Influence of No-Till and Conventional Soil Treatment on Pseudo-Total and Mobile Cu, Zn, and Cd in Carbonate Chernozems: The Case of Northern Kazakhstan

Figure 1. Monthly precipitation and air temperature during the growing seasons of 2020–2022, compared with long-term mean values

2.2 Soil sample collection and preparation

Soil samples were collected from the 0–20 cm soil layer. Within each plot, 10 soil cores were collected using a soil auger at points uniformly distributed across the area and combined into a single composite sample. Each composite represented the sampled soil layer at the plot level rather than an individual sampling point.

The common 0–20 cm sampling interval provided an integrated basis for comparing the upper soil layer under both tillage systems. Compositing did not allow separate assessment of the 0–10 and 10–20 cm sublayers and could therefore mask near-surface metal stratification under NT. The findings characterize the combined 0–20 cm layer rather than depth-specific metal distributions.

Prior to the analysis, soil samples were air-dried and sieved through a 1 mm sieve. Metal concentrations were expressed in mg kg⁻¹ of air-dry soil.

2.3 Sample extraction, instrumental analysis, and quality control

Cu, Zn, and Cd in the soil extracts were determined by flame atomic absorption spectrometry using an AA-140 spectrometer (Varian), an air–acetylene flame, and element-specific hollow cathode lamps. The analytical wavelengths were 324.8 nm for Cu, 213.9 nm for Zn, and 228.8 nm for Cd.

Calibration was performed using a series of standard solutions. Calibration stability was checked by analyzing a control standard after every five determinations. All determinations were repeated at least three times. Blanks were analyzed alongside the soil samples. These laboratory determinations were distinct from the field replicates and the annual treatment means used in the statistical analysis.

2.3.1 Pseudo-total, acid-soluble metals

Analytical recoveries for pseudo-total Cu, Zn, and Cd were 98.0%, 98.0%, and 97.0%, respectively. The corresponding recoveries for the mobile fractions were 97.0%, 98.0%, and 97.0%. Method detection limits (MDLs), expressed on an air-dry soil mass basis, were 0.050, 0.020, and 0.005 mg kg⁻¹ for pseudo-total Cu, Zn, and Cd, respectively, and 0.100, 0.040, and 0.010 mg kg⁻¹ for the corresponding mobile fractions.

Pseudo-total concentrations of Cu, Zn, and Cd were operationally defined as the acid-soluble fractions recovered by extraction with 5 M HNO₃. The extraction principle is described in RD 52.18.191-2018 [26]; the sample mass and heating conditions used here are specified below and differ from that reference procedure. These measurements do not represent total elemental concentrations obtained by complete soil digestion.

A 10 g portion of air-dry soil was placed in a conical flask and treated with 50 mL of 5 M HNO₃, giving a soil-to-solution ratio of 1:5. The suspension was heated for 3 h in a sand bath maintained at 95–100 ℃. After cooling, it was filtered through an ash-free “white tape” filter, and the first portions of the filtrate were discarded. Metal concentrations in the filtrate were determined by flame atomic absorption spectrometry.

2.3.2 Mobile metals

Mobile Cu, Zn, and Cd were operationally defined as the fractions extracted with ammonium acetate buffer at pH 4.8 using the two-step extraction described in Section 4.5.3 of the CINAO guidelines [27].

A 10 g portion of air-dry soil was combined with 50 mL of buffer and shaken for 1 h. The suspension was filtered while retaining as much soil as possible in the extraction flask. A second 50 mL portion of buffer was added to the remaining soil, and extraction was repeated for 1 h. The second suspension was transferred to the filter, and both extracts were combined for flame atomic absorption analysis. The soil-to-extractant ratio was 1:5 at each step, with a cumulative addition of 100 mL of buffer per 10 g of soil.

Throughout the study, “mobile” refers to the fraction recovered by this extraction, not to a direct measurement of plant uptake. The analytical scheme distinguished two operationally defined extractable fractions rather than providing a complete sequential fractionation of soil metals.

2.4 Statistical analysis

All calculations and statistical processing were performed in the R 4.5.1 environment. Observations within each year were averaged to obtain one annual mean concentration for each tillage × fertilization combination. The dataset therefore contained 24 annual treatment means per response variable: two tillage systems × four fertilization treatments × three years. These annual means, rather than individual analytical determinations, were used as observations in the reported ANOVA.

Separate ANOVA models were fitted for the pseudo-total and mobile concentrations of each metal. The models included tillage, fertilization, their interaction, and Year as an additive categorical blocking factor:

$Y_{i j k}=\mu+T_i+F_j+(T \times F)_{i j}+B_k+\varepsilon_{i j k}$       (1)

Here, Y_ijk is the annual mean concentration for tillage system i, fertilization treatment j, and year k; μ is the overall mean; T_i and F_j are the tillage and fertilization effects; (T × F)_ij is their interaction; B_k is the Year effect; and ε_ijk is the residual term. This balanced model had 14 residual degrees of freedom. Interactions with Year, field-block effects, and within-plot temporal correlations were not modeled separately.

Potential outliers were screened separately for each of the six response variables using externally studentized residuals from the corresponding ANOVA model. Annual treatment means with absolute externally studentized residuals greater than 3 were flagged for further examination rather than automatic exclusion. No annual treatment means exceeded this threshold, and all 24 annual treatment means per response variable were retained. This screening applied to the annual treatment means used in the ANOVA, not to individual field or laboratory observations.

Concentrations were summarized as means ± standard errors across the three annual means for each treatment combination:

$S E=\frac{S}{\sqrt{n}}, n=3$         (2)

Here, s is the sample standard deviation of the annual means. The SE therefore describes variation among annual treatment means rather than analytical repeatability or within-year variation among field replicates. The significance level was set at p ≤ 0.05.

The MF was calculated separately for each metal and treatment combination as the ratio of the corresponding three-year mean concentrations:

$M F (\%)=100 \times \frac{\bar{C}_{m o b}}{\bar{C}_{p t}}$    (3)

Here, $\bar{C}_{m o b}$ and $\bar{C}_{pt}$ denote the three-year mean mobile and pseudo-total concentrations, respectively. MF was used as a descriptive index of the mobile concentration relative to the acid-soluble concentration. This ratio of means is distinct from the arithmetic mean of three annual MF values. The ANOVA described above concerned metal concentrations, not MF.

Differences between tillage and fertilization treatments depended on both the metal and the fraction measured. Mobile Cu and pseudo-total Cd concentrations were higher under NT, whereas fertilized treatments had higher pseudo-total Zn means (Tables 2–4). These contrasting responses show why acid-soluble concentrations and mobile fractions should be considered together: differences in one fraction were not necessarily accompanied by corresponding differences in the other.

The higher mobile Cu concentrations under NT agree with the findings of de Santiago et al. [16], who reported greater DTPA-extractable Cu and Zn under NT after 21 years of contrasting tillage treatments on a calcareous Vertisol. They associated these differences with greater surface organic matter and microbial activity, although plant concentrations did not increase. Organic matter offers a plausible explanation for the Cu response in the present study because it can promote both retention on solid surfaces and mobilization through dissolved organic complexes. Lair et al. [15] demonstrated strong Cu retention by soil organic matter, while Hernández-Soriano et al. [28] and Lee and Chen [29] showed that dissolved organic compounds can enhance Cu solubilization. Thus, greater extractable Cu under NT can coexist with similar pseudo-total concentrations without implying an equivalent increase in plant uptake.

For Zn, the higher acid-soluble concentrations in fertilized treatments contrasted with the limited differentiation of mobile concentrations (Tables 2 and 4). This pattern is consistent with retention processes that restrict the transfer of Zn into the extracted mobile fraction. In a 38-year field experiment, Lair et al. [15] found that soil pH, organic matter, particle size, and mineral composition influenced metal retention and partitioning. Unlike the increased DTPA-extractable Zn reported by de Santiago et al. [16], mobile Zn in the present study showed little separation between tillage systems. Differences in soil properties, sampling depth, and extraction procedures provide reasonable explanations for this contrast.

Higher pseudo-total Cd under NT, alongside similar mobile concentrations, also has parallels in long-term tillage research. Düring et al. [14] reported increased aqua regia-extractable Cd in the 0–25 cm layer of a Eutric Cambisol under NT, together with near-surface enrichment of pollutants. Dubovik et al. [24] likewise documented vertical differentiation of trace elements in typical chernozem, with enrichment concentrated nearer the surface under direct sowing than under plowing. These studies identify retention and redistribution as plausible contributors to the Cd pattern observed here. However, redistribution within the sampled layer alone would not explain a higher layer-average concentration; differences in retention and metal inputs or losses must also be considered.

Interannual variation was evident across the metal profiles, particularly in the common increase in mobile Cd after 2020 (Figures 2 and 3). Differences in precipitation and temperature provide an environmental context for these changes (Figure 1). Hernández-Soriano et al. [28] found that metal solubility responded to moisture conditions and organic matter transformations, with lower mobility under saturation than at field capacity in their experiments. Hooda and Alloway [30] also examined redistribution among metal fractions during a two-year incubation of sludge-amended soils. These studies support moisture-dependent changes in soil chemistry as possible explanations for annual variation. The year factor nevertheless represents the combined conditions of each observation year rather than the separate contribution of rainfall or temperature.

The sensitivity of Cd can be interpreted through the competing effects of carbonate retention, solution chemistry, and pH. Davis et al. [31] demonstrated that calcite retains Cd through surface sorption followed by incorporation into mixed Cd–Ca carbonate solids. Conversely, Temminghoff et al. [32] showed in sandy soil that dissolved Ca competed with Cd for sorption sites and that decreasing pH reduced Cd retention. Calcium therefore has different roles depending on whether it occurs in a reactive carbonate surface or in soil solution. These opposing processes help explain why carbonate-rich soils can retain Cd while still exhibiting variation in its extractable concentration as chemical conditions change.

Redox processes provide a further comparison. Fulda et al. [33] found that reducing conditions in paddy soil promoted Cd-sulfide formation when sufficient sulfate was available, whereas reoxidation returned Cd to more soluble forms. Although flooded paddy conditions differ from the upland chernozem studied here, this finding illustrates why moisture-related changes in Cd mobility need not follow a single direction. Carbonate reactions, pH, competing ions, and redox conditions can favour different forms of Cd retention or release.

MF provides a complementary comparison of relative extractability. Its higher values for Cu under NT were consistent with the higher mobile Cu concentrations, whereas Zn MF remained low across treatments. Cd had the highest MF, at 27.14–37.70%, but this did not mean that most Cd occurred in the extracted mobile fraction. The lower Cd ratios under NT also reflected higher pseudo-total concentrations alongside similar mobile concentrations (Table 5; Figure 4). Meers et al. [34], comparing 13 extraction procedures across 28 soils, demonstrated that measured Cd extractability depends on the extractant and soil properties. Accordingly, MF should be interpreted as a method-specific ratio rather than a direct estimate of the proportion available for plant uptake.

Long-term studies further show that changes in metal extractability do not follow a universal aging pattern. Lu et al. [35] and Guo et al. [36] observed redistribution of added Cu towards more strongly retained fractions. McBride and Cai [37] also reported declining Cu and Zn extractability over a decade, although phytotoxicity persisted at high metal additions. In contrast, McGrath et al. [38] found no consistent decline in Zn and Cd extractability during 23 years after sludge application ceased. These contrasting outcomes indicate that soil properties, amendment history, and metal loading influence the persistence of extractable fractions. The differences observed in the present study therefore should not be assumed to disappear simply with increasing duration of NT management.

The distinction between soil extractability and crop response is also supported by studies of biosolid-amended soils. Black et al. [39] observed different temporal responses among free-ion activity, chemical extractions, and plant concentrations. Luo and Christie [40] reported increased Cu and Zn uptake associated with greater plant growth without corresponding increases in tissue concentrations, while studies by Ogbazghi et al. [41] and Kim et al. [42] highlighted the importance of water management and soil pH. For the present system, higher mobile Cu under NT therefore warrants comparison with crop Cu status rather than an immediate conclusion about improved nutrition. Combining consistent soil sampling with fraction-specific analysis and crop-tissue measurements would provide a stronger basis for micronutrient management and Cd monitoring than acid-soluble concentrations alone.

Composite sampling of the 0–20 cm layer may have obscured vertical stratification, particularly near-surface enrichment under NT. Future studies should distinguish the 0–10 and 10–20 cm layers and include deeper sampling to assess redistribution beyond the upper soil layer. The three-year observation period also limits conclusions about long-term accumulation and does not establish changes occurring immediately after the adoption of NT.

The soil characterization included pH and humus content but did not document their annual changes alongside metal concentrations. The Year factor therefore cannot separate the contributions of precipitation, temperature, and soil chemical conditions. Longer-term monitoring should combine metal measurements with soil moisture, organic carbon, pH, and CaCO₃. The two extraction procedures distinguished acid-soluble and mobile fractions but did not identify individual binding phases or directly measure plant uptake. Sequential fractionation, soil-solution analysis, and crop-tissue measurements would strengthen the interpretation of metal retention, mobility, and transfer to crops.

Also, no certified reference material was available, and this is acknowledged as a limitation.

The SE and ANOVA estimates were derived from digitized annual profiles and do not incorporate digitization uncertainty or variability among individual field plots. Retaining plot-level observations in subsequent analyses would improve estimates of treatment variability and allow repeated measurements from the same plots to be accounted for explicitly.