Twelve sites (vineyards) located in a valley in Trentino (Northeast Italy) were investigated. The area covers 230 km² in a N-S direction and is crossed by the Adige River. The south-western part of the area is characterised by the presence of several lakes. According to the Köppen-Geiger climate classification, the entire area is classified as being warm-temperate (Cfb). The sites differ in their associated bedrock material and the soils are classified as four types:

The soil types were also characterised according to their elevation, available water capacity (AWC) and soil texture (Figure S1). For each soil type, three previously uninvestigated sites were identified in different geographic positions (Figure S2, Table S1). Six replicates per site were chose at the centre of the inter-rows. A random sampling design was used and the edges of the vineyard were omitted from the sampling. The Figure S3 shows the sampling design. Sampling was conducted in spring 2017. Table S1 shows some details that help contextualise the sites: municipality, site elevation, soil texture class and AWC (data collected from Trentino Soil Map - http://meteogis.fmach.it/cartaSuoli/indexnew.php), vineyard planting year, agronomic management, soil history (data collected from the registers of agronomic practices supplied by farmers) and the average of meteorological parameters over the previous ten years (data collected from the network of meteorological stations of Trentino - https://meteo.fmach.it/meteo/index.php). In all the vineyards, the ground was covered with permanent grass.

For microarthropod analysis, soil cubes of (10 × 10 × 10) cm³ were sampled between the rows. For the chemical and metataxonomic analyses, the soil was collected from the 0-20 cm layer after removal of the first few millimetres of the surface organic layer. Each sample (replicate per site) consisted of four cores sampled from around the sampled cube for microarthropod analysis and homogenised to obtain a representative sample. For each analysis type, 72 samples were assessed.

Chemical and texture analyses were carried out on the air-dried fine soil fraction (< 2 mm). Soil texture was determined as percentage of sand (2 mm - 50 µm), silt (50 µm - 2 µm) and clay (< 2 µm) by wet sieving and hydrometric assay of soil dispersant solution (sodium hexametaphosphate). The pH was measured using a pH meter on the soil-water suspension (w/v 1:2.5). Total carbonates were determined from powdered samples by volumetric method, measuring evolved CO₂ after the addition of HCl. Organic carbon content was obtained by calculating the difference between total carbon - measured by Dumas combustion of powdered soil and thermal conductivity detection using a CN analyser - and total carbonates. Soil organic matter (SOM) content was calculated using the determined organic carbon content (conversion factor 1.724). Total nitrogen (N) was measured simultaneously with total carbon. The available fraction of Cu, Zn, Pb and Cd were extracted in a DTPA solution and their concentrations were detected by ICP-OES.

For the soil microbial community analysis, total DNA was extracted from 0.25 g of fresh fine earth fraction (< 2 mm) that had been frozen at -80 °C immediately after sampling. To this end, a Power Soil DNA isolation kit (MO BIO Laboratories Inc., CA, USA) was used. Soil DNA amplification was performed using primers 515F (50-GTGYCAGCMGCCGCGGTAA-30) and 806R (50-GGACTACNVGGGTWTCTAAT-30) specific to the bacterial and archaeal 16S rRNA gene (Caporaso et al., 2011), and ITS1F (50-CTTGGTCATTTAGAGGAAGTAA-30) (Gardes et al., 1993; Berruti et al., 2017) and ITS2 (50-GCTGCGTTCTTCATCGATGC-30) (White et al., 1990) specific to the fungal ITS1 region. DNA purification, indexing, quantification, preparation of libraries for Illumina MiSeq (PE300) sequencing, data pre-processing and subsequent taxonomic classifications of operational taxonomic units (OTUs) were performed as previously described by Coller et al. (2019). The de novo greedy algorithm (MICCA) with 97 % identity was used to cluster the filtered sequences into OTUs. Subsequently, the Ribosomal Database Project (RDP) Classifier v2.11 was applied for taxonomic classification (Wang et al., 2007) as described by Coller et al. (2019). Finally, a total of 3,551,358 amplicon sequences of the V4 region of the 16S rRNA gene and 4,585,869 sequences of the ITS region were acquired from 72 soil samples. The 16S and ITS sequences were clustered into 24,355 and 7,856 OTUs (97 % identity), respectively. The 16S and ITS sequences were rarefied evenly at 16,000 and 22,000 reads per sample, respectively. After rarefaction, the dataset consisted of 21,840 prokaryotic OTUs (henceforth referred to as “bacterial”) and 6,929 fungal OTUs. The OTUs identified 409 bacterial and 314 fungal genera.

The microarthropod community was assessed by taxa composition, abundance (number of individuals per m²) and QBS-ar quality index. The latter is based on a direct correlation between soil quality and the number of microarthropods well-adapted to the soil habitat. QBS-ar divides the organisms into morphological classes according to their capacity to adapt to the environment and each “biological form” receives an ecomorphological score (EMI) within a range of 1 to 20, commensurate with its adaptation level (Menta et al., 2008). The microarthropods were extracted from the soil samples using a Berlese-Tullgren funnel (Wallwork, 1970). The extracted individuals were observed using a stereomicroscope and assigned to the suitable biological form. The QBS-ar score represented the total of the highest EMI values of each taxon (Parisi et al., 2005).

R software version 4.2.1 was used for the statistical analysis. The Wilcoxon non-parametric statistical test - with FDR correction following the Benjamini-Yekutieli procedure - was run to identify which parameters differed significantly under different conditions (p ≤ 0.05). The same italic letters indicate non-significant differences. The richness and diversity indices (number of taxa - S, Chao1 index, Shannon index - H' and Pielou index - J) were calculated by the functions of the vegan R package. Site characteristics (elevation, available water capacity- (AWC) and texture) were determined using factor analysis for mixed data (FAMD) (FactoMineR package). The variance in microbial (bacterial and fungal OTUs) and microarthropod (taxa) composition was analysed by non-metric multidimensional scaling (NMDS) based on Bray-Curtis dissimilarity and 999 permutations (vegan R package). The significance of the differences between the communities of various soil types was tested using analysis of similarity (ANOSIM) (vegan R package). Non-parametric correlation test (Spearman correlation) was used to estimate the relationships between edaphic parameters and the biota. We considered correlations with p ≤ 0.05 and ρ_S ≥ │0.5│ to be significant. The Mantel test was performed to evaluate Spearman rank correlations between single factors. Functional prediction of bacteria and fungi was calculated using Faprotax v1.2.10 (Louca et al., 2016) based on 16S and ITS data, respectively (microeco R package).

The texture of the four soil types in the twelve sites varied from loam to silt loam to sandy loam. In detail, SLA was characterised by a higher silt level than the others, DCC and DMG soils had a higher clay content, while MG soils showed a lower silt and a higher sand content than the other soil types (Table 1). All soil types were subalkaline. DCC and MG soils had the highest levels of carbonates and organic matter, while DMG soils showed the lowest total carbonate values and C/N ratio, and organic matter and total N that were half of DCC soils (Table 1). The available Cu concentration in the four soil types was on average in the range of 76-136 mg kg^₋1 d.w., and MG soils had the highest mean value of all soils. DCC and MG soils contained the highest available fraction of Zn. The available Pb concentration was higher in the two glacial soils (MG and DMG) than in SLA and DCC soils. In all sites, the available fraction of Cd was less than 0.21 mg kg^₋1 d.w. (Table 1).

Overall, 314 genera of fungi were detected (191 in SLA, 210 in DCC, 200 in MG and 238 in DMG). Nine genera of fungi represented 70 % of the total frequency of genera in all soil types: Mortierella (29.2 %), Cryptococcus (8.0 %), Alternaria (7.12 %), Mycoarthris (6.2 %), Tetracladium (5.9 %), Davidiella (4.8 %), Ilyonectria (3.9 %), Exophiala (2.9 %) and Clonostachys (2.5 %). Considering the genera with a frequency greater than 1.0 %, Mortierella, Tetracladium, Alternaria, Ilyonectria and Davidiella were the predominant fungal genera in each soil type. Minimedusa was more abundant in DCC and Cylindrocarpon was more abundant in SLA than in the other soils (Figure 1a). Cryptococcus and Mycoarthris were mainly found in DCC and glacial soils, Clonostachys and Exophiala in glacial soils (MG and DMG) and SLA and Botrytis, Nectria and Metarhizium in SLA and DCC (Figure 1a). The biodiversity indices S, Chao1 and J were found to be statistically different between the various soil types. In particular, S and Chao1 indices were higher in DMG and DCC soils than in the others (Table 2). The NMDS and ANOSIM analyses showed that the composition of the fungal community of SLA soils differed from the other soil types (Figure 2a, Table S2) and available Cd, total carbonates, pH, total N, SOM and Zn were the chemical variables that contributed the most to separate groups (Table S3). Several correlations were found between the fungal community and the chemical parameters (Table S5) and also among the different fungal genera. Mortierella in glacial soils (MG and DMG) showed negative correlations with metals (Cu, Zn and Cd), total N, SOM and clay. Additionally, Mortierella in glacial and DCC soils was found to be negatively correlated with some phytopathogenic genera (Davidiella in MG – ρ_S : -0.74, DMG – ρ_S : -0.68 and DCC – ρ_S : -0.72; Fusarium in MG – ρ_S : -0.62, DMG – ρ_S : -0.83; Alternaria in MG – ρ _S: -0.88). Tetracladium was found to be negatively correlated with the available fraction of Cu, Zn and Cd in all soil types, except for Cu in DCC. A negative correlation was also found with Pb in DMG. In DCC and DMG soils, Ilyonectria showed a significant positive correlation with Cu and negative Spearman coefficients for SOM, total N and Cd in DCC only; meanwhile, Davidiella displayed positive correlations with CaCO₃, SOM, total N and Pb in both the soil types and opposite correlations with Cu between the two soil types. Cylindrocarpon, despite only characterising SLA soils, showed an opposite correlation with SOM and total N between SLA and DCC and a negative correlation with Pb in both the latter soil types. Botrytis, which characterised the DCC soils, displayed negative correlations with Cu concentrations and with the genus Minimedusa (ρ_S : -0.58) in this soil type. In all soil types, endophytic fungi were more frequent than epiphytic ones (Table S4b). The most frequent functional classes of the fungal community were, in order of abundance, saprotrophs/decomposers, pathogens/parasites and mycorrhizae (Table S4b). The arbuscular mycorrhizae were less frequent in DCC and MG.

A total of 409 genera of bacteria were retrieved (335, 357, 336 and 351 in SLA, DCC, MG and DMG soils, respectively) and five genera (Gp6, Nitrososphaera, Gp4, Gp16 and Gp17) represented 40 % of the total frequency of genera in all soil types. Regarding the genera with a frequency greater than 1.0 % in each soil type (Figure 1b), these five genera characterised all soil types, whereas Gaiella characterised only SLA, DCC and DMG soils, Gp7 SLA and DMG soils and Pirellula DCC soils. The analysis of the biodiversity indices (Table 2) showed that those of S, Chao1 and J were higher in DCC than in the other soil types. No significant differences were found between SLA and glacial soils (MG and DMG). The bacterial community composition of the SLA soils, similarly to the fungal community, differed from the other soil types (Figure 2b, Table S2). The chemical parameters that contributed the most to differentiating the groups were available Cd, total carbonates, available Zn, total N and SOM (Table S3). Statistical correlations were found between some bacterial genera and the chemical parameters (Table S5). Negative and positive correlations were observed between the genera belonging to the phylum Acidobacteria (Gp16, Gp17, Gp4, Gp7 and Gp6) and SOM, total N and pH, in DCC and DMG soils, respectively. Furthermore, several positive correlations were observed between Acidobacteria and metals in all soil types. Available Cu positively affected the genus Gp4 in all soil types, Gp7 in SLA and DCC, Gp6 and Gp17 in DCC and Gp16 in DMG soils. Available Zn correlated positively with the genus Gp4 in SLA and the glacial soils (MG and DMG), and with Gp16 in SLA and DMG and Gp7 in SLA. Moreover, Gp7 showed a negative correlation with available Zn in MG. In DCC and MG soils, Nitrososphaera showed positive correlations with SOM, total N and available Cd, a negative correlation with the C/N ratio and an opposite correlation with available Cu and CaCO₃. Moreover, this genus correlated positively with available Zn and negatively with available Pb in MG soils. The genus Gaiella was found to be strongly related to edaphic parameters, particularly in DCC and DMG soils, and it exhibited negative correlations with metals, especially in DMG soils. A negative correlation between available Cu and Gaiella was also found in SLA and DCC soils. In the latter soils, Pb correlated positively with Gaiella. The genus Pirellula genus was positively related to SOM and total N in DCC and MG soils. The functional analysis (Table S4a) showed that the most represented bacterial functional groups in all soil types were those involved in C degradation and the N cycle. C degradation groups (cellulolysis, chitinolysis, hydrocarbon degradation and xilanolysis processes) were more frequent in CCD. Other C cycle-related functional groups (methanogenesis and utilisation of low molecular weight hydrocarbons) and denitrifiers were also among the highest in DCC.

A total of 42,412 specimens belonging to sixteen taxa were collected in all investigated soils (SLA: 8,732 individuals and 15 taxa, DCC: 8,458 individuals and 16 taxa, MG: 15,056 individuals and 15 taxa, DMG: 10,167 individuals and 13 taxa). The sixteen taxa found were Acari, Araneidi, Chilopoda, Coleoptera (adults and larvae), Collembola, Diplopoda, Diplura, Diptera (larvae), Hemiptera, Hymenoptera, Isopoda, Lepidoptera, Pauropoda, Protura, Symphyla and Thysanoptera. Acari (56.6 % ± 5.8 %), Collembola (31.5 % ± 6.0 %) and Hymenoptera (6.5 % ± 1.2 %), which were the most abundant groups in each soil type. Diplopoda were found in SLA and DCC soils only, Diplura in DCC and MG soils and Isopoda in all soil types except for DMG (Figure 1c). The composition of microarthropod communities showed no differences among the soil types (Figure 2c, Table S2). All biological indices of microarthropods (abundance, S, Chao1, H′, J and QBS-ar) showed the lowest values in DCC compared to the other soil types. The J and H′ indices were higher in SLA than in the other soil types, abundance and QBS-ar were higher in MG (Table 2). DCC soils were found to contain the microarthropod community (comprising Isopoda, Diplura, Pauropoda, Symphyla, Hemiptera, Coleoptera and Hymenoptera) that had the most negative correlations with available Cu (Table S5). Hemiptera and Coleoptera in SLA and Hemiptera and the larvae of Lepidoptera in MG soils also correlated negatively with this metal. By contrast, a positive correlation was found between Pauropoda, Coleoptera and Hymenoptera and available Cu in DMG soils. Hymenoptera, Diplura, the larvae of Diptera and Isopoda in DCC soils showed a higher number of correlations with other chemical parameters than the other soil types (Table S5).

Figure 1. Relative abundance of (a) fungal, (b) bacterial and microarthropod communities at genus level for the first two and class/order level for the third, based on Illumina sequencing of the four soil types. “Other” includes taxa with average relative abundance < 1 %.

Figure 2. Ordination plot generated by nonmetric multidimensional scaling (NMDS) based on Bray-Curtis dissimilarity, displaying the soil biota composition of the sites investigated, (a) Fungi, (b) Bacteria, and (c) Microarthropods and the contribution of the edaphic variables. Red vectors contributed significantly (p ≤ 0.05) to the separation of soil types, blue vectors did not contribute significantly. p is based on random data permutations.

Table 2. Fungal, bacterial and microarthropod biodiversity indices (mean ± standard error) in four soil types. Different letters indicate significant differences between soil types (Wilcoxon test with BY correction, p ≤ 0.05).

INDEX	SLA	DCC	MG	DMG
Fungi
S	69 ± 2 (b)	75 ± 4 (ab)	73 ± 1 (b)	83 ± 3 (a)
Chao1	73 ± 2 (b)	85 ± 4 (a)	79 ± 2 (b)	89 ± 3 (a)
ƒ	2.59 ± 0.05	2.49 ± 0.06	2.51 ± 0.08	2.59 ± 0.05
J	0.61 ± 0.01 (a)	0.58 ± 0.01 (b)	0.59 ± 0.02 (ab)	0.59 ± 0.01 (ab)
Bacteria
S	175 ± 4 (b)	197 ± 5 (a)	184 ± 2 (ab)	180 ± 3 (b)
Chao1	218 ± 5 (b)	240 ± 6 (a)	218 ± 3 (b)	218 ± 5 (b)
H′	2.67 ± 0.03	2.84 ± 0.06	2.66 ± 0.03	2.72 ± 0.05
J	0.52 ± 0.00 (b)	0.54 ± 0.01 (a)	0.51 ± 0.01 (b)	0.52 ± 0.01 (ab)
Microarthropods
S	9.2 ± 0.7 (a)	6.7 ± 0.9 (b)	8.8 ± 0.8 (ab)	8.3 ± 0.5 (a)
Chao1	11 ± 1	8 ± 1	11 ± 1	10 ± 1
H′	1.2 ± 0.05 (a)	0.68 ± 0.09 (c)	0.87 ± 0.07 (bc)	0.98 ± 0.04 (b)
J	0.57 ± 0.03 (a)	0.38 ± 0.04 (b)	0.42 ± 0.03 (b)	0.47 ± 0.02 (b)
Abundance (No m-2)	30893 ± 6792 (ab)	29927 ± 6268 (b)	53273 ± 9714 (a)	35974 ± 8453 (ab)
QBS-ar	104 ± 6 (ab)	77 ± 12 (c)	107 ± 7 (a)	90 ± 6 (b)

The Spearman correlation analysis highlighted the effects of Cu and Zn on soil biodiversity in the four soil types (Table 3). The biodiversity indices for SLA and MG soils displayed fewer significant correlations than for the other soil types. In DCC, available Cu concentrations showed a negative impact on microarthropod, bacterial and fungal biological indices. In DMG, negative correlations were found between the available fractions of Cu and Zn and fungal and bacterial indices, whereas positive correlations were observed between these metals and the microarthropod indices. In MG, available Zn concentrations showed negative correlations with the number of microarthropod taxa (S index).

The available fraction of Cd in all soil types exhibited low concentrations, showing the natural origin of this metal (Kabata-Pendias & Pendias, 2010). Available Pb concentrations were similar to those found in non-acidic Italian agricultural soils (Barbafieri et al., 1996) (range 0.5-5.0 mg kg^₋1), indicating that they could have both natural and anthropic origins. In the twelve sites investigated, the available Zn concentrations were found to be within the range reported for non-acidic Italian wine-growing soils (2-30 mg kg^₋1; Barbafieri et al., 1996). The presence of the Zn in the present study may derive from either natural sources or the application of chemical products. Furthermore, the highest values of available Zn were found in DCC and MG soils, where carbonates had the highest levels, confirming that Zn in soil is strongly associated with the carbonate matrix (Lestan et al., 2003). On average, available Cu in all soil types had concentrations above the mean value for Italian vineyards (30 mg kg^₋1; Barbafieri et al., 1996). These levels of copper may result from anthropogenic activity due to the application of this metal as a fungicide in viticulture. It is important to emphasise that levels of available Cu above 100 mg kg^₋1 can influence the soil biota (Keiblinger et al., 2018), and the European Union has thus restricted the application of plant protection products containing copper compounds to a maximum of 28 kg ha^-1 of copper over a 7-year period (Regulation EU 1981/2018); this aims to minimise potential accumulation in the soil and reduce exposure to non-target organisms. Research also showed the possibility of reducing the amount of copper used for vine protection and developing innovative formulations (Cabús et al., 2017; Battiston et al., 2019).

The NMDS and ANOSIM analyses showed that the fungal communities of the twelve sites studied were affected by pedological origin. Moreover, the dissimilarity observed between fungal communities of different soil types highlighted how edaphic factors are essential in the selection of community structure (Zheng et al., 2019). Among the most representative fungal genera of all four soil types were Mortierella, Tetracladium, Alternaria, Ilyonectria and Davidiella. The high frequency of Mortierella confirmed that it is one of the most abundant genera of fungi in agricultural and forest soils at all latitudes (Simon et al., 2017; Ozimek & Hanaka, 2021). Mortierella, having shown negative relationships with several phytopathogenic genera in this study, could have an aptitude for suppressing fungal plant diseases. Several studies have demonstrated that it is a PGPF (plant growth-promoting fungi) and is able to improve the bioavailability of certain elements, especially P (Tamayo‐Velez & Osorio, 2017) and Fe (Ozimek & Hanaka, 2021), as well as plant growth, stress tolerance (Yu et al., 2016; Ozimek & Hanaka, 2021) and the protection of crops from pathogens (Jaroszuk-Sciseł et al., 2011). Mortierella, like other genera belonging to the phylum Zygomycota, plays a key role in the decomposition of plant material (Alexopoulos et al., 1996) and the accumulation of organic carbon in the soil (Zhu et al., 2022). In glacial soils, Mortierella has been found to have a negative correlation with SOM and N. In particular, MG soils were characterised as having a high SOM content (5 %) and a high C/N ratio (> 11), suggesting that they undergo slow mineralisation, contain stabilised carbon and have low nitrogen availability. The positive correlation with C/N suggests the involvement of Mortierella in the decomposition of organic matter more susceptible to mineralisation and a potential reduction in its ability to humify SOM due to low mineral nitrogen availability. Despite being a fungal genus mainly observed in aquatic environments, Tetracladium has also been found in agricultural soils, where it is involved in the degradation of plant debris (Klaubauf et al., 2010) and is a typical root fungus capable of enhancing plant growth and nutrient uptake (Wang et al., 2018). Cu and Zn appeared to adversely affect Tetracladium, although in aquatic habitats some species have shown wide variability in their susceptibility to metal pollution (Miersch et al., 2005). The main species of Ilyonectria in soils are associated with root rot disease in numerous woody and herbaceous plants (Liao et al., 2019; Sánchez et al., 2019). In this study, this fungal genus characterised all soil types but appeared to be influenced only by the chemical characteristics of DCC and DMG soils, which had a similar texture but differed in other edaphic variables. In both soil types, Ilyonectria showed a positive relationship with Cu, demonstrating the ability of this fungus to tolerate metals (Torres-Cruza et al., 2018). Ilyonectria species are ubiquitous soilborne fungi, associated with the “black foot disease” of grapevine (Petit, 2017; Viret & Gindro, 2025). They can persist in a latent state, infecting young vines under unfavourable soil and climatic conditions (e.g., compacted or poorly drained soils; Petit, 2017; Viret & Gindro, 2025). With these characteristics, the widespread presence of this fungus in vineyards could be mainly due to the repeated cultivation of grapevine for long periods, rather than to the copper content. The genus Davidiella, whose members are mainly phytopathogens, characterised all the soil types, but showed a different response to edaphic characteristics depending on soil type. The abundance of arbuscular mycorrhizae appears to reflect the soil organic matter (SOM) content, which was higher in DCC and MG soils. However, in these soils, this functional group was less abundant compared to other soil types. A high SOM content likely indicates higher concentrations of nutrients, including available phosphorus (P). In soils rich in available P, the growth of arbuscular mycorrhizae may be reduced, as the host plant is not metabolically advantaged enough to establish symbiosis with the endophyte (Nouri et al., 2015).

The pedological origin of the investigated sites also affected the bacterial community composition, in agreement with the NMDS and ANOSIM analyses, and Bray-Curtis analysis showed how soil parameters drive the bacterial community structure (Zheng et al., 2019). Five of the most frequent genera in the four soil types (Gp4, Gp6, Gp7, Gp16 and Gp17), with a total average frequency of over 32 %, belong to the phylum Acidobacteria. This bacterial group is one of the most common in soils (Dunbar et al., 2002), and its abundance and diversity are driven by environmental factors, such as pH and nutrients (Kielak et al., 2016). In this study, positive correlations between Acidobacteria and the available fraction of metals, particularly Cu and Zn, highlighted that the use of Cu in viticulture may have affected the soil bacterial community and confirmed the ability of this phylum to tolerate pollutants (Kielak et al., 2016). Nitrososphaera was a dominant group in all soil types and showed positive correlations with total N in DCC and MG soils. It has been shown that this genus is closely related to organic and mineral forms of N (Zhalnina et al., 2013), and in agricultural soils it is among the main contributors to the oxidation of ammonium to nitrite (Xia et al., 2011), the first step in the nitrification process. This evidence is confirmed by functional analyses, showing that aerobic ammonia-oxidising bacteria were among the most abundant groups in all soil types and their frequency reflected the trend in total N in the diverse soil types. In this study, Nitrososphaera displayed both negative and positive relationships with metals. Several studies have demonstrated the tolerance of this genus of Archaea to high Cu concentrations, having been found in extreme environments such as mining sites with high metal concentrations (Maezato & Blum, 2012). Pirellula characterised the DCC soils, where SOM and total N exhibited the highest values compared to other soil types and were positively correlated with this genus. These results are consistent with those of Buckley et al. (2006), who found an over-representation of Pirellula in soils amended with organic fertilisers, in which an increase in SOM was noted. Moreover, the chitinolysis process was more pronounced in DCC soils than in others, and Rabus et al. (2002) found that Pirellula can grow with the chitin monomer N-acetylglucosamine as the sole source of C and N under aerobic conditions. Methanogens, methanotrophs and denitrifiers were found mainly in DCC soils, where clay, SOM and total N were higher than in the other soil types. This may indicate that temporary anaerobic conditions occur in DCC soils due to water saturation or oxygen consumption from SOM decomposition, inducing anaerobic processes such as methanogenesis and denitrification (Seo et al., 2014).

This study confirms that Acari, Collembola and Hymenoptera are the predominant groups in vineyard soils (Gagnarli et al., 2015). Diplopoda were predominant in SLA and DCC soils, where total N and C/N ratio were higher than in the other soil types. This is in line with the relationship between Diplopoda, N availability and C/N ratio found by Aerts et al. (2012), who observed that N content influences the decomposition rate of leaf litter, the main nutritional source for Diplopoda. Diplura were affected by high metal concentrations in DCC and MG soils, as also reported by Sendra et al. (2021). The lower amount of SOM and total carbonates in DMG soils could have limited the development of Isopoda in this soil type, as also shown by Hadjicharalampous et al. (2002). Conversely, DMG soils were hospitable for Diptera adults, which were absent in the other soil types. Moreover, a higher clay level and a greater number of fungal genera found in DMG soils are consistent with the positive relationship with the presence of Diptera (Garrido-Jurado et al., 2011).

The concentrations of available Zn and Cu are thought to be linked to viticultural practices. The highest values of these elements were found in soils where vines were also the previous crop. These two elements are, in fact, active components of pesticides or fertilisers, which are widely used in viticulture, and it is highly likely this is where they originate from, given the concentrations found in the soils of the investigated vineyards. Cu and Zn are essential micronutrients for all plants and animals, although they become toxic at concentrations above certain values (Jaishankar et al., 2014). The relationships between these metals and biodiversity indices are determined by soil types (Table 3), emphasising that tolerance of microorganisms to pollutants may depend on other soil characteristics, which could reduce their mobility and interaction with organisms or increase the adaptability and the liveability of the microbial community (Keiblinger et al., 2018). In this study, Cu content showed a negative impact on bacterial and fungal biodiversity indices in DCC and DMG soils, as also demonstrated by Wang et al. (2018). Berg et al. (2012) also found that long-term Cu exposure causes changes in soil bacterial community composition. Zn concentrations in DMG soils negatively affected fungal richness and indices of bacterial biodiversity. Several studies have shown a reduction in OTU microbial diversity in soils with high Zn concentrations (Golebiewski et al., 2014; Hur et al., 2011). The biological indices of microarthropods were negatively affected by Cu in DCC and in SLA soils. Similar results were found by Pedersen et al. (1999) in a study conducted in Denmark on long-term Cu-contaminated agricultural soils: overall, DCC soils were the worst at promoting microarthropod biodiversity, showing the lowest values of the biological indices, despite the high level of some chemical quality indicators (SOM, total N and C/N ratio). The factors behind the limitation in arthropod biodiversity are probably related to the physical structure of DCC soils, which are characterised by a high skeleton content, thus reducing arthropod mobility (Duyar, 2018).