOVA and data checked for normality by assessing residual plots. The data was not required to be transformed and the effects were considered significant if they differed at the probability level of 5% based on Fisher’s unprotected least significant difference (LSD) test. The relationship between the concentration of Cd in the tubers and soil characteristics was determined by linear and stepwise (forward) multiple regression analyses. All analyses were performed using GENSTAT for Windows v18.

Transfer coefficient

The transfer coefficient (TC) is a very commonly used measure to assess the transfer of inorganic contaminants such as Cd from soil to different plant organs (Alloway 1990). In this study it was the Cd concentration in the potato tuber divided by the total soil Cd concentration Equation (1):(1) $\mathrm{T}\mathrm{C}=\mathrm{C}\mathrm{d}\mathrm{t}\mathrm{u}\mathrm{b}\mathrm{e}\mathrm{r}(\mathrm{m}\mathrm{g}\mathrm{k}{\mathrm{g}}^{-1}\mathrm{D}\mathrm{W}\left)/\mathrm{C}\mathrm{d}\mathrm{s}\mathrm{o}\mathrm{i}\mathrm{l}\right(\mathrm{m}\mathrm{g}\mathrm{k}{\mathrm{g}}^{-1}\mathrm{D}\mathrm{W})$(1)

Equation (1) can be rearranged to indicate the soil concentration at which a specific plant concentration is reached at a given site, assuming conditions such as soil properties (and therefore TC) do not change. Using the food standard as the target plant concentration (Cdtuber limit), a conservative estimate of the soil concentration at which the food standard can be reached (nCdfood standard) can be calculated (Equation 2). As the food standard is expressed as fresh weight, this requires conversion to a Cd concentration based on dry weight. The mean and standard error of the dry weight of potato samples measured in this study was 19.9% (0.3). This is consistent with the dry weight value (20%) which has been used to transform fresh weight to dry weight in potatoes in several other studies (Oporto et al. 2007; Fan et al. 2009; Tack 2014; Ashrafzadeh et al. 2017). A dry weight of 20% was therefore used in Equation 2.(2) $\text{nCdfood standard}\left(\mathrm{m}\mathrm{g}{\mathrm{k}\mathrm{g}}^{-1}\right)={\text{Cdplant limit (mg kg}}^{-1}\text{DW) / TC}$(2) These values are not intended as threshold limits, but rather provide an insight into soil properties influencing plant uptake and the Cd concentrations at which management to mitigate the risk of exceeding food standards may be considered.

Results and discussion

Effect of cultivar type – 2015

Cadmium concentrations in potato tuber cultivars ranged between 0.040 and 0.275 mg kg⁻¹ DW (n = 70) (Table 2), with a mean concentration of 0.135 mg kg⁻¹ DW (0.032 mg kg⁻¹ FW). The mean concentration is similar to that previously reported (0.140 mg kg⁻¹ DW) in a survey of potatoes sold commercially in the Waikato region (largely grown in Pukekohe) (Kim 2005), but higher than an earlier study (0.100 mg kg⁻¹ DW) of potatoes grown at Pukekohe (Roberts et al. 1995). Although, if only considering potatoes from the Pukekohe site, the mean concentration in the present study (0.070 mg kg⁻¹ DW) is lower than both previous studies. Cadmium concentrations are within the range of mean values reported in several overseas studies (0.02–0.40 mg kg⁻¹ DW) (Thomas et al. 1972; Wolnik et al. 1983; Zurera et al. 1987; Oborn et al. 1995; McLaughlin et al. 1997; Karavoltsos et al. 2002; Radwan and Salama 2006; Fan et al. 2009; Luis et al. 2014; Lin et al. 2015; Norton et al. 2015). As reported the moisture content in our potatoes was 80% (w/w). Therefore, the FW Cd concentrations of the tubers ranged from 0.009 to 0.069 mg kg⁻¹. These Cd concentrations are lower than the FSANZ standard 1.4.1 ML for Cd in potatoes of 0.1 mg kg⁻¹ FW, which is the same as the Codex Standard 193–1995 (CAC 2009) and European Commission food standards (EC 1661/2006) for peeled potatoes.

Table 2. Potato tuber Cd concentrations (mg kg⁻¹ DW) in different cultivars at Pukekohe, Waikato and Lincoln.

Cultivar	Pukekohe	Waikato	Lincoln
Agria	0.063abA	0.133aB	0.195abC
Bondi	0.073ab		0.150c
Crop 39	0.053bc		0.200ab
Desiree	0.063ab		0.275d
Ilam Hardy			0.158c
Innovator			0.205a
Moonlight	0.080aA	0.103bB	0.185bC
Nadine	0.113d		0.160c
Russet Burbank	0.070ab		0.123e
Russet Ranger	0.040c		0.115e

Note: Mean followed by the same lowercase letter within a site are not significantly different (P < 0.05) or not significantly different between sites (capital letters).

In some instances, potatoes are consumed without removing the peel. It is therefore important to also know what the Cd concentration is in the potato peel, and the relative mass of the peel to the whole potato. In a subset of samples, the potato peel and the flesh were separated and analysed for Cd. Results showed that in all instances, Cd concentrations in the potato peel were higher (1.3–2.3 fold) than the flesh (Figure 1). This is consistent with what others have previously reported for Cd (Corguinha et al. 2012; Šrek et al. 2012; Norton et al. 2015), and other trace metals such as Zn, Fe, manganese, nickel and copper (Subramanian et al. 2011; Šrek et al. 2012).

Figure 1. Comparison of cadmium concentrations (mg kg⁻¹ DW) in the flesh or peel in potato cultivars from the Lincoln research trial site.

Although the peel contained higher Cd concentrations (0.041–0.080 mg kg⁻¹ FW) than the flesh, it was determined for a range of cultivars (and different sized potatoes), the peel constitutes only 10–17% of the mass of the whole potato (data not shown). As a result, even if potatoes were prepared and eaten without removing the peel, Cd concentrations in the whole potato would still remain compliant with the ML of 0.1 mg kg⁻¹ FW.

There were significant differences in tuber Cd concentrations between cultivars (Table 2). At Pukekohe and Lincoln, Cd concentrations varied 2.8 and 2.4 fold respectively, although the difference was less between the two cultivars grown at the Waikato site. Several studies have previously reported variation in Cd concentrations between potato cultivars (Roberts et al. 1995; McLaughlin et al. 1997; Jönsson and Asp 2011; Corguinha et al. 2012; Luis et al. 2014). In some of these studies, Cd concentrations between cultivars varied by a similar magnitude as found in this study (Dunbar et al. 2003; Mengist et al. 2017), although others have reported a larger range. For instance, Ashrafzadeh et al. (2017) showed Cd concentrations in potato tubers ranged between 0.05 and 0.21 mg kg⁻¹ DW in 10 cultivars grown at a single site in Lincoln, New Zealand.

It is also important to note that while there were differences in Cd concentrations between cultivars, they were often inconsistent. At the Lincoln site for example, Desiree had the highest Cd concentration, but had one of the lowest Cd concentrations at the Pukekohe site. Nadine had the highest Cd concentration at Pukekohe but had one of the lower Cd concentrations at Lincoln. An exception was Russet Ranger, which had the lowest Cd concentration at both Lincoln and Pukekohe. It is not clear at present why there was this inconsistency. A next step would be to collect data on Cd concentrations in these cultivars across a greater number of sites and seasons to determine if this inconsistency is still maintained.

Compared to other crops (i.e. wheat, rice), there have been few studies that have investigated the mechanisms that control the variation in Cd concentrations in different potato cultivars. Dunbar et al. (2003) and Mengist et al. (2018, 2017) both suggest differences are due in part to differential partitioning of Cd between organs (tuber, roots, stem and leaves) within the plant. Mengist et al. (2017) for instance found differences in both the shoot Cd concentration and leaf Cd mobilisation between a low and a high Cd accumulating cultivar. The high Cd accumulator had a higher shoot Cd concentration and remobilisation of Cd from leaves to tubers, thought to be driven by differences in shoot biomass (Mengist et al. 2018). Although, unlike Dunbar et al. (2003), Mengist et al. (2017) also indicate that differences in total plant uptake of Cd are important, with higher uptake in the high Cd accumulating cultivar, although the mechanism was not known. Regardless of the mechanism, the variation in Cd uptake in potatoes highlights the potential to use low Cd accumulating cultivars (e.g. Russet Ranger) at sites where there is high Cd uptake. Additionally, including Cd accumulation as one of the traits in future cultivar evaluation trials, which at present is not common practice (Clemens et al. 2013), alongside more commonly assessed traits such crop yield and disease resistance should be considered.

A comparison of tuber Cd concentrations in two cultivars (Agria and Moonlight) grown at all three sites showed that Cd concentrations were highest at Lincoln, then Waikato and lowest at the Pukekohe site (Table 2). Cadmium concentrations were not related to the total soil Cd concentration, as the Lincoln site had the lowest soil Cd concentration (0.18 mg kg⁻¹), although it did have the lowest pH (6.0), a soil factor known to effect Cd availability in soils (Alloway 1990). The comparatively low Cd concentrations in potatoes at Pukekohe may be a result of the high clay content (59%) and pH (6.5), both properties shown to limit Cd availability in soils (Chaney and Hornick 1978; Tiller et al. 1994). The small number of sites and narrow range of soil properties in the 2015 trials limits any comparison between Cd concentrations and soil properties. As a result, a larger comparison between Cd concentrations and soil properties was undertaken and reported below.

Survey of Cd concentrations in potatoes and soil properties – 2016 and 2017

Soil properties

Total soil Cd concentrations ranged between 0.09 and 0.86 mg kg⁻¹ (Figure 2a). The mean concentration of 0.36 mg kg⁻¹ is slightly higher than the mean Cd concentration of 0.28 mg kg⁻¹ reported for cropping soils in New Zealand (Cavanagh 2014). Most samples (88%) had soil Cd concentrations below the TFMS tier 1 trigger of 0.6 mg kg⁻¹. There were significant differences in total Cd concentrations between regions, with Waikato and Pukekohe higher than Canterbury and Manawatū-Whanganui (Figure 2a), in line with what is observed in larger surveys (Cavanagh 2014). Other soil properties recognised as affecting plant concentrations of Cd also varied and are reported in Table 3. Soil pH ranged from low to high, while total C and CEC were low to medium (Blakemore et al. 1987). The majority of soils were classified as silt loams.

Figure 2. (A) Total soil cadmium concentration (mg kg⁻¹) in each region (MW = Manawatū–Wanganui; Puke = Pukekohe). The dashed line is the TFMS trigger 1 value. (B) Mean cadmium concentrations (mg kg⁻¹ DW) in potatoes for each region. (C) The transfer coefficient for each region. Boxes depict the 25th and 75th percentile values, with horizontal lines plotted within boxes representing the median value. Whiskers show the 10th and 90th percentile values and the points the 5th and 95th percentile values.

Table 3. Selected soil properties from the survey sites (0–15 cm).

Site	pH	Olsen P	Total P	Total C	Total N	K	Mg	Ca	Na	CEC	Total Cd	Total Zn	Cl	Sand	Silt	Clay
		mg kg^−1	mg kg^−1	(%)	(%)	cmol kg^−1	cmol kg^−1	cmol kg^−1	cmol kg^−1	cmol kg^−1	mg kg^−1	mg kg^−1	mg kg^−1	%	%	%
1	6.0	28	n.d.	2.5		0.2	9.3	0.7	0.19	16	0.18	55	51	23	54	23
2	6.0	25	835	1.9	0.2	0.1	6.9	0.5	0.14	13	0.12	72	13	7	70	23
3	5.3	20	604	2.0	0.2	0.2	6.0	0.3	0.16	12	0.14	48	39	35	47	18
4	5.0	44	1031	3.5	0.3	0.5	6.8	0.5	0.20	17	0.19	79	24	11	69	20
5	5.4	36	740	1.6	0.2	0.6	6.5	0.9	0.09	17	0.14	71	12	15	62	23
6	5.7	47	1030	3.3	0.3	0.5	8.2	0.8	0.24	15	0.16	82	38	7	67	26
7	5.4	30	947	2.5	0.3	0.4	6.6	0.7	0.12	14	0.14	79	38	13	62	25
8	5.6	25	819	2.1	0.2	0.4	6.4	0.5	0.13	12	0.13	84	30	10	68	22
9	5.5	29	1074	2.5	0.3	0.8	6.0	0.7	0.09	13	0.16	86	33	13	73	14
10	5.5	50	774	1.3	0.1	0.5	4.2	0.5	0.10	9	0.11	72	58	28	55	17
11	6.1	48	892	2.4	0.2	0.4	6.6	1.3	0.19	13	0.15	90	29	13	66	21
12	5.2	51	1156	2.5	0.3	0.7	9.3	1.5	0.19	18	0.24	96	57	6	67	27
13	5.0	78	1852	7.7	0.6	0.9	16.9	2.2	0.24	36	0.25	100	160	1	35	64
14	5.6	84	2141	7.8	0.6	0.5	23.9	1.6	0.22	38	0.21	88	70	1	43	56
15	5.2	55	1954	8.7	0.6	0.6	18.8	2.1	0.19	38	0.29	95	37	1	34	65
16	7.2	144	3418	2.4	0.2	0.9	12.9	0.7	0.08	15	0.30	103	4	12	61	27
17	5.4	92	2296	7.3	0.6	1.8	14.2	2.7	0.15	34	0.27	105	18	1	28	71
18	6.6	76	n.d.	1.6	n.d.	0.9	10.4	1.1	0.20	16	0.27	51	<10	21	20	59
19	6.9	303	3971	1.5	0.1	1.1	13.8	1.6	0.12	19	0.69	136	13	4	21	75
20	6.8	255	3161	2.3	0.2	1.3	13.9	1.5	0.11	19	0.54	139	49	4	24	72
21	6.0	136	2033	2.7	0.2	0.8	8.1	0.8	0.11	17	0.56	82	21	10	29	61
22	6.4	76	1732	3.7	0.3	0.7	11.9	0.8	0.11	19	0.40	82	17	8	34	58
23	6.8	146	2259	3.0	0.3	1.6	15.9	1.2	0.15	22	0.46	128	18	8	29	63
24	5.9	190	2839	2.2	0.2	0.8	9.2	1.1	0.21	19	0.54	122	28	7	30	63
25	6.4	171	4878	4.4	0.4	1.7	14.5	1.9	0.16	27	0.54	134	19	12	48	40
26	6.1	150	2540	2.5	0.2	1.6	10.3	1.5	0.07	20	0.52	125	12	7	30	63
27	5.9	74	1944	7.1	0.6	1.1	19.4	1.5	0.26	30	0.33	77	43	44	42	14
28	5.8	24	1073	6.0	0.5	0.5	13.0	0.9	0.22	25	0.43	71	34	8	31	61
29	6.6	54	n.d.	6.3	n.d.	0.8	13.5	1.4	0.10	22	0.77	76	<10	40	41	19
30	6.7	53	2590	4.0	0.4	0.8	13.1	1.2	0.07	20	0.74	98	42	31	47	22
31	6.0	65	1947	2.8	0.3	0.7	9.4	1.0	0.12	17	0.41	144	46	28	39	33
32	6.7	31	1611	3.6	0.4	0.8	10.9	1.3	0.07	18	0.32	68	n.d.	27	50	23
33	6.2	41	1953	4.2	0.4	1.0	7.1	1.2	0.05	17	0.50	83	3	28	51	21
34	6.9	73	2445	3.6	0.4	1.2	11.8	1.4	0.05	18	0.77	98	6	30	50	20
35	5.9	70	1981	3.6	0.4	1.0	7.5	1.0	0.08	19	0.53	125	9	27	47	26

n.d. not determined

Cadmium concentration in potatoes

Potato tuber Cd concentrations ranged between 0.004 and 0.574 mg kg⁻¹ DW, with a mean concentration of 0.129 mg kg⁻¹ DW (0.026 mg kg⁻¹ FW) (Figure 2b), although 92% of samples had Cd concentrations < 0.20 mg kg⁻¹ DW. With the exception of two of the three replicates from a single site at Pukekohe, Cd concentrations in potato tubers were well below the FSANZ limit of 0.5 mg kg⁻¹ DW (assuming a dry weight content of 20%). The reason for the higher values at one of the Pukekohe sites is unknown as there were no obvious soil differences. Unlike the other sites surveyed, potatoes were the first crop at this site after it had been in long-term pasture.

There was no significant difference (P = 0.080) in the mean Cd concentrations between Innovator (0.111 mg kg⁻¹ DW) and Moonlight (0.134 mg kg⁻¹ DW). Furthermore, there was no significant difference in tuber Cd concentrations between regions (Figure 2b). This is in contrast with soil Cd concentrations and resulted in a marked regional variation in the transfer coefficient (Figure 2c), highlighting that the differences in tuber concentrations could be expected to be attributed to other soil properties affecting Cd phyto-availability or total plant uptake. Very few studies have reported TC values for potatoes to allow comparison with the present study. An exception is Ashrafzadeh et al. (2017) who reported TC values between 0.095 and nearly 6.0 for ten cultivars from the same site.

Relationship between potato Cd concentrations and soil properties

There were no significant relationships between potato tuber Cd concentrations and soil factors commonly shown to affect Cd concentrations, such as organic C, CEC, total Cd, total Zn, and soil Cl concentration (McLaughlin et al. 1996) (Table 4). The only significant variables were total soil Al, extractable Na and base saturation, that individually explained at most 8% of the variation in potato Cd concentration. The majority (86%) of the data used for exploring the relationships were from a single cultivar, discounting plant factors confounding the relationships. As reported, comparison of Cd concentrations between the two cultivars were not significantly different. Hence, in the relationship analysis, cultivar difference would be small as one cultivar dominated the data and Cd concentrations did not differ between the two cultivars. Stepwise forward multiple regression analysis indicated a subset of six soil properties i.e. Al + Mn + Na + Fe + P + Zn accounted for 56.9% of the variation in potato tuber Cd concentration (Table 5).

Table 4. Single relationships of Cd concentration in potatoes (DW) with soil properties.

Soil property	Significance level (p-value) of the relationship
pH	0.065
Olsen P	0.189
Exchangeable K	0.058
Exchangeable Ca	0.573
Exchangeable Mg	0.371
Exchangeable Na	0.002 S
CEC	0.156
Base saturation	0.023 S
Total Cd	0.709
Total Zn	0.578
Total P	0.052
Total N	0.056
Total C	0.070
C/N	0.122
Total Fe	0.477
Total Mn	0.529
Total S	0.621
Total Al	0.010 S

Note: S indicates a statistically significant relationship at the 5% level.

Significant relationship were positive except for base saturation.

Table 5. Coefficients of soil properties and their significance levels

Soil property	Coefficient	SE	% variation
Al	0.1275	0.0139	41.5***
Mn	−0.0535	0.0113	0.3***
Na	0.0199	0.0069	9.5**
Fe	−0.0479	0.0105	0.7***
P	−0.0677	0.0103	4.8***
Zn	0.0304	0.0112	0.1**

**, ***Significant at P < 0.01 and 0.001 levels, respectively.

It is uncertain why there were no significant relationships between soil properties such as pH and total Cd which have been previously shown to be important for Cd uptake in potatoes, although the absence of relationships is not unique to this study (e.g. Norton et al. 2015). It may simply be that compared to studies that have reported relationships (McLaughlin et al. 1997), soil properties in the current study are a lot narrower in range. The lack of a relationship between tuber Cd concentration and CEC may in part be related to the method used to measure CEC. Cation exchange capacity measured using neutral ammonium acetate opposed to the extracting the soil at its field pH, may have overestimated CEC, particularly in the variable charged soils in this study (i.e. Allophanic, Organic). It is also possible that site specific management (e.g. irrigation and fertiliser (quantity, type) application) that has been shown to influence Cd uptake in other crops such as wheat and flax (Grant et al. 1996; Jiao et al. 2004; Li et al. 2011) had some effect on Cd uptake between sites. Another factor that may have confounded results was differences in crop yield between sites. An increase in yield for example for any given cultivar may result in a decrease in that cultivar's Cd concentration. Although requested, yield data was not consistently provided by growers to allow testing of a crop yield effect.

Management of Cd concentrations in potatoes

To provide an approximation of the total soil Cd concentrations above which food standards might be exceeded, nCdfood standard values were derived at individual sites. The mean and 95th percentile range were 0.7 mg kg⁻¹ (0.31–1.4) in Canterbury, 2 mg kg⁻¹ (0.4–3.54) in Pukekohe, 2 mg kg⁻¹ (0.7–7.04) in Manawatū-Whanganui and 4.7 mg kg⁻¹ (1.19–16.7) in the Waikato. There was large variability in values, both within and between regions. This large variability has been noted in other studies. For example, De Vries and McLaughlin (2013) reported critical limits for potatoes in Australia between 0.147 and 0.959 mg kg⁻¹, reflecting the effect of soil pH and soil texture on Cd uptake. Despite the variability, and with the exception of Canterbury, potatoes grown in soils within the upper acceptable limit of the TFMS may be expected to remain within food standards.

In regions such as Canterbury where Cd uptake is high at relatively low soil Cd concentrations, and there is an absence of relationships with soil properties, cultivar management is potentially the best option for ensuring Cd concentrations in potatoes remain compliant with food standards if soil Cd concentrations continue to increase over time. Further, while there is no immediate urgency to identify low-Cd-accumulating cultivars as tuber concentrations are generally much lower than the FSANZ food standard, it would be prudent that potato breeding and selection trials should include Cd uptake as a plant trait to ensure that if high Cd accumulating cultivars are identified across a range of sites, that they do not become widely grown.

Conclusions

This study indicates there is variability in Cd concentrations in potatoes throughout New Zealand, and generally concentrations are much lower than the FSANZ ML. There were significant differences in Cd concentrations between potato cultivars. There were no significant relationships between soil properties such as pH and total Cd which have been previously shown to be important for predicting Cd concentrations in potatoes. With the exception of Canterbury, the derived soil Cd concentrations at which ML for Cd in potatoes might be reached are greater than the upper acceptable limit of the TFMS. Management of Cd in potatoes is likely to be most effective through the use of low-Cd-accumulating cultivars or at least avoiding the wide-spread growing of high-Cd-accumulating cultivars if they are identified from further screening studies. Further studies to determine if other management factors such as irrigation or fertiliser type and rate on Cd concentrations in potatoes are needed.

Acknowledgements

Chikako van Koten provided statistical advice and analysis. Two anonymous referees provided constructive comment that improved the manuscript.

Disclosure statement

No potential conflict of interest was reported by the authors.

ORCID

Colin W. Gray http://orcid.org/0000-0002-5397-8243

Niklas J. Lehto http://orcid.org/0000-0001-8563-2469

Brett H. Robinson http://orcid.org/0000-0003-0322-0255

Kiran Munir http://orcid.org/0000-0001-6600-5734

Jo-Anne E. Cavanagh http://orcid.org/0000-0003-3476-7472

Additional information

Funding

This project was part of a larger project funded by the Ministry for Primary Industries and the Fertiliser Association of New Zealand, with support from Taranaki Regional Council, Manaaki Whenua − Landcare Research, Foundation for Arable Research, Potatoes NZ, Waikato Regional Council, DairyNZ, Flour Millers Association, Bakers Industry Research Trust, Greater Wellington Regional Council, Ministry for the Environment, Beef + Lamb NZ, Bay of Plenty Regional Council, Vegetable New Zealand, OnionsNZ, Gisborne District Council, Environment Canterbury and Marlborough District Council.