Soil bacterial communities exhibit systematic spatial variation with landform across a commercial potato field

Geoderma 335 (2019) 112–122

Contents lists available at ScienceDirect

Geoderma journal homepage: www.elsevier.com/locate/geoderma

Soil bacterial communities exhibit systematic spatial variation with landform across a commercial potato field

T

⁎

Saraswoti Neupane1, Claudia Goyer , Bernie J. Zebarth, Sheng Li, Sean Whitney Fredericton Research and Development Centre, Agriculture and Agri-Food Canada, Fredericton, NB, Canada

A R T I C LE I N FO

A B S T R A C T

Keywords: Bacterial diversity Topographic features Soil physico-chemical properties 16S rRNA Transect

Topography drives spatial variation of soil edaphic factors at the landscape scale however, it is unclear how it influences the spatial distribution of bacterial communities in distances relevant to agro-ecosystem management. This study examined the influence of soil physico-chemical properties and topographic features on bacterial communities and diversity in a commercial potato field with a rolling landform. Eighty-three soil samples were systematically collected across a transect 1100 m long. A significant negative correlation (r = −0.73) between soil pH (range 4.3–7.0) and slope gradient (range 1.8–11.9%) was observed. Regressions and/or a canonical correspondence analysis showed that pH, slope gradient and organic carbon were the major factors influencing bacterial α-diversity based on 16S rRNA gene sequences. Semivariogram analyses revealed that the bacterial αdiversity, the relative abundance of most phyla, pH and slope gradient showed strong to medium spatial autocorrelations with a range between 20.8 and 217.8 m. These results evidenced that soil pH and slope gradient were the major factors explaining variation in the spatial structure of the bacterial community. Our results showed that the soil bacterial communities varied in a systematic and predictable pattern in an agricultural field in response to variation in soil physico-chemical properties and topographic features.

1. Introduction Soil physico-chemical properties are well-recognized factors influencing the soil microbial community composition and diversity in terrestrial ecosystems (Fierer and Jackson, 2006; Lauber et al., 2009). The extent to which these factors affect microbial community composition has implications on ecosystem functioning. In agricultural ecosystems, the response of microbial communities to soil physico-chemical properties and other environmental factors determines the sustainability of agriculture, where microbial communities play a pivotal role by decomposing soil organic matter and mineralizing nutrients (Gougoulias et al., 2014). The shift in soil bacterial communities in response to soil edaphic factors is well recognized in different types of ecosystems (Lauber et al., 2008; Rousk et al., 2010). Agricultural practices alter soil physicochemical properties, which in turn can influence bacterial community composition and diversity. Cropping systems (Bossio et al., 1998), soil type and crop management practices (Jangid et al., 2008; Cederlund et al., 2014; Ramirez et al., 2010) were shown to change soil bacterial abundance and diversity through changes in soil pH and soil organic carbon (SOC) availability (Rousk et al., 2010; Peacock et al., 2001).

Topography often drives spatial variation of soil edaphic factors at the landscape scale (Moore et al., 1993) through its effects on the distribution of various factors, such as hydrological processes and soil erosion processes (Li et al., 2008). Solar radiation and temperature are affected by slope aspect (Holland and Steyn, 1975; Dubayah and Rich, 1995). Soil moisture regimes also vary with upper slope positions usually drier and better drained whereas lower slope positions are wetter and may be poorly drained (Moore et al., 1993). The majority of eroded soil is deposited locally in the lower-slope positions or depressional areas, resulting in changes in soil texture and soil organic matter. These processes often result in the development of different soils and in the formation of a soil catena along the slope (Pennock et al., 1987; Li et al., 2011). Various geostatistical techniques were used to describe the spatial variability of bacterial community diversity and their spatial relationship with other variables (Nunan et al., 2005). There is evidence of a spatial structure to soil microorganisms, with scales ranging from millimeters to hundreds of meters. Several studies have examined bacterial diversity at small (cm to < 10 m) scales (Nunan et al., 2005; Franklin and Mills, 2003; Ritz et al., 2004; O'Brien et al., 2016). Other studies have reported on spatial structure of microbial communities at spatial

⁎

Corresponding author. E-mail address: [email protected] (C. Goyer). 1 Present address: Department of Entomology, Kansas State University, Manhattan, KS, USA. https://doi.org/10.1016/j.geoderma.2018.08.016 Received 20 March 2018; Received in revised form 7 August 2018; Accepted 9 August 2018 0016-7061/ © 2018 Published by Elsevier B.V.

Geoderma 335 (2019) 112–122

S. Neupane et al.

2.3. Field soil sampling procedure

scales of km to hundreds of km (Zinger et al., 2011; Griffiths et al., 2011). In contrast, few studies have examined taxonomic variation in microbial (Osborne et al., 2011; Rosenzweig et al., 2016) or functional (Enwall et al., 2010; Philippot et al., 2013) communities at the scale at which agricultural management practices are performed (e.g., ~10 to 100 m scale) within individual fields (Cambouris et al., 2014). Therefore, the spatial structure of microbial communities within individual agricultural fields remains largely unexplored. The aim of this study was to examine the influence of soil physicochemical properties and topographic features on bacterial community composition and diversity using a transect in a commercial potato field with a rolling landform. We hypothesized that bacterial community diversity exhibits systematic spatial patterns, and that local soil physico-chemical properties and topographic features would directly influence the structure of the soil bacterial communities.

The experimental transect consisted of 83 sampling locations, with the distance between adjacent sampling locations ranging from 0.9 to 86.7 m (Fig. 1B). Locations of soil sampling points were determined following a stratified random sampling method described by Pennock et al. (2008). In brief, the aforementioned landform element was used as the strata and points were randomly assigned to each landform element at a probability proportional to its area. Points were sparse after the initial allocation. To facilitate spatial analysis for short ranges, one slope was randomly selected and additional sample points were assigned, again based on the stratified random sampling method. This nested sampling allowed the coverage for the whole field as well as enough short-range point pairs with a manageable sample size. A nonuniform distance between sampling locations also facilitated the development of semi-variograms. Soil samples were collected shortly before vine desiccation on September 15, 2014. One composite soil sample was collected from each sampling location. Each composite soil sample consisted of 10 soil cores collected from 0 to 15 cm depth in the potato hill. Soil was thoroughly mixed, and ~10 g of soil was subsampled, placed in a 15 ml falcon tube, stored under cool conditions during transport, and stored at −80 °C until used to evaluate the diversity of bacterial communities. The remainder of the soil sample was stored under cool conditions during transport, and passed through a 4.5 mm sieve. A subsample of approximately 20 g was oven dried at 105 °C to determine gravimetric water content (GWC). The remaining soil was dried at 30 °C for 72 h then stored at room temperature until further soil analyses.

2. Materials and methods 2.1. Field site description This study used samples collected from a transect in a commercial potato field located in New Brunswick, Canada in 2014. The field has grassed terrace diversions at approximately 60 m intervals to reduce water erosion. The transect was approximately 1100 m long and located between two terrace diversions near the lower portion of the field (Fig. 1A). Two grassed waterways were present at the lowest points along the transect, oriented perpendicular to the direction of the transect. Soils at the field site were generally developed in loamy glacial till and classified as podzols based on Food and Agriculture Organization of the United Nations. The field was cropped to potatoes in 2014 using standard production practices (Atlantic Canada Potato Guide; http://publications.gc.ca/ collections/collection_2015/aac-aafc/A53-1281-1967-eng.pdf). The preceding crop was spring barley (Hordeum vulgare L.). The field was limed in spring. Potatoes (cv. Russet Burbank) were planted in May and hilled in July 2014. Fertilizer was banded at planting at a rate of 1350 kg ha−1 (analysis N:P:K 17:12:17). Standard commercial herbicides, pesticides and fungicides were applied to control diseases, insects, and weeds. All crop inputs were applied uniformly throughout the field. No irrigation was applied as is common in this region. Daily average air temperature and total precipitation during the crop growing season (25 May 2014–15 September 2014) were 17 °C and 495 mm, respectively, as measured at the nearest weather station (St Leonard CS, New Brunswick; http://climate. weather.gc.ca/).

2.4. Soil physico-chemical properties Soil pH was determined in soil water suspension (1:1) (Hendershot et al., 2008). Soil texture was assessed using the pipette method following organic matter removal (Kroetsch and Wang, 2008). The soil organic carbon (SOC) and total nitrogen (TN) concentrations were measured by dry combustion (Skjemstad and Baldock, 2008) using a VarioMacro (Elementar Americas Inc., Mt. Laurel, New Jersey).

2.5. DNA extraction, library preparation and 16S rRNA gene sequencing Total DNA was extracted from 0.5 g of each soil sample using the method described by Griffiths et al. (2000). The genomic DNA was purified using PowerClean®DNA clean up kit (MoBio Laboratories, Carlsbad, CA, USA) and quantified using Fluoroskan Ascent™ Microplate Fluorometer (ThermoFisher Scientific, MA, USA). The bacterial V3-V4 region of 16S rRNA gene was amplified using the 341F and 806R primers with Illumina adapter sequences (Klindworth et al., 2013). Library preparation and sequencing were carried out following the instruction for 16S metagenomic sequencing library preparation and sequencing protocol available at the Illumina website (http://support. illumina.com/downloads/16s_metagenomic_sequening_library_ preparation.html). Briefly, amplification was carried out in a total volume of 25 μl with the following component: 12.5 ng DNA template, 12.5 μl KAPA HiFi HotStart Ready Mix, 5 μM of each primer. The PCR condition was 95 °C for 3 min and 25 cycles of 95 °C for 30 s, 55 °C for 30 s and 72 °C for 30 s and 5 min at 72 °C for elongation. The PCR product was purified using AMPure XP beads (Beckman Coulter, Pasadena, CA, USA) following the manufacturer's instruction. For each of the 83 samples, a dual indexed library was prepared using Nextera®XT Index Kit. An equal amount of DNA from each library was pooled then sequenced using the Illumina MiSeq Sequencer and the MiSeq Reagent kit v3 and following the 2 × 300 bp paired-end sequencing protocol (Illumina Inc., USA).

2.2. Topographic features Imagery was collected with an Ebee Unmanned Aerial Vehicle (UAV) equipped with a CANON 110 camera in spring of 2014, and a Digital Elevation Model (DEM) was created in Pix4D using the point cloud and the DEM routines. The original DEM was in 0.075 m resolution. The DEM was imported into ArcGIS (version 10.1), coarsened to 2 m (nearest neighbor), and then smoothed for 5 times (9 nodes average) to eliminate random errors. Topographic features including slope gradient, slope curvature, and aspect were determined using the Slope, Curvature and Aspect tool, respectively, in the Arc Toolbox of ArcGIS (version 10.4.1). The slope curvature was a single value which reflects both plan and profile curvature. The smoothed DEM was imported into LandMapR tool kit to divide the field into five landform elements (crest, upper-slope, mid-slope, lower-slope and depression) following the procedure described by Li et al. (2011). The geographic coordinates of each sampling location were recorded using a Trimble GeoXH GPS unit with Can-Net correction service (10 cm horizontal accuracy). The topographic features for each sampling point were extracted in ArcGIS based on their coordinates. 113

Geoderma 335 (2019) 112–122

S. Neupane et al.

Fig. 1. Aerial photo of the field (A) and spatial distribution of elevation and landform elements of the transect (B) with sampling locations indicated as black dots.

Wunsch, 1970). Sequences were checked for chimera using UCHIME (Edgar et al., 2011) within the Mothur pipeline and chimeric sequences were removed. For further cleaning, the ribosomal database project (RDP) reference database (Cole et al., 2009) was used to classify the sequences using the naïve Bayesian classifier (Wang et al., 2007) algorithm with 80% bootstrap cutoff. Sequences that were unclassified or assigned to Eukaryota, Archaea, mitochondria, and chloroplast were removed. After cleaning, there was a total of 7,712,711 sequence reads from the 83 samples. The sequences were clustered into operational taxonomic units (OTUs) with sequence identity of 97% using average neighbor algorithm within the Mothur pipeline. The consensus taxonomy for each OTU was assigned using the Ribosomal Database

2.6. Sequence analysis Image analysis, base calling, error estimation, quality filter and demultiplexing were performed using the CASAVA pipeline (version 1.8). Demultiplexed paired-end fastq sequences were analyzed in the Mothur software pipeline (version 1.34.1, Schloss et al., 2009). Briefly, paired-end sequences were assembled using the default criteria in make.contigs. Low quality sequences included ambiguous bases, erroneous length (> 450 bp) and homopolymers (> 8) were removed. High quality 8,389,093 sequence reads were aligned against Mothur compatible SILVA reference aligned database (Yilmaz et al., 2014) using the Needleman-Wunsch global alignment method (Needleman and 114

Geoderma 335 (2019) 112–122

S. Neupane et al.

very different parameter values for the same dataset. As a result, if different models were used for different variables, the parameter values will not be directly comparable. Therefore, in this study, a spherical model was used to fit the semivariogram in all cases. The spherical model was selected because it was the default model for the majority of variables tested in the study. Parameters of the model were used to evaluate the spatial dependence of the variables. The nugget value (C0) represents measurement errors or spatial sources of variation at distances smaller than the sampling interval or both. The range value defines the distance at which the data is no longer autocorrelated, and the sill value (C0 + C) represents the random variation beyond the range of autocorrelation. The sampling transect features small knolls approximately 100 m to 200 m wide to the east and a long slope to the west (Fig. 1B). The semivariograms all showed similar patterns of waves beyond 100 m distance. To avoid the effect of this large scale pattern on the range and sill parameters when fitting the models, the maximum lag distance was set at 100 m or 150 m. The spatial dependency was considered negligible if the R2 of the model was < 0.1. If the R2 of the model was > 0.1 but < 0.5 or the nugget over sill ratio (C0 / (C0 + C)) was > 0.75, the spatial dependency of the data was considered weak. If the R2 of the model was > 0.5, the spatial dependency of the data was considered intermediate when the nugget over sill ratio value was between 0.25 and 0.75 and strong when the nugget over sill ratio was < 0.25 (adapted from Cambardella et al., 1994; Duffera et al., 2007). Sequence data produced in this study were deposited to the European Nucleotide Archive Short Read Archive under the accession numbers ERPO22799.

Project database as a reference (Cole et al., 2009). To minimize the bias due to the differences in sequence depth across the samples, the sequence depth was equalized to the size of the smallest sample (i.e., 14,788 sequence reads) through random subsample method. Low abundance OTUs (≤0.005%) were removed as described by Bokulich et al. (2013). 2.7. Statistical analyses of relationships between bacterial diversity, soil physico-chemical properties and topographic features Pearson's correlation coefficients were calculated in Hmisc package within the R software platform (R Development Core Team, 2013) to determine the relationships between soil physico-chemical properties (pH, GWC, SOC, TN, C:N ratio, silt, sand and clay), and topographic features (elevation, slope gradient, slope curvature and aspect). Pearson's correlation coefficients were calculated to examine the relationships between the relative abundance of major bacterial phyla and classes (Proteobacteria, α-, β-, γ-, δ- Proteobacteria, Actinobacteria, Acidobacteria, Firmicutes, Verrucomicrobia, Chloroflexi, Planctomycetes, Gemmatimonadetes and Bacteroidetes) and soil physico-chemical properties and topographic features. Species richness, species richness estimator (Chao1), species diversity index (Shannon index), and Pielou's evenness were calculated for OTU dataset using vegan package within the R software platform (Oksanen et al., 2015; R Development Core Team, 2013). Bacterial βdiversity was assessed as Bray-Curtis dissimilarity index (Bray and Curtis, 1957) based on the abundance of OTUs for each pair of samples. Second-degree polynomial regressions were used to determine the relationships between α- or β-diversity and pH or slope. Differences in the mean relative abundance among pH and SOC categories of the 50 most abundant OTUs (phylotype), selected across all samples, were visualized using a heatmap. The samples were divided into six categories based on soil pH and SOC. Each category had 13 or 14 samples ensuring that the results were representative. Analysis of variance was performed to examine if the relative abundance of OTUs differed significantly in each category. Canonical correspondence analysis (CCA) was performed to examine the relationships between bacterial community composition using 16S rRNA gene sequencing and soil properties (pH, GWC, sand, clay, SOC, and TN) and topographic features (slope gradient and slope curvature). Environmental vectors (pH, GWC, sand, clay, SOC, TN, slope gradient and slope curvature) were fitted in CCA ordination. The strength and correlations of environmental vectors with CCA ordination were determined using a permutation based significance test with 999 permutations and plotted to visualize the relationships of bacterial communities and individual factors.

3. Results 3.1. Topographic features and soil physico-chemical properties The transect was located between two terrace diversion strips (Fig. 1A) and was characterized by an undulating landscape (Fig. 1B). Most of the landscape consisted of upper and mid-slope positions, with the lower slope positions mostly located beyond the field boundaries. The slope was mainly south facing. Elevation ranged from 117.6 to 141.0 m above sea level (Table 1). The slope gradient ranged from approximately 1.8 to 11.9%, and in many cases the direction of maximum slope was approximately perpendicular to the transect. As a result of the dominant slope within the field, there was relatively limited variation in aspect with a range from 2.3° to 179.9°. The slope curvature ranged from approximately −0.9 to 0.9 where negative values indicate concave curvature and positive values indicate convex curvature Table 1 Physico-chemical properties and topographic features of the experimental transect.

2.8. Statistical analyses of spatial variability of bacterial diversity, soil physico-chemical properties and topographic features Semivariograms were used to examine the spatial structure of topographic features, soil physico-chemical properties, and diversity of bacterial communities using the α-diversity (Shannon index) and the relative abundance of major phyla. The analysis was conducted in GS+ software (Gamma Design Software, LLC, Plainwell, MI, USA). In the analysis, the lag class distances were defined using non-uniform intervals, with the upper bounds at 3 m, 6 m, 10 m, followed by 5 m intervals until 60 m and then 10 m intervals until the maximum lag distance. Numbers of pairs were 23 and 38 for the first two lag classes, respectively and ranged from 51 to 147 for all other lag classes within 300 m. The smaller intervals (narrow bin widths) at shorter distances were designed to increase resolution while maintaining acceptable sample sizes for these lag classes. Four types of models (Linear, Spherical, Exponential and Gaussian) are available in the GS+ program, and by default, the model with the lowest value of Residual Sum of Square is selected as the best fit model. However, these models often produce

Characteristics

Mean (SD)

Range

Physico-chemical Clay content (g kg−1) Sand content (g kg−1) Silt content (g kg−1) Soil pH Soil organic carbon (g kg−1) Soil total nitrogen (g kg−1) Soil C:N Gravimetric water content (g g−1)

129.0 (16.3) 422.5 (39.4) 448.5 (33.3) 5.4 (0.7) 24.3 (6.4) 2.3 (0.4) 10.4 (1.7) 0.21 (0.03)

82.6–173.2 312.1–548.7 368.7–532.8 4.3–7.0 14.0–43.4 1.6–3.7 7.5–15.3 0.13–0.30

Topographic Elevation (m) Slope gradient (%) Slope curvature (% m−1) Aspect (°)

122.0 (5.1) 6.8 (2.5) 0.0 (0.4) 121.0 (39.4)

117.6–141.0 1.8–11.9 −0.9 - +0.9 2.3–179.9

SOC = soil organic carbon, TN = total nitrogen, C:N = organic carbon to nitrogen ratio, GWC = gravimetric water content. 115

Geoderma 335 (2019) 112–122

S. Neupane et al.

Table 2 Pearson correlation coefficients (r) between soil physico-chemical properties and topographic features. Elevation Slope Curvature Aspect Clay Sand Silt pH SOC TN C:N GWC

Slope

Curvature

Aspect

Clay

Sand

Silt

pH

SOC

TN

C:N

0.18 0.11 0.53⁎⁎⁎ −0.04 −0.21 −0.73⁎⁎⁎ −0.28⁎⁎ −0.05 −0.42⁎⁎⁎ −0.22

−0.03 −0.03 0.49⁎⁎⁎ −0.56⁎⁎⁎ −0.35⁎⁎ −0.42⁎⁎⁎ −0.41⁎⁎⁎ −0.28⁎⁎ −0.47⁎⁎⁎

0.09 −0.27⁎ 0.27⁎ −0.1 0.04 0.15 −0.09 0.13

−0.53⁎⁎⁎ 0.14 −0.46⁎⁎⁎ −0.07 0.18 −0.26⁎ 0.07

−0.91⁎⁎⁎ −0.22 −0.55⁎⁎⁎ −0.59⁎⁎⁎ −0.33⁎⁎ −0.61⁎⁎⁎

0.48⁎⁎⁎ 0.68⁎⁎⁎ 0.60⁎⁎⁎ 0.52⁎⁎⁎ 0.68⁎⁎⁎

0.42⁎⁎⁎ 0.17 0.53⁎⁎⁎ 0.49⁎⁎⁎

0.82⁎⁎⁎ 0.82⁎⁎⁎ 0.75⁎⁎⁎

0.35⁎⁎ 0.66⁎⁎⁎

0.57⁎⁎⁎

⁎

0.24 0.39⁎⁎⁎ −0.29⁎⁎ 0.13 0.29⁎⁎ −0.41⁎⁎⁎ −0.30⁎⁎ −0.29⁎⁎ −0.35⁎⁎⁎ −0.12 −0.33⁎⁎

Significance levels, ⁎ P < 0.05, ⁎⁎ P < 0.01, ⁎⁎⁎ P < 0.001; Slope = slope position, Curvature = Slope curvature, SOC = soil organic carbon, TN = total nitrogen, C:N = soil organic carbon to nitrogen ratio, GWC = gravimetric water content.

2208 OTUs. A total of 110 OTUs were highly abundant and were found in all sampling locations across the transect. Species richness, measured as the number of observed OTUs, averaged 1381 ± 190 (sd) and varied from 877 to 1640 across sampling locations (Table S1). Species evenness ranged from 0.75 to 0.90, α-diversity measured as Shannon index ranged from 5.22 to 6.59 while species richness (Chao1) ranged from 957 to 1979 (Table S1). Bacterial community composition differed across the transect. Variations in the relative abundances were observed in bacterial phylotypes taxonomically affiliated to Proteobacteria (26.1 to 47.7%), Actinobacteria (13.3 to 36.8%) and Acidobacteria (7.6 to 24.6%). Relative abundances of low abundant phyla ranged from 0.6 to 7.4% of Chloroflexi, 0.7 to 4.6% of Planctomycetes, 0.5 to 6.4% of Firmicutes, 0.5 to 9.2% Verrucomicrobia, 0.2 to 4.2% of Bacteroidetes, and 7.2 to 13.6% of other bacteria among the samples (data not shown).

(Table 1). The landscape segmentation analysis showed that most of the field was classified as mid-slope probably due to the high slope gradient values (Fig. 1B). The physico-chemical properties of soils varied along the transect. Soil texture ranged from 312.1 to 548.7 g kg−1 sand, 368.7 to 532.8 g kg−1 silt, and 82.6 to 173.2 g kg−1 clay (Table 1). SOC concentration ranged from 14.0 to 43.4 g kg−1 dry soil and soil TN concentration ranged from 1.6 to 3.7 g kg−1 dry soil. Soil pH ranged from 4.3 to 7.0, a surprisingly wide range given the history of uniform lime applications. Gravimetric water content measured at the time of soil sampling ranged from 0.13 to 0.30 g g−1. Correlations between measured topographic features and soil physico-chemical properties were explored (Table 2). Soil sand content was strongly positively correlated with slope curvature (r = 0.49; P < 0.001) (i.e., greater sand content with more convex curvature), while silt content was strongly negatively correlated (r = −0.56; P < 0.001) with slope curvature. SOC and TN were both negatively correlated with slope curvature while soil C:N was negatively correlated with slope gradient. Soil pH was strongly negatively correlated with slope gradient (r = −0.73; P < 0.001) (Fig. 2). Surprisingly, there was a positive correlation between clay content and slope gradient, perhaps due to greater clay content in subsoils. SOC and TN were positively correlated with gravimetric water content (r = 0.75; P < 0.001 and 0.66; P < 0.001, respectively).

3.3. Relationships between relative abundances of phyla and topographic features or soil physico-chemical properties Significant Pearson correlations (P < 0.001) between the relative abundance of all bacterial phyla and pH were observed (Table 3). The relative abundance of most bacterial phyla were also strongly correlated with slope gradient, which is not surprising given the strong relationship between soil pH and slope gradient (Fig. 2). The relative abundance of the majority of the phyla were significantly correlated with SOC, C:N ratio and soil clay content, however in all cases the correlation was less than for soil pH (Table 3). There was no correlation or weak correlations between the relative abundance of bacterial phyla and other topographical features (e.g., elevation, aspect and slope curvature), soil TN or sand content. Bacterial α-diversity calculated as Shannon index was negatively correlated with slope (r = −0.64, P < 0.0001) (Fig. S1), but positively with SOC and C:N ratio (r = 0.52 and 0.64, respectively) (Table 3). Bacterial α-diversity increased with increasing soil pH up to about pH 6, but did not increase further for additional increases in pH (Fig. 3A.). Similarly, the relationship between bacterial β-diversity (Bray-Curtis dissimilarity index) and soil pH was statistically significant (R2 = 0.55, P < 0.0001) (Fig. 3B). Bacterial β-diversity was greater at pH below 4.5 and pH > 6.4 compared to soil pH between pH 4.5 and 6.4 indicating that bacterial diversity was more dissimilar in more acidic soils or in soils close to neutrality.

3.2. Bacterial community composition and diversity The 16S rRNA gene sequences from 83 samples were clustered into

3.4. Factors influencing bacterial diversity The CCA showed that soil physico-chemical properties (pH, sand, clay, GWC, TN and SOC) and topographic features (slope gradient and slope curvature) explained 37.2% of the total variation in species (Fig. 4). Permutation based correlations were used to determine the significance of soil physico-chemical properties and topographic

Fig. 2. The relationship between slope gradient and soil pH based on linear regression analysis. 116

Geoderma 335 (2019) 112–122

S. Neupane et al.

Table 3 Pearson correlation coefficients between soil physico-chemical properties or topographic features and relative abundance of major bacterial phyla or α-diversity (Shannon index). Soil physico-chemical properties

Phyla or classes: Acidobacteria Actinobacteria Proteobacteria α-Proteobacteria β-Proteobacteria δ-Proteobacteria γ-Proteobacteria Bacteroidetes Chloroflexi Firmicutes Planctomycetes Verrucomicrobia Shannon Index

Topographic features

Clay

Sand

pH

TN

SOC

C:N

GWC

Slope curvature

Slope gradient

Elevation

Aspect

−0.37⁎ 0.28 0.05 −0.27 −0.45⁎⁎ −0.49⁎⁎ 0.34⁎ −0.35⁎ 0.45⁎⁎ 0.49⁎⁎ −0.26 −0.41⁎⁎ −0.37⁎

−0.22 0.14 0.13 −0.09 0.12 −0.1 0.12 −0.13 0.07 0.01 −0.22 −0.13 −0.13

0.81⁎⁎⁎ −0.53⁎⁎⁎ −0.33⁎ 0.53⁎⁎⁎ 0.59⁎⁎⁎ 0.84⁎⁎⁎ −0.73⁎⁎⁎ 0.66⁎⁎⁎ −0.79⁎⁎⁎ −0.79⁎⁎⁎ 0.61⁎⁎⁎ 0.72⁎⁎⁎ 0.78⁎⁎⁎

0.23 0.07 −0.23 0.24 −0.15 0.18 −0.26 0.19 −0.26 −0.04 0.12 0.13 0.23

0.49⁎⁎ −0.32⁎ −0.24 0.48⁎⁎ 0.06 0.51⁎⁎⁎ −0.48⁎⁎ 0.46⁎⁎ −0.51⁎⁎⁎ −0.33⁎ 0.48⁎⁎ 0.51⁎⁎⁎ 0.52⁎⁎⁎

0.6⁎⁎⁎ −0.62⁎⁎⁎ −0.18 0.56⁎⁎⁎ 0.21 0.66⁎⁎⁎ −0.54⁎⁎⁎ 0.59⁎⁎⁎ −0.59⁎⁎⁎ −0.49⁎⁎ 0.70⁎⁎⁎ 0.72⁎⁎⁎ 0.64⁎⁎⁎

0.39⁎ −0.21 −0.16 0.39⁎ 0.10 0.38⁎ −0.37⁎ 0.39⁎ −0.46⁎⁎ −0.29 0.28 0.34⁎ 0.38⁎

−0.40⁎⁎ 0.21 0.12 −0.04 −0.07 −0.27 0.15 −0.20 0.24 0.15 −0.23 −0.20 −0.24

−0.65⁎⁎⁎ 0.47⁎⁎ 0.26 −0.50⁎⁎⁎ −0.39⁎ −0.73⁎⁎⁎ 0.61⁎⁎⁎ −0.52⁎⁎⁎ 0.68⁎⁎⁎ 0.66⁎⁎⁎ −0.58⁎⁎⁎ −0.63⁎⁎⁎ −0.64⁎⁎⁎

−0.22 0.02 0.04 −0.02 −0.12 −0.20 0.09 −0.20 0.18 0.04 0.04 0.02 −0.09

−0.14 0.08 0.12 −0.11 −0.07 −0.17 0.18 0.04 0.16 0.14 −0.15 −0.17 −0.20

Significance levels: ⁎ P < 0.05, ⁎⁎ P < 0.01, GWC = gravimetric water content.

⁎⁎⁎

P < 0.001, SOC = soil organic carbon to nitrogen ratio, TN = total nitrogen, C:N = SOC to TN ratio,

Fig. 4. Canonical correspondence analyses (CCA) of bacterial diversity based on OTUs of the 83 samples (plotted as symbols) and soil properties or topographic features plotted as vectors. Soil properties included soil pH (pH), gravimetric water content (GWC), sand, clay, soil organic carbon (SOC), and total nitrogen (TN) and topographic features include slope gradient (slope) and slope curvature (curvature). The direction and length of arrows show the relative importance of soil physico-chemical properties or topographic features explaining the variability in bacterial community diversity.

of bacterial communities in the studied transect (Fig. 4). The opposing direction of the pH and slope vectors (Fig. 4) is consistent with the strong negative correlation between the two variables (Table 2). The vector for SOC was generally opposite to the sand vector (Fig. 4), which is consistent with SOC generally being associated with finer soil particles. Fig. 3. Second-degree polynomial regressions between bacterial (A) α-diversity (Shannon index) (B) β-diversity (Bray-Curtis dissimilatory index) and soil pH.

3.5. Effect of pH and SOC on dominant OTUs

features of the CCA ordination. The results showed that soil pH (R2 = 0.83, P < 0.0001), slope (R2 = 0.55, P < 0.0001) and SOC (R2 = 0.30, P < 0.0001) were major factors explaining the variability

The most abundant phylotype OTUs consisted of unclassified bacteria, Rhizobiales, Gaiella, γ-Proteobacteria and Acidobacteria Gp-6 with relative abundances between of 5.2% and 9.4% (Fig. 5). Out of the 50 most abundant OTUs, 16 OTUs had significant differences in relative abundances among six categories based on pH and SOC categories 117

Geoderma 335 (2019) 112–122

S. Neupane et al.

Fig. 5. Heatmap of the top 50 most abundant OTUs across 6 categories of pH and soil organic carbon (SOC). Ranges of pH and SOC were selected to ensure a similar number of samples in each category. The mean of the relative abundance of 13 or 14 samples (in each category) for individual OTUs were plotted where columns and rows represent category and OTUs. Taxonomic affiliations of each OTU were assigned to the genus level when possible or to the closest known taxonomic level. The color scale bar shows the percentages of OTU in the categories.

4 was significantly different between low and high SOC at the lower pH but not at medium or high pH. The relative abundance of most OTUs was greater at high SOC compared to low SOC. This is visually more evident for OTUs classified as Acidobacteria Gp-4 and Verrucomicrobia subdivision 3 at lower pH, Acidobacteria Gp-16 and Spartobacteria at medium pH and Spartobacteria and candidate division WPS.2 at high pH (Fig. 5).

(Fig. 5). No significant difference was observed in the relative abundance of unclassified bacteria, Gaiella and γ-Proteobacteria however the relative abundances of the most abundant OTUs classified as Rhizobiales, and Acidobacteria Gp-6 differed significantly between pH and/ or SOC categories (P ≤ 0.001). Sixteen OTUs had significantly greater relative abundances at high pH compared to low pH including Rhizobiales, Acidobacteria Gp-6 and Gp-16, Spartobacteria and β-Proteobacteria (Fig. 5). Several OTUs had greater relative abundances at low pH (< 4.95) compared to medium (4.95 to 5.68) and high pH (> 5.68) including OTUs classified as γ-Proteobacteria, Acidobacteria Gp-1 and GP-3, and Xanthomonadaceae however these trends were not statistically significant (Fig. 5). Significant differences in the relative abundance between low and high SOC were observed for ten OTUs at low, medium and high pH including OTUs classified as Rhizobiales (P ≤ 0.0001), Acidobacteria Gp-16 (≤0.0001), Spartobacteria and candidate division WPS.2. The relative abundance of Acidobacteria GP-

3.6. Spatial patterns of topographic features, soil physico-chemical properties and bacterial diversity Semivariogram analyses showed that all soil physico-chemical properties, except for sand content, were strongly autocorrelated in space with high R2 values and low C0 / (C0 + C) values (Fig. 6A and B; Table 4). Sand content also displayed intermediate spatial dependence. Nugget values were very low, indicating that the sampling interval was 118

Geoderma 335 (2019) 112–122

S. Neupane et al.

Fig. 6. Semi-variograms of slope and curvature (A) pH (B), α-diversity of bacteria community (Shannon index) (C) and the relative abundance of major bacterial phyla (D) and Proteobacteria classes (E, F).

curvature, and pH showed similar patterns to the distance of 600 m (data not shown). A further examination of phylum Proteobacteria revealed that the relative abundances of its classes showed very different spatial dependence patterns where β-Proteobacteria showed negligible spatial dependence, γ-Proteobacteria showed intermediate level of spatial dependence, and α- and δ-Proteobacteria both showed strong spatial dependence. However, α- and δ-Proteobacteria had very different autocorrelation ranges of 31.6 and 217.7 m, respectively (Fig. 6E and F; Table 4).

short enough to capture most of the in-field variations. Sand and clay content showed short-range autocorrelations with 17.7 m and 24.8 m, respectively. Slope gradient, slope curvature, pH and silt content showed medium-range (from 44.7 m to 53.7 m) autocorrelations, whereas GWC, TN, SOC and C:N all showed long-range (> 100 m) autocorrelations (Table 4). The bacterial α-diversity as measured by the Shannon index showed a strong spatial autocorrelation with R2 of 0.91 and the semivariogram showed a medium range of 50.8 m (Fig. 6C, Table 4). Furthermore, the relative abundance of all bacterial phyla showed strong spatial autocorrelation with R2 ranging between 0.77 and 0.98 with the exception of the Proteobacteria, which showed negligible spatial dependence with a R2 of 0.01 (Fig. 6E and F; Table 4). The semivariograms of the relative abundance of all bacterial phyla, except for Proteobacteria, had medium ranges (33.4 m to 53.2 m), similar to those semivariograms for slope gradient, slope curvature, pH and silt content. Furthermore, the shape of the semivariograms for these phyla and slope gradient, slope

4. Discussion 4.1. Spatial distribution of topographic features and soil physico-chemical properties Topography is an important factor controlling soil and hydrological processes in the landscape (Li et al., 2008). Given the importance of 119

Geoderma 335 (2019) 112–122

S. Neupane et al.

Table 4 Statistics and parameters for the spherical semivariogram models of soil properties, relative abundance of bacterial phyla and bacterial α-diversity (Shannon index). C0

C0 + C

Range (m)

R2

C0 / (C0 + C)

Spatial dependence

Soil and landscape properties: Slope gradient (%) Slope curvature (% m−1) pH GWC (g g−1) Sand (g kg−1) Silt (g kg−1) Clay (g kg−1) TN (g kg−1) SOC (g kg−1) C:N

0.01 0.0066 0.0001 0.000065 7 167 0.2 0.0053 0.1 0.01

6.041 0.1662 0.2732 0.00158 1159 831 192.5 0.1836 77.6 5.043

44.7 53.7 47.7 140.1 17.7 47.2 24.8 113.5 148.8 162.5

0.98 0.84 0.95 0.90 0.68 0.79 0.76 0.89 0.90 0.94

0.002 0.04 0 0.041 0.006 0.201 0.001 0.029 0.001 0.002

3 3 3 3 2 3 3 3 3 3

Relative abundance (%): Proteobacteria α-Proteobacteria β-Proteobacteria γ-Proteobacteria δ-Proteobacteria Actinobacteria Acidobacteria Firmicutes Verrucomicrobia Chloroflexi Planctomycetes Gemmatimonadetes Bacteroidetes

4.64 2.87 0.001 11.3 0.055 8.63 4.15 0.224 0.16 0.171 0.101 0.076 0.146

22.4 9.62 1.787 41.23 0.921 32.5 17.37 2.002 3.641 1.942 1.154 1.001 0.756

3.3 31.6 2.5 20.8 217.8 40 49 33.4 50.1 35.5 50 53.2 36

0.01 0.77 < 0.01 0.59 0.97 0.93 0.94 0.70 0.96 0.89 0.98 0.95 0.83

0.207 0.298 0.001 0.274 0.06 0.266 0.239 0.112 0.044 0.088 0.088 0.076 0.193

0 3 0 2 3 3 3 2 3 3 3 3 3

Diversity: Shannon index

0.0379

0.1518

50.8

0.91

0.25

3

C0 = nugget, C0 + C = Sill, C0 / (C0 + C) = relative nugget effect, SOC = soil organic carbon, TN = total nitrogen, C:N = organic C to nitrogen ratio, GWC = gravimetric water content.

matter is important in controlling both GWC and TN within a landscape. These results indicate that soil physico-chemical properties played an important role in shaping the soil bacterial community composition and diversity in this study, and that the topographic features of the landscape controlled the spatial distribution of the soil physico-chemical properties. Significant Pearson correlations between the relative abundance of the bacterial phyla and soil physico-chemical properties (pH and SOC) or topographic features (slope gradient) were observed further supporting the importance of these parameters in influencing bacterial communities. Furthermore, significant changes in OTUs relative abundances among pH and SOC categories were observed in this study. Historically, soils in this region are naturally acidic because they would have been under coniferous forests for thousands of years before these soils were used for agriculture. The inherent acidity of these soils might have promoted species that were better adapted to grow at acidic pH. Other studies have shown the effect of pH on the relative abundance of bacterial phyla but not at a finer taxonomic scale in their study systems (Griffith et al., 2011; Lauber et al., 2009; Rousk et al., 2010; Fierer et al., 2007; Eilers et al., 2010). SOC is generally low in this field site with an average of 2.4% due to intensive agriculture. Despite this, differences in relative abundances of the predominant bacteria between SOC levels were observed. Almost all bacteria showed differences between SOC categories at all pH categories. pH strongly influences carbon availability (Kemmitt et al., 2006) which in turns could influence the abundance of some species.

topography on soil formation and soil catena development (Pennock et al., 1987), it is to be expected that topography changes soil physicochemical properties at the landscape scale (Moore et al., 1993; Seibert et al., 2007). Topography was reported to have a more pronounced effect in a rolling landscape than in a flat landscape (Birkeland, 1999). Potato growing areas in New Brunswick, Canada are characterized by undulating landforms with podzolic soils (Zebarth et al., 2002). In this study, topographic features correlated with soil physicochemical properties. There was a strong negative correlation between slope gradient and soil pH, an effect attributed to increased soil erosion by water and tillage at locations with greater slope gradient resulting in exposure of the low pH subsoil. Concave slope curvature was associated with increased GWC, increased accumulations of silt and SOC and with a decrease in sand content. A study of Zebarth et al. (2002) in a potato field reported similar results. Clay is typically deposited in lower slope positions, however no increase in soil clay content was associated with concave slope curvature in this study. Rather, soil clay content increased with increasing slope gradient, which may be due to an increase in clay content in the subsoil. 4.2. Effect of topographic features and soil physico-chemical properties on bacterial communities and diversity Soil pH, slope gradient and SOC were the three most important drivers for the α-diversity and bacterial community composition determined by linear or second-degree polynomial regressions and CCA. It is well known that soil pH shapes bacterial community composition and diversity in soils (Rousk et al., 2010; Lauber et al., 2009; Fierer and Jackson, 2006). Bacterial α-diversity was also increased with increasing SOC, and increased SOC was associated with concave curvature which is indicative of lower slope positions. Similarly, bacterial α-diversity in bulk soil was greater in lower-slope positions with concave curvature than in upper-slope positions with convex curvature under perennial and annual agricultural crops fields (Hargreaves et al., 2015). The SOC is commonly used as a measure of soil organic matter, and soil organic

4.3. Spatial distribution of bacterial communities In agricultural ecosystems, agricultural management practices increase the complexity of the edaphic factors that influence the spatial distribution of bacterial communities by changing soil physico-chemical properties (Kosmas et al., 2001). Agricultural management practices are typically performed at a ~10 to 100 m scale within individual fields (Cambouris et al., 2014), and thus a medium spatial 120

Geoderma 335 (2019) 112–122

S. Neupane et al.

scale might be more appropriate in studying bacterial distribution in agri-ecosystem. In this study, soil bacterial communities had a clear spatial pattern in their distribution in a commercial potato field. The αdiversity and relative abundance of all bacteria phyla (except for Proteobacteria) showed strong spatial autocorrelations in ranges similar to soil pH, slope gradient and slope curvature, further supporting the importance of these factors in driving diversity of bacterial communities. Although the relative abundance of phylum Proteobacteria showed negligible spatial autocorrelation, the most abundant class, the α-Proteobacteria, showed strong spatial dependence and the range was similar to those of the other bacteria phyla. The different patterns of spatial dependence for other classes under phylum Proteobacteria suggest that other unknown factors are involved. A few studies have investigated the spatial distribution of bacterial community diversity in agricultural fields (Franklin and Mills, 2003; O'Brien et al., 2016; Osborne et al., 2011; Rosenzweig et al., 2016). However, none of these studies evaluated the possible interactions between topographic features and soil physico-chemical properties on the spatial structure of soil bacterial community diversity. The only study that had a similar aim showed that soil bacterial community diversity under annual and perennial agricultural crops changed between plots positioned on the upper, mid and lower slope positions despite having similar soil physico-chemical properties (Hargreaves et al., 2015). However, the range at which the bacterial community was spatially structured in the landscape was not evaluated. This study demonstrated that the soil bacterial diversity and relative abundance of bacterial phyla varied in a systematic and predictable pattern in response to variations in topography and soil physico-chemical properties in a potato field with a rolling landform. The bacterial community composition and diversity were greatly influenced by topographic features and soil physico-chemical properties despite intensive tillage practices and uniform inputs. The study of spatial distribution of bacterial communities brings new insights into how communities develop in soil systems, and what factors may be important in maintaining and regulating soil ecosystem function and crop productivity. The two major factors influencing bacterial communities, soil pH and SOC, are ones that can be managed within agricultural fields through application of lime and application of organic amendments (manure, compost), respectively. Thus, it may be possible to manipulate the diversity of bacterial communities in agricultural fields at spatial scales of 10–100 m that are relevant to management of agricultural fields. Supplementary data to this article can be found online at https:// doi.org/10.1016/j.geoderma.2018.08.016.

Cambouris, A.N., et al., 2014. Precision agriculture in potato production. Potato Res. 57, 249–262. Cederlund, H., et al., 2014. 2014. Soil carbon quality and nitrogen fertilization structure bacterial communities with predictable responses of major bacterial phyla. Appl. Soil Ecol. 84, 62–68. Cole, J.R., et al., 2009. The ribosomal database project: improved alignment and new tools for rRNA analysis. Nucleic Acids Res. 37, D141–D145. Dubayah, R., Rich, P.M., 1995. Topographic solar radiation models for GIS. Int. J. Geogr. Inf. Syst. 9, 405–419. Duffera, M., White, J.G., Weisz, R., 2007. Spatial variability of Southeastern U. S. Coastal Plain soil physical properties: implication for site-specific management. Geoderma 137, 327–339. Edgar, R.C., et al., 2011. UCHIME improves sensitivity and speed of chimera detection. Bioinformatics 27, 2194–2200. Eilers, K.G., Lauber, C.L., Knight, R., Fierer, N., 2010. Shifts in bacterial community structure associated with inputs of low molecular weight carbon compounds to soil. Soil Biol. Biochem. 42, 896–903. Enwall, K., et al., 2010. Soil resources influence spatial patterns of denitrifying communities at scales compatible with land management. Appl. Environ. Microbiol. 76, 2243–2250. Fierer, N., Jackson, J.A., 2006. The diversity and biogeography of soil bacterial communities. Proc. Natl. Acad. Sci. U. S. A. 103, 626–631. Fierer, N., Bradford, M.A., Jackson, R.B., 2007. Toward an ecological classification of soil bacteria. Ecology 88, 1354–1364. Franklin, R.B., Mills, A.L., 2003. Multi-scale variation in spatial heterogeneity for microbial community structure in an eastern Virginia agricultural field. FEMS Microbiol. Ecol. 44, 335–346. Gougoulias, C., Clark, J.M., Shaw, L.J., 2014. The role of soil microbes in the global carbon cycle: tracking the below-ground microbial processing of plant-derived carbon for manipulating carbon dynamics in agricultural systems. J. Sci. Food Agric. 94, 2362–2371. Griffiths, R.S., et al., 2000. Rapid method for co-extraction of DNA and RNA from natural environments for analysis of ribosomal DNA- and rRNA-based microbial community composition. Appl. Environ. Microbiol. 66, 5488–5491. Griffiths, R.I., et al., 2011. The bacterial biogeography of British soils. Environ. Microbiol. 13, 1642–1654. Hargreaves, S.K., William, R.J., Hofmockel, K.S., 2015. Environmental filtering of communities in agricultural soil shifts with crop growth. PLoS One 10, e0134345. Hendershot, W.H., Lalande, H., Duquette, M., 2008. Soil reaction and exchangeable acidity. In: Carter, M.R., Gregorich, E.G. (Eds.), Soil Sampling and Methods of Analysis. 2008. CRC Press, Taylor & Francis Group, Boca Raton, pp. 173–178. Holland, P.G., Steyn, D.G., 1975. Vegetational responses to latitudinal variations in slope angle and aspect. J. Biogeogr. 2, 179–183. Jangid, K., et al., 2008. Relative impacts of land-use, management intensity and fertilization upon soil microbial community structure in agricultural systems. Soil Biol. Biochem. 40, 2843–2853. Kemmitt, S.J., Wright, D., Goulding, K.W.T., Jones, D.L., 2006. pH regulation of carbon and nitrogen dynamics in two agricultural soils. Soil Biol. Biochem. 38, 898–911. Klindworth, A., et al., 2013. Evaluation of general 16S ribosomal RNA gene PCR primers for classical and next generation sequencing-based diversity studies. Nucleic Acids Res. 41, e1. https://doi.org/10.1093/nar/gks808. Kosmas, C., Gerontidis, S., Marathianou, M., Detsis, B., Zafiriou, T., Muysen, W.V., Govers, G., Quine, T., Vanoost, K., 2001. The effects of tillage displaced soil on soil properties and wheat biomass. Soil Tillage Res. 58, 31–44. Kroetsch, D., Wang, C., 2008. Particle size distribution. In: Carter, M.R., Gregorich, E.G. (Eds.), Soil Sampling and Methods of Analysis. CRC Press, Taylor & Francis Group, Boca Raton, pp. 714–720. Lauber, C.L., et al., 2008. The influence of soil properties on the structure of bacterial and fungal communities across land-use types. Soil Biol. Biochem. 40, 2407–2415. Lauber, C.L., et al., 2009. Pyrosequencing-based assessment of soil pH as a predictor of soil bacterial community composition at the continental scale. Appl. Environ. Microbiol. 75, 5111–5120. Li, S., et al., 2008. Patterns of tillage and water erosion on topographically complex landscapes in the North America Great Plains. J. Soil Water Conserv. 63, 37–46. Li, S., et al., 2011. Extracting topographic characteristics of landforms typical of Canadian agricultural landscapes for agri-environmental modeling. I. Methodology. Can. J. Soil Sci. 91, 251–266. Moore, I.D., et al., 1993. Soil attribute prediction using terrain analysis. Soil Sci. Soc. Am. J. 57, 443–452. Needleman, S.B., Wunsch, C.D., 1970. A general method applicable to the search for similarities in the amino acid sequence of two proteins. J. Mol. Biol. 48, 443–453. Nunan, N., et al., 2005. Links between plant and rhizoplane bacterial communities in grassland soils, characterized using molecular techniques. Appl. Environ. Microbiol. 71, 6784–6792. O'Brien, S.L., et al., 2016. Spatial scale drives patterns in soil bacterial diversity. Environ. Microbiol. 18, 2039–2051. Oksanen, J., et al., 2015. Vegan: community ecology package. In: R Package Vegan Version 2.3-0, . http://www.R-project.org/. Osborne, C.A., et al., 2011. The influence of sampling strategies and spatial variation on the detected soil bacterial communities under three different land-use types. FEMS Microbiol. Ecol. 78, 70–79. Peacock, A.D., et al., 2001. Soil microbial community responses to dairy manure or ammonium nitrate applications. Soil Biol. Biochem. 33, 1011–1019. Pennock, D.J., Zebarth, B.J., de Jong, E., 1987. Landform classification and soil distribution in hummocky terrain, Saskatchewan, Canada. Geoderma. 40, 297–315. Pennock, D.J., Yates, T., Braidek, J., 2008. Soil sampling designs. In: Carter, M.R.,

Acknowledgement Thanks to Ginette Decker, Karen Terry and Christian Gallant for invaluable help in soil sampling and soil analysis. Thanks for Dr. Aaron Mills for critically reading the manuscript. Thanks to the potato grower for granting access to his field. This work was supported by the Agri-Innovation Program of Agriculture and Agri-Food Canada. There are no conflicts of interest to be reported. References Birkeland, P.W., 1999. Soils and Geomorphology. Oxford University Press, New York. Bokulich, N.A., et al., 2013. Quality-filtering vastly improves diversity estimates from Illumina amplicon sequencing. Nat. Methods 10, 57–59. Bossio, D.A., et al., 1998. Determinants of soil microbial communities: effects of agricultural management, season, and soil type on phospholipid fatty acid profiles. Microb. Ecol. 36, 1–12. Bray, J.R., Curtis, J.T., 1957. An ordination of upland forest communities of southern Wisconsin. Ecol. Monogr. 27, 325–349. Cambardella, C.A., et al., 1994. Field-scale variability of soil properties in central Iowa soils. Soil Sci. Soc. Am. J. 58, 1501–1511.

121

Geoderma 335 (2019) 112–122

S. Neupane et al.

Schloss, P.D., et al., 2009. Introducing mothur: open-source, platform-independent, community-supported software for describing and comparing microbial communities. Appl. Environ. Microbiol. 75, 7537–7541. Seibert, J., Stendahl, J., Sorensen, R., 2007. Topographic influence on soil properties in boreal forest. Geoderma 121, 139–148. Skjemstad, J.O., Baldock, J.A., 2008. Total and organic carbon. In: Carter, M.R., Gregorich, E.G. (Eds.), Soil Sampling and Methods of Analysis. CRC Press, Taylor & Francis Group, Boca Raton, pp. 225–237. Wang, Q., et al., 2007. Naïve Bayesian classifier for rapid assignment of rRNA sequences into the new bacterial taxonomy. Appl. Environ. Microbiol. 73, 5261–5267. Yilmaz, P., et al., 2014. The SILVA and “all-species living tree project (LTP)” taxonomic frameworks. Nucleic Acids Res. 42, D643–D648. Zebarth, B.J., et al., 2002. Soil variation within a hummocky podzolic landscape under intensive potato production. Geoderma 110, 19–33. Zinger, L., et al., 2011. Contrasting diversity patterns of Crenarchaeal, bacterial and fungal soil communities in an alpine landscape. PLoS One 6, e19950.

Gregorich, E.G. (Eds.), Soil Sampling and Methods of Analysis, 2nd edn. CRC Press, Boca Raton, pp. 1–14. Philippot, L., Spor, A., Henault, C., et al., 2013. Loss in microbial diversity affects nitrogen cycling in soil. ISME J. 7, 1609–1619. R Development Core Team, 2013. R: A Language and Environment for Statistical Computing. R Foundation for statistical computing, Vienna, Austria. http://www.Rproject.org/. Ramirez, K.S., et al., 2010. Consistent effects of nitrogen fertilization on soil bacterial communities in contrasting systems. Ecology 91, 3463–3470. Ritz, K., et al., 2004. Spatial structure in soil chemical and microbiological properties in an upland grassland. FEMS Microbiol. Ecol. 49, 191–205. Rosenzweig, N., et al., 2016. Geostatistical approach to visualize the diversity of soil inhabiting bacteria and edaphic qualities in potato (Solanum tuberosum) production systems. Am. J. Potato Res. 93, 518–532. Rousk, J., et al., 2010. Soil bacterial and fungal communities across a pH gradient in an arable soil. ISME J. 4, 1340–1351.

122