Bacterial and fungal communities respond differently to varying tillage depth in agricultural soils

In arable cropping systems, reduced or conservation tillage practices are linked with improved soil quality, C retention and higher microbial biomass, but most long-term studies rarely focus on depths greater than 15 cm nor allow comparison of microbial community responses to agricultural practices. We investigated microbial community structure in a long-term field trial (12-years, Lincoln, New Zealand) established in a silt-loam soil over four depth ranges down to 30 cm. Our objectives were to investigate the degree of homogenisation of soil biological and chemical properties with depth, and to determine the main drivers of microbial community response to tillage. We hypothesised that soil microbiological responses would depend on tillage depth, observed by a homogenisation of microbial community composition within the tilled zone. Tillage treatments were mouldboard plough and disc harrow, impacting soil to ∼20 and ∼10 cm depth, respectively. These treatments were compared to a no-tillage treatment and two control treatments, both permanent pasture and permanent fallow. Bacterial and fungal communities collected from the site were not impacted by the spatial location of sampling across the study area but were affected by physicochemical changes associated with tillage induced soil homogenisation and plant presence. Tillage treatment effects on both species richness and composition were more evident for bacterial communities than fungal communities, and were greater at depths <15 cm. Homogenisation of soil and changing land management appears to redistribute both microbiota and nutrients deeper in the soil profile while consequences for soil biogeochemical functioning remain poorly understood.

72 in untilled soil to communities in soil tilled to depths of either 10 or 20 cm. We also expected 73 fungal communities to be more prone to disturbance from ploughing because of their extensive 74 hyphal networks (Wardle 1995). Therefore, our objectives were to investigate the degree of 75 homogenisation of soil biological and chemical properties to depths of 30 cm, and to identify the 76 main drivers of microbial community responses to tillage intensity. 89 Intensive tillage (Ii): cultivation to ~20 cm using a mouldboard plough, followed by secondary 90 cultivation (one pass with a spring tined implement followed by harrowing and rolling twice). 91 All tillage operations were carried out using standard commercial equipment. Spring-sown main 92 crops rotation included barley, wheat, and peas. They were followed by winter-grazed (sheep) 93 cover crops (oats or forage brassicas). All crops were sown using a Great Plains direct drill.
94 Fertiliser (N and P) were applied to the spring crops to ensure these nutrients were not limiting.
119 Soil chemical analysis 120 Gravimetric soil moisture content was determined by the mass difference before and after drying 121 at 105°C for 16 h. The pH of each sample was determined using a glass electrode at 1:2 field 122 moist sample to water ratio (Hendershot et al. 2008). Bulk density (< 4 mm) was calculated from 123 the weight of field-moist soil of known volume, corrected for its stone and moisture contents.
124 Aggregate stability or mean weight diameter (MWD) was determined by first separating 2-4 mm 125 aggregates from whole soil by sieving, and then air-drying them at 25˚C before aggregate 126 stability determination using a wet-sieving method (Kemper & Rosenau 1986). The air-dried 2−4 127 mm aggregates (50 g) were sieved underwater for 20 minutes on a nest of sieves (2.0, 1.0 and 0.5 128 mm diameter). The soil remaining on each sieve was weighed after oven drying at 105˚C. The 133 Exchangeable acidity (Exch. Acid.) and aluminium (Exch. Al.) was determined by extraction 134 using 1 M KCl. The amount of H + and Al 3+ in the extracts was determined by titration as 135 described by Sims (1996). Total carbon (C) and nitrogen (N) contents were determined by the 136 Dumas dry combustion method at 950°C using a Truspec C/N analyzer (LECO, St. Joseph, 137 Michigan, USA). Manuscript to be reviewed 139 Microbial biomass C (MBC) and N (MBN) were determined by chloroform fumigation-140 extraction as described by Sparling & West (1988). Pre-and post-fumigation extracts were 141 analysed for organic C by combustion catalytic oxidation using a TOC-V CSH analyzer (Shimadzu 142 Corporation, Kyoto, Japan) and for organic N by the persulfate oxidation method described by We used the aov function in R version 2.14 (R Core Team 2012) to perform analyses of variance 211 on soil chemical data using a two-way layout (treatment; depth), with interaction terms.
212 Canonical redundancy analysis (RDA) and was used to summarise variation in the bacterial and 213 fungal community data that could be explained by our set of explanatory variables (e.g., pH, soil 214 water content). Variance partitioning was then performed using the function varpart.MEM in R,   Figure S1). With the exception of soil water content, the greatest difference 250 among treatments was again between the non-till control treatments Pp and Pf, and the biggest 251 difference among depths was between 0 -7.5 cm and 15 -25 cm, noting that chemical data was 252 never obtained from the deepest (25-30 cm) samples. Previous research at this field site has 253 indicated that crop productivity is not influenced by tillage, neither is nutrient input (pers. Manuscript to be reviewed 300 Two-way ANOVA confirmed that relative bacterial OTU richness (variety or number of 301 OTUs) in the 0 -7.5 cm depth was greater than at lower depths ( Fig. 4; P < 0.001), but did not 302 differ among treatments. In contrast, fungal richness did not significantly differ by depth or 303 treatment perhaps also explaining why variance partitioning showed that soil chemical properties