Introduction

Soil is one of the most important environments for life on Earth. Microorganisms in the soil are responsible for the principal role of mineralization of soil organic carbon (SOC) and stabilization of organic forms from different carbon inputs (Jansson and Hofmockel 2020). Soil microorganisms are the main contributors of nutrient cycling, fertilization maintenance, and sequestration of carbon (Jacoby et al. 2017). Soils contain a wide variety of microbial habitats with different humidity, salinity, texture, pH, and other abiotic characteristics (Fierer 2017). The various habitats that the soil provide for microorganism can support a huge microbial diversity; even one gram of soil may contain 109 bacterial cells (Paul and Frey 2024).

Different agricultural managements can significantly influence the composition of bacterial communities and can cause the enrichment of certain taxa (Lee et al. 2020). The decrease in plant diversity due to different land uses can affect soil functionality by changing the structure of soil microbial communities (Wen et al. 2020). One of the most common methods of agricultural soil management is tillage practice, which can cause similar levels of erosion to vineyards as water (Blavet et al. 2009; Li et al. 2007). The no-tillage method can significantly reduce soil erosion compared with the conventional tillage method in vineyards (Cerdà et al. 2020), especially if vegetation cover, such as natural grass or cover crop, is present. A meta-analysis from the USA confirmed that the no-tillage systems enhanced the diversity of soil microbial communities by providing nutrient and stabilized habitats compared to conventional tillage and deep tillage (Nunes et al. 2020). On the other hand, reduced tillage had no such effect on the number of microbes, only on diversity, because bacterial species can have overlapping functions in the soil, so different microbial processes can be maintained (Li et al. 2020).

The common grape (V. vinifera L.) is an agriculturally important plant, so it is important to study the effect of tillage and no-tillage on the structure of soil bacterial communities, and consequently on grape health. As a primary reservoir, the soil plays an important role in the development of plant microbiota (Zarraonaindia et al. 2015). Based on the study of Gupta et al. (2019), the diverse and responsive soil microbiome gives the grape a more diverse and responsive plant microbiome, which can provide more effective help against abiotic and biotic stress, as well. Soil bacterial communities can facilitate many plant processes, for instance, hormone secretion, solubilization of iron or phosphorus, N fixation, and alert plant immune system (Tkacz and Poole 2021). Some soil surface bacteria are capable to initiate the production of rotundone (Gupta et al. 2019) responsible for the “peppery” taste of wine, thus affecting the taste of the final product (Herderich et al. 2012). The holobiont concept highlighted the importance of the core microbiome present in grape plants (Bettenfeld et al. 2020, 2022). From a biogeographical aspect, Gobbi et al. (2022) proved that spatial distance had a huge impact on both fungal and bacterial diversity of the vineyard soil. The greater distance resulted in more diverse communities, which also established the microbial terroir of the vineyards at a global scale (Gilbert et al. 2014). Due to the importance of viticulture and the aforementioned reasons, it is important to study the microbiota of grapevines. During the research of recent years, Kovács et al. (2020) found that in the Badacsony wine region, different agricultural managements (intensive, extensive, abandoned) influenced the fungal diversity of the grapevine rhizosphere, based on cultivation and sequencing methods. In the study of Knapp et al. (2021), a shared core fungal diversity was explored in the aboveground parts of Furmint grapes in the Hungarian Tokaj wine region. Another study focused on the noble rot development and the changes in the berry physical, chemical properties and the fungal microbial communities on grape berries. Using cultivation and molecular sequencing methods, it was found that the richness of the fungal community most likely changed with the physical properties (Hegyi-Kaló et al. 2020).

Based on previous studies, the widely used tillage methods in vineyards could influence the soil bacterial diversity (Lupwayi et al. 1998; Legrand et al. 2018; Khan et al. 2023). We hypothesized that tillage inter-row management with no vegetation present and soil redistribution due to erosion also have a significant effect on the soil bacterial communities in vineyards. The aim of the study was therefore to analyze the soil physical and chemical parameters affecting the microbial diversity and community structures and to study the taxonomic composition and changes of bacterial communities under different inter-row management in a vineyard situated in the Balaton Uplands.

Materials and methods

Sampling site description

The sampling sites are in vineyards of Balatoncsicsó area in the Balaton Uplands, Hungary (Fig. 1). The studied vineyard is in a catchment of the Csorsza creek (21 km2) with high agriculture usage, where the continental climate is affected by both Oceanic and Mediterranean influences (Horel and Zsigmond 2023). Vineyards in this area are usually located on a moderate to steep slopes which can lead to erosion problems. The site is located on an approximately 7–8% ascent, with rows parallel to the slope. Riesling grapes (V. vinifera L.) are grown on the site, which is a type of grape to make white wine. The cultivation and land-use method has not changed in recent years, and weeding is being carried out as necessary (Horel et al. 2022). The soil of the area is loess, which is characterized by a mainly homogenous distribution from the top few centimeters (Horel and Zsigmond 2023). The general soil type of the area is Cambisol or Luvisol developed from loess soil (Mantel et al. 2023).

Fig. 1
figure 1

The geographical locations of the sampling sites and slope inclines of the different inter-row managed parcels. (Vineyard N: tilled inter-row, Vineyard O: no-tilled (perennial grass) inter-row. Picture from Google Earth.)

Full size image

Soil sampling

The sampling was performed in 2020 in a vineyard where the different vine inter-rows were treated with tilled (N) and no-tilled (O) cultivation methods. The tilled inter-rows were plowed in the spring to approximately 20 cm depth. The tilled site had no vegetation cover only bare soil was present during the study year. The no-till inter-row had perennial grass cover, with occasional weeding performed. Three soil samples were taken from each sampling site, two at the rows and one at the inter-rows. The sampling sites were in at different heights, in the upper (U), hotspot/intermediate (H, HL), and lower (L) parts of the slopes. While the upper and lower points were adjusted to the edge of the vine row, the hotspots were determined based on the sudden change in the slope angle of inclination in the middle of the row. This sudden change means a concave soil situation, where soil deposition is likely due to erosive effects, and therefore, changes in the soil physical, chemical, and biological properties may occur. There were two sampling occasions, one in July 2020 (signed as 0720) when the grapes were in flower and another in October 2020 (signed as 1020) around harvest. (The sample names follow the order: cultivation method, slope position, and sampling time, sign of the sampling point, e.g., NH0720A, which is tilled, hotspot samples collected in 2020 July, the first of the 3 replicates.) The soil samples originated from around the top 10 cm were placed into single-use plastic bags and stored at 6–8 °C until laboratory processing.

Soil chemical and physical analyses and soil CO 2 flux measurements

The collected soil samples were air-dried and sieved (2 mm) before further analyses for chemical (NH4+–N, NO3−–N, total nitrogen (ISO 11261:1995), soil organic carbon (SOC) content (Tyurin method), electrical conductivity, pHH2O (MultiLine P4)) and physical parameters (grain size distribution of clay, silt, and sand contents) as described in more details by Horel and Zsigmond (2023). Soil CO2 fluxes (EGM5, PP Systems) and soil water contents (Hydrosense II, Campbell Scientific) were measured at each sampling points for both sampling events following the method described by Dencső et al. (2021).

Environmental DNA isolation and 16S rRNA gene-based amplicon sequencing

Following thorough mixing and removal of plant debris, 0.3 g of soil per sample was used for environmental DNA isolation. The DNA isolation process was based on the protocol of the DNeasy PowerSoil Kit (Qiagen, Hilden, Germany) with the modification that the cells were lysed for 2 min at 30 Hz in a Retsch Mixer Mill MM301 cell mill (Retsch, Haan, Germany) and performed in a final volume of 100 µl by the QiaCube (Qiagen, Hilden, Germany). The concentration of the extracted DNA was measured with Qubit™ fluorometer (Thermo Fisher Scientific, USA). The DNA was stored at – 70 °C until further procession.

The V3–V4 region of the 16S rRNA gene was amplified by polymerase chain reaction (PCR) using the CS1-TS-B341F (5’-CCTACGGGNGGCWGCAG-3’) and the modified CS2-TS-B805NR (5’-GACTACNVGGGTATCTAATCC-3’) primers (Herlemann et al. 2011). Amplifications were performed in duplicate, and the final volume was 20.0 µl from the following reaction components: 10.0 µl Master mix, 0.2 µl from both primer, 8.6 µl PCR pure water, and 1.0 µl from the DNA isolates. The PCR protocol started with primary denaturation at 98.0 °C for 5 min, after followed by 25 cycles of denaturation: 95.0 °C for 0.4 min, annelation: 55.0 °C for 0.30 min, extension: 72.0 °C for 1 min, and a final extension at 72.0 °C for 10 min. The PCR products were quality controlled in 1% agarose gel. A Qubit™ fluorometer (Thermo Fisher Scientific, USA) was used to check the concentration of the PCR products. The normalized, pooled, purified, and high-quality libraries were sequenced in a 2 × 250 bp paired-end format using a MiSeq v2 500 cycle reagent cartridge on the Illumina MiSeq platform (Illumina Inc., San Diego, USA).

Bioinformatic and statistical analysis

Raw sequencing data are available at sequence read archive (SRA) under accession number PRJNA782964. The bioinformatic analysis was the same as described in detail by Borsodi et al. (2024). All statistical calculations were performed using the software package R (R Core Team, Version 4.0.2). Package ggplot2 (Wickham 2016) was used for the visualization of stacked bar plots from the relative abundance data.

The Chao1 and ACE are nonparametric methods and estimate the richness with the calculation of expected OTUs based on observed OTUs (Kim et al. 2017). The Simpson index is a representative microbial index, as it measures the diversity and evenness of distribution (Feranchuk et al. 2018). The UPGMA (unweighted pair group method with arithmetic mean) clustering method based on Bray–Curtis dissimilarity indices was used for the dendrogram plot. This clustering method can preserve most of the initial distances (Mouchet et al. 2008). The Bray–Curtis dissimilarity index is probably the most used one in microbial ecology for relative abundance data, because it is sensitive to the differences in the dataset in the case of the abundant and less abundant OTUs as well (Modin et al. 2020). The non-metric multidimensional scaling (NMDS) is reliable statistical tool to help visualize complex multivariate data in a reduced number of dimension (usually 2–3) to see patterns in the dataset (Dexter et al. 2018). Since this nonlinear method uses a rank-based approach, it is less sensitive for the outlier data (Armstrong et al. 2022). This method was used for the visualization of the NMDS data, also based on the Bray–Curtis dissimilarity indices.

The effects of soil physical and chemical parameters, slope positions, and inter-row cultivation method of perennial grass or tillage on soil microbial communities were analyzed using nonparametric statistical analyses of the Wilcoxon and Kruskal–Wallis tests for the non-normally distributed datasets. Principal component analysis (PCA) was applied to explore the factor pattern of the selected soil and plant parameters. Pearson’s correlation coefficient (r) was used to evaluate the linear correlation between the soils’ physical and chemical characteristics, soil CO2 fluxes, and bacterial taxonomy data. Statistical significance of the data sets was determined at p < 0.05.

Results

Soil physical and chemical characteristics

Differences in soil physical and chemical parameters were observed for the different inter-row managed sites. Table 1 shows the soil physical and chemical parameters along with each points’ CO2 fluxes and soil water content. Among the sampling points, we found no significant differences between SWC, soil temperature, or CO2 flux data; however, higher SWCs were observed for sites with high clay contents (p > 0.05, U(df) = 5). Soil physical parameters showed some distinguishable differences among the sites, and the higher slope positions revealed a significantly lower sand and significantly higher clay contents (Table 1). Soil chemistry also differed among slope positions and soil management methods. In general, lower pH was observed for the perennial grassed inter-row compared to the tilled slope. SOCs and total nitrogen contents were significantly higher for the upper slope position compared to the hotspots or lower points (p < 0.05, U(df) = 2 for SOC and p < 0.05, U(df) = 2 for total N). P2O5 contents on the other hand showed significantly lower values for the upper points compared to the lower ones (p < 0.05, W = 2).

Table 1 Chemical and physical characteristics of the soils collected at different slope positions (upper. hotspot and lower) in 2020 July and October (mean, ± SD; n = 6)
Full size table

The PCA analyses revealed that the first principal component (PC1) accounted for 44.28% of the variation caused by the interaction, while PC2 accounted for 20.05% (Fig. 2). The data show clear partitioning of both upper and lower slope positions and sampling collection times. We found strong positive correlations between K2O and SOC or total nitrogen contents of the soils (r > 0.93 and r > 0.71; p < 0.001; n = 18 for tilled and no-tilled sites, respectively). Similarly strong negative connections between P2O5 and clay or positive with silt contents were also observed (r = − 0.74 and r = 0.85; p < 0.01; n = 18). Samples collected in July showed correlations between soil texture and SWC (r = 0.65 and r = − 0.72; p < 0.01; n = 18, for clay and sand, respectively), while in October these connections were less pronounced. Soil CO2 fluxes showed no strong correlation to most of the soil physical and chemical properties; however, the range of the measured values was relatively low (Fig. 2).

Fig. 2
figure 2

Principal component analysis (PCA) ordination biplot for the tilled (full) and no-tilled (empty) inter-row managed vineyards with environmental variables (selected soil physical and chemical parameters) represented as vectors, and the pairwise representation of the study sites. (SWC: soil water content; SOC: soil organic carbon content; Total N: total nitrogen content; n = 18. Sample signs for methods: N, tilled (filled); O, no-tilled (empty); for slope positions: U, upper (square); H, hotspot/intermediate (circle); L, lower (triangle); for sampling times: 0720 (blue) represents July 2020; 1020 (red) represents October 2020; for parallels: A, B, C)

Full size image

Diversity and taxonomic composition of soil bacterial communities

The sequencing resulted in a total of 1,983,467 bacterial sequences from the 36 soil samples, and it was assigned into 13,184 operational taxonomic units (OTUs). The coverage values were at least 0.97, so the sequences represented the bacterial communities well (Supplementary Table 1). For the species richness and diversity indices, the three parallels were used according to sampling time, slope position, and soil management. The Sobs, ACE, Chao1, and Inverse Simpson indices were calculated from the OTUs (Supplementary Table 1). The bacterial communities of the no-tilled soil samples were more similar to each other than the tilled ones, according to all indices (Supplementary Fig. 1 A-D). The bacterial community indices of the tilled soil samples from the upper part of the slope were significantly lower than those from the hotspot and the bottom of the slope at both sampling times (July and October). For both the tilled and no-tilled soil samples, the differences in the diversity indices among the samples from different slope positions (upper, hotspot, and lower) were larger in October than in July. The highest diversity values for both the tilled and no-tilled soil samples were detected in the slope hotspot sampling site in October.

As a results of sequencing, altogether 41 bacterial phyla or phylum-level taxonomic units were identified. The phyla for which the average relative abundance values per sample type exceeded 1% were Pseudomonadota (formerly Proteobacteria), Acidobacteriota (formerly Acidobacteria), Bacteroidota (formerly Bacteroidetes), Verrucomicrobiota (formerly Verrucomicrobia), Actinobacteriota (formerly Actinobacteria), Gemmatimonadota (formerly Gemmatimonadetes), Planctomycetota (formerly Planctomycetes), Myxococcota, Chloroflexota (formerly Chloroflexi), an unclassified phylum (Bacteria_unclassified), Bdellovibrionota, Elusimicrobiota, Armatimonadota (formerly Armatimonadetes) and Abditibacteriota (Fig. 3). Based on the relative abundance values of the dominant phyla, the bacterial communities of the soil samples were clustered according to the sampling times, using the UPGMA method with Bray–Curtis dissimilarity values. Of the phyla with the highest relative abundance, Pseudomonadota and Acidobacteriota were slightly more abundant in October, while Bacteroidota and Verrucomicrobiota in July. The Gemmatimonadota abundance was much higher in October. The hotspot and lower slope position soil samples from the tilled sampling sites (NH and NL samples) formed separated clusters in both sampling times.

Fig. 3
figure 3

Percentage distribution of 16S rRNA gene amplicon sequences among the bacterial phyla with relative abundance higher than 1% in at least one sample, together with the UPGMA dendrogram based on Bray–Curtis dissimilarity indices, in the soil samples of the vineyards. (Phyla with a relative abundance below 1% are grouped as Others. Sample signs are detailed in Fig. 2.)

Full size image

The stacked bar plot (Supplementary Fig. 2) based on bacterial orders with a relative abundance exceeding 1% in samples resulted in 45 orders. The five most abundant orders were Vicinamibacterales (Acidobacteriota), Chitinophagales (Bacteroidota), Burkholderiales (Pseudomonadota), Pedosphaerales (Verrucomicrobiota), and Sphingomonadales (Pseudomonadota), and their relative abundance was higher than 2% in all samples. The bacterial order-based UPGMA dendrogram shows the same sampling time-based primary separation as the bacterial phyla. The highest abundance of Vicinamibacterales showed seasonal changes, with a lower abundance in July than in October, while Chitinophagales has a higher abundance in the July samples.

The NMDS analysis was based on the relative abundance of the genera, where the abundance was higher than 1% at least one of the samples. On the NMDS plot (Fig. 4), the seasonal differences are clearly noticeable as expected (stress level = 0.147). The October soil samples were more similar to each other than the July ones. The parallels were relatively close to each other, which indicates their similarity. The exceptions to this were the no-tilled lower, hotspot, upper samples in July, and the tilled hotspot samples, where the parallels showed differences. The distribution of genera involved in the separation of the soil samples was quite diverse. Some genera showed strong correlations with certain sample types, such as Ramlibacter (Pseudomonadota) with the tilled upper slope positions October samples as well as Solirubrobacter (Actinobacteriota) with the tilled hotspot slope position in October samples. Some genera correlate highly with the sampling times, like the Sphingomonas (Pseudomonadota), Gemmatimonas (Gemmatimonadota), YC-ZSS-LKJ147 (Gemmatimonadota) with the October samples, Abditibacterium (Abditibacteriota), BIyi10 (Pseudomonadota), Adhaeribacter (Bacteroidota) with the July samples. In addition to the uncultured ones, the four most abundant genera and candidate genera were Vicinamibacteraceae_ge (Acidobacteriota), RB41 (Acidobacteriota), Sphingomonas (Pseudomonadota) Pedosphaeraceae_ge (Verrucomicrobiota), and their relative abundance was higher than 1% in almost all samples (Supplementary Fig. 3). The majority of the OTUs could only be identified at a taxonomic level higher than genus, such as order or higher. The NMDS plot of soil bacterial communities with the fitted environmental parameters is presented in Supplementary Fig. 4. None of the examined soil properties showed a significant positive or negative correlation with any bacterial taxon characteristic of any soil sample type.

Fig. 4
figure 4

NMDS (stress value = 0.147) plot from the relative abundance of the bacterial genera, where the abundance values are higher than 1% in at least one sample, the lower genera are combined as others group. (Sample signs are detailed in Fig. 2. The taxonomic names represented as numbers: 1, Vicinamibacteraceae_ge; 2, Ellin6055; 3, uncultured_Bacteria; 4, RB41; 5, uncultured_ge; 6, Pedosphaeraceae_ge; 7, Ellin6067; 8, Blastocatellaceae_unclassified; 9, Sphingomonas; 10, Pseudarthrobacter; 11, Ellin517; 12, Opitutus; 13, Xanthomonadaceae unclassified; 14, MND1; 15, Steroidobacter; 16, Subgroup 7 ge; 17, Massilia; 18, YC-ZSS-LKJ147; 19, Flavisolibacter; 20, Chthoniobacter; 21, Adhaeribacter; 22, env.OPS 17 ge; 23, Caenimonas; 24, Ramlibacter; 25, Puia; 26, Comamonadaceae unclassified; 27, Candidatus Udaeobacter; 28, Solirubrobacter; 29, PLTA13_ge; 30, Pseudomonas; 31, Chitinophagaceae unclassified; 32, Subgroup 10; 33, Pedosphaeraceae unclassified; 34, Terrimonas; 35, Gemmatimonas; 36, SC-I-84 ge; 37, Bryobacter; 38, Bacteria unclassified; 39, KD4-96_ge; 40, Streptomyces; 41, Microvirga; 42, Gemmatimonadaceae unclassified; 43, Subgroup 5 ge; 44, Ferruginibacter; 45, 67–14 ge; 46, TRA3-20 ge; 47, Haliangium; 48, Gaiella; 49, S0134 terrestrial group ge; 50, WD2101 soil group ge; 51, BIyi10; 52, Methylotenera; 53, Pedobacter; 54, Chitinophaga; 55, Edaphobaculum; 56, Microscillaceae unclassified; 57, Ohtaekwangia; 58, Flavobacterium; 59, Abditibacterium; 60, Sphingobacteriaceae unclassified; 61, OM190 ge; 62, Stenotrophomonas; and 63, Others)

Full size image

Discussion

Soil physical, chemical, and biological parameters were studied under different inter-row soil management methods in a sloping vineyard site. We aimed at finding the most important environmental driving factors responsible for changes in microbial community structures and diversities.

Our results indicated that the studied soil samples had high bacterial richness and diversity. The bacterial phylum composition of the studied vineyards was similar to that of other agricultural areas, with the dominance of the phyla Pseudomonadota and Acidobacteriota (formerly Acidobacteria) (Lupatini et al. 2014; Delgado-Baquerizo et al. 2018; Coller et al. 2019; Bettenfeld et al. 2022). The phyla Pseudomonadota and Actinobacteriota (formerly Actinobacteria) which were also present in the samples play a crucial role in the soil carbon cycle and producing secondary metabolites (Jenkins et al. 2009). The high relative abundance of the phylum Acidobacteriota was not surprising, since members of this phylum are widespread on Earth (Kalam et al. 2020). Interestingly, their abundance in bulk soil is usually higher than in the rhizosphere, but they take part in soil recovery and plant growth after disturbance (Kielak et al. 2016). Similar to the results of Berlanas et al. (2019), phyla detected from the studied small catchment vineyards with relative abundance higher than 1% can be regarded as a core component of grapevine rhizosphere soils. The elevated abundance of the phyla Bacteroidota and Gemmatimonadota could be the result of soil fertilization (Nemergut et al. 2008). The differences were mainly relied on the sampling time (July and October). The richness, evenness, and diversity of the bacterial community change during the different plant development phases except in the seedling stage (Bettenfeld et al. 2022; Chaparro et al., 2014). Marasco et al. (2018) reported that the bacterial diversity of the grape rhizosphere, however, mainly depends on the root genotype. The grapes of the investigated area belong to the same genotype, as all investigated varieties were Riesling, but other types of grapes can also be found nearby. Corneo et al. (2013) found that microbial communities in vineyard soils are relatively conserved over the seasons because changes in soil physicochemical properties can have a greater effect on them than temperature. In addition, monoculture can help maintain a similar bacterial community structure throughout the year (Dequiedt et al. 2011). The similarity between the hotspot and lower part samples can be the result of the less steep slope position; in this part of the incline the water runoff could be slower and less destructive, so less bacterial cell, but from the same accumulated composition could be transported away (Du et al. 2020).

The order and genus-level taxonomic analysis showed the same sampling time separation of the studied vineyard. At the order-level taxonomic composition, the relative abundance of Vicinamibacterales was higher in October than in July, while Chitinophagales and Pedosphaerales were the opposite. That could be caused by the soil moisture changes in the soil, which has an immense influence on the soil bacterial communities (López-Piñeiro et al. 2013). There may have been other reasons for this difference, such as changes in temperature and different elements such as potassium and phosphorus in the soil (Kennedy et al. 2005). This is supported by the fact that the result of the physical–chemical properties of the soil also showed a separation according to the sampling time in addition to the slope position. The genus Massilia, which correlated the tilled July samples in the upper position, also correlated with soil clay and potassium content. Species belonging to the genus Massilia can be observed in the rhizosphere of grapes (Compant et al. 2011). Soil clay particles provide a suitable habitat for microorganisms (Dong 2012). The genus Stenotrophomonas is also known as a plant growth promoting species and can be isolated from the rhizosphere and is usually a dominant taxon in soil (Ryan et al. 2009). Its abundance was, however, high only in the cultivated upper July samples of the soils of the Balaton Uplands vineyard. Both genera, Stenotrophomonas (Xiao et al. 2009) and Massilia (Cardinale et al. 2019), have outstanding phosphate solubilizing properties that could be the reason for the strong negative correlation with P2O5 content. However, the microbial diversity barely changed with tillage compared to the no-tillage.

The diversity indices showed that different trend prevailed in the samples. The tilled July samples represent an increasing tendency from the upper slope position to the lower slope position. This tendency was observable in the case of the estimated number of OTUs, the two estimated richness indexes and the Inverse Simpson diversity. The reason can be the accelerated bacterial colonization in the new niches formed due to the erosion and deposition of soil. The depositional zones, where the sediment accumulates, can provide more nutrient rich, homogenous, and moist habitat for the soil microorganisms, thereby microbial diversity can increase (Du et al. 2020). The bacterial cells could migrate on the surface of the slope caused by precipitation (Abu-Ashour and Lee 2000). Due to the accelerated erosion caused by the long-term tillage, which can be higher than erosion due to rainfall alone (Van Oost et al. 2006), the soil aggregates can break down because of the shred force of the rain and expose the microorganisms to the air and lower the soil water holding capacity, creating a stressed environment for the soil bacterial communities (Li et al. 2015; Wei et al. 2016). In contrast, the no-tilled areas were covered with vegetation in both July and October, which could significantly reduce the effect of runoff and water erosion (Zuazo and Pleguezuelo 2008). Different covering plants or crops can reduce soil erosion by up to 68% (Novara et al. 2011). The precipitation in 2020 July was lower than in October in the catchment area (Horel et al. 2022), and thus, the before-mentioned effects were more significant in October. This corresponds to the tendencies from the upper slope position to the hotspot for the tilled terroir. However, the October lower position samples showed less diversity than the October hotspot samples based on all diversity indices. It may have been caused by geographical attributes, which indicate that the upper slope position located in the detachment zone, while the hotspot could be in the starting point of the depositional zone, where the water runoff and erosion are slower (Novara et al. 2011, 2015). Thus, less bacterial washed out to the lower sampling site. This effect is also noticeable in the case of the no-tilled areas, but with a lesser extent due to the covering vegetation. The exact effect of tillage on the decrease in microbial diversity is still unclear as this practice increased the bacterial diversity in some soils (Wu et al. 2015; Degrune et al. 2016). However, it is difficult to compare these studies due to the different soil physical, chemical, and locational properties, which significantly influence the diversity of microorganisms (Legrand et al. 2018).

Conclusions for future biology

In this article, next-generation amplicon sequencing was used for a comparative analysis of soil bacterial community composition in vineyards from different slope positions and different tillage practices in the Balaton Uplands (Hungary). The results of the study suggest that accelerated erosion due to conventional tillage has affected the diversity of soil bacterial communities through increased accumulation in the depositional zones of slopes. Although accumulation also occurred in no-tillage sites, the rate was slower, and the amount was smaller due to the cover vegetation. In this way, the use of cover crops can help to reduce soil loss between vines, thereby also reducing microbial diversity.