Introduction

The structural changes in Central European agriculture and the increase in price of fuel have forced the replacement of cost intensive mouldboard ploughing by simplified tillage systems (Nail et al. 2007; Koch et al. 2009). A reduction in tillage intensity by non-inversion systems promotes the activity of earthworms (Ehlers 1975), which increases the water infiltration capacity of soils and reduces erosion risk (Tebrügge and Düring 1999; Soane et al. 2012). This is especially important for sugar beet cultivation, which leaves large soil areas uncovered in late spring during a period when heavy rainfall events regularly occur in central Europe (Clemens and Stahr 1994; Prasuhn 2012). This problem is intensified by the fact that sugar beet is mainly cropped on silt loams, especially sensitive to water erosion (Koch et al. 2009). Moreover, non-inversion tillage offers the potential to sequester organic C (Freibauer et al. 2004; Jacobs et al. 2009) and often increases the stocks of soil microbial biomass (Stockfisch et al. 1999; Heinze et al. 2010a), but not in all cases (Ahl et al. 1998).

Non-inversion tillage systems generally seem to promote fungi (Hendrix et al. 1986; Frey et al. 1999), especially biotrophic arbuscular mycorrhizal fungi (AMF) (Kabir et al. 1998; Kabir 2005), in most cases also saprotrophic fungi (Ahl et al. 1998), but not always (Heinze et al. 2010a; Strickland and Rousk 2010; Jacobs et al. 2011). Fungi are generally able to take up large amounts of S into their biomass (Banerjee and Chapman 1996), and thus, a close relationship has been observed between microbial biomass S and the fungal biomarker ergosterol (Heinze et al. 2010b). The most important indicators for saprotrophic fungi are currently the cell-membrane components linoleic acid (18:2ω6,9) and ergosterol (Joergensen and Wichern 2008). However, ergosterol is more specific for saprotrophic fungi and simpler to measure than linoleic acid. Neither of these cell-membrane components are suitable to estimate the biomass of AMF (Olsson et al. 2003; Ruess and Chamberlain 2010).

In contrast, the cell-wall component glucosamine occurs in the chitin of AMF and saprotrophic fungi, as well as in the murein of bacteria (Amelung 2001; Appuhn and Joergensen 2006). The bacterial contribution to the glucosamine content of soils can be estimated by the highly specific muramic acid, which occurs only in bacteria (Engelking et al. 2007). As cell-wall components accumulate in soil organic matter (SOM) (Amelung 2001), they are very useful to identify the specific contribution of fungal and bacterial residues to the C sequestration potential of soils (Joergensen and Wichern 2008; Liang and Balser 2011; Miltner et al. 2012). Fungi are said to have a higher C sequestration potential than do bacteria (Guggenberger et al. 1999; Bailey et al. 2002; Jastrow et al. 2007) due to the higher substrate use efficiency, i.e. the lower metabolic quotient of fungi (Sakamoto and Oba 1994). Consequently, amino sugar analysis could serve as a time-integrated biomarker for the contribution of these two main microbial groups to SOM in tillage experiments (van Groenigen et al. 2010; Liang et al. 2011; Martins et al. 2012). This is especially true for soil layers below 30 cm, which have received substantially less attention (Moritz et al. 2009; Murugan and Kumar 2013), although tillage may have strong impact on microbial processes and C sequestration in the subsoil (Wright et al. 2007).

Our study is based on the following three hypotheses: (1) A reduction in tillage intensity (mouldboard plough > grubber > no tillage) generally increases stocks of SOM and especially microbial biomass, independently of site-specific differences in environmental conditions. (2) In the top (0–5 cm) soil layers, this increase is caused by the preferential accumulation of fungal residues. (3) In the bottom (30–40 cm) soil layers, a reduction in tillage intensity generally promotes AMF at the expense of saprotrophic fungi. Our objectives were to test these hypotheses by measuring fungal biomass (ergosterol) and C, N and S stored in the soil microbial biomass and in SOM, as well as by estimating the co-accumulation of fungal and bacterial residues in loess-derived soils down to 40 cm of a mature, 15-year-old on-farm tillage experiment (Koch et al. 2009). The experimental arable sites were located in eastern and southern Germany and characterized by large-scale plots. This made it possible to investigate the interactions between tillage treatments and site effects, which have been shown by Heinze et al. (2010a) to override tillage effects.

A specific focus of our experiment was to consider the vertical and horizontal spatial variability for all soil properties analysed. For this reason, samples were taken once with many replicates per site and tillage treatment and at a high resolution down the profiles in early spring. This is a period at field capacity, before strong root growth takes place and when the ambient temperature is close to the average annual temperature (Anderson and Domsch 1989, 1990). For these reasons, early spring is recommended for taking representative soil samples in soil microbiological monitoring programmes (Höper and Kleefisch 2001; Bloem et al. 2006).

Materials and methods

Experimental site and investigation design

Three tillage systems were investigated at four sites (Friemar, Grombach, Lüttewitz and Zschortau) established in the early 1990s by the agricultural division of Südzucker AG, Mannheim/Ochsenfurt, and the Institute of Sugar Beet Research, Göttingen, Germany, as on-farm long-term field experiments (Koch et al. 2009). The mean annual temperature and precipitation ranges from 7.8 to 9.3 °C and 512 to 776 mm, respectively. Site characteristics, soil pH and texture of the sites are given in Table 1 and Koch et al. (2009).

Table 1 Climatic and soil characteristics (3–27 cm) of the four experimental sites
Full size table

At each site, three similar-sized tillage treatment plots were formed on one large field with spatially homogeneous soil properties. The different tillage systems were annual mouldboard ploughing to a depth of 30 cm (MBT), grubber tillage to a depth of 10–15 cm (GRT) and no-tillage (NT) with direct seed drilling. Before sugar beet sowing, 3–5 cm deep seedbed cultivation was introduced in the NT treatment to improve sugar beet crop establishment. Depending on the site, tillage plot size ranged from 2.5 to 8 ha per treatment. The crop rotation consisted of sugar beet (Beta vulgaris L.)/winter wheat (Triticum aestivum L.)/winter wheat at all sites for the past 15 years (Koch et al. 2009). White mustard (Sinapis alba L.) was sown after harvest of the second wheat as green manure. Crop residues were left on the field, and sugar beet was sown in March–April using a single-seed drill adapted to crop residues lying on the soil surface. The crop management was carried out following the regional standards of agricultural practice, including the use of non-selective herbicides in MBT and NT treatments. Based on the infestation level between treatments, sugar beet selective herbicides, molluscicides and rodenticides were used (Koch et al. 2009). Application of N fertilizer varied between the sites but was identical for all treatments at one site. Nitrogen is applied as calcium ammonium nitrate (~27 % N), potassium as KCl (60 % K2O) and phosphorus as triple superphosphate (46 % P2O5). The mean annual crop yield and the mean annual N fertilization rate from 1999–2010 are presented in Table 2.

Table 2 Treatment-specific mean grain yields for winter wheat and taproot yields for sugar beet; mean N fertilization (nitrogen is applied as calcium ammonium nitrate ~27 % N) rates for all sites and treatments from the year 1999–2010 (in brackets N fertilization rates in 2010, the year of soil sampling)
Full size table

Soil sampling and chemical analysis

All soil samples were taken in April 2010 when the sites were grown with winter wheat crop sown in October 2009. At each large-scale plot of the tillage treatment, three equal blocks were randomly specified. Soil samples were taken from three points 10 m apart in a triangle within each of these three blocks, resulting in nine samples per treatment per site. Samples were taken at 0–5, 5–10, 10–15, 15–20, 20–30, and 30–40 cm depth, using a steel corer with 4 cm diameter. All samples were passed through a 2-mm sieve and stored at 4 °C for a few weeks until the assessment of microbiological properties (Anderson and Domsch 1989, 1990). A field moist soil sample was used to analyse pH (1:2.5 soil–water ratio). Dried (24 h at 105 °C) and finely ground samples were used for chemical analyses (C, N, P and S). Total C and N were determined by gas chromatography using a Vario EL (Elementar, Hanau, Germany) analyser. Concentrations of total P and S were measured after HNO3/pressure digestion (Chander et al. 2008) by ICP-ACS (Spectro Analytical Instruments, Kleve, Germany).

Microbial activity and biomass indices

The basal respiration of soil was measured by the incubation of 60 g soil sample for 7 days at 22 °C with 40 % water holding capacity. The CO2 evolved was trapped in 0.5 M NaOH, and the excess NaOH was back-titrated using 0.5 M HCl after the addition of saturated BaCl2 solution. Fumigated (24 h with ethanol-free CHCl3 at 25 °C) and non-fumigated 5-g samples were extracted with 20 ml of 0.5 M K2SO4 by 30 min horizontal shaking at 200 rpm and filtered (hw3, Sartorius Stedim Biotech, Göttingen, Germany) to measure microbial biomass C and N (Brookes et al. 1985; Vance et al. 1987). Organic C in the extracts was measured as CO2 by infrared absorption after combustion at 850 °C using a Dimatec 100 automatic analyser (Dimatec, Essen, Germany). Microbial biomass C was calculated as EC/k EC, where E C = (organic C extracted from fumigated soil) − (organic C extracted from non-fumigated soil) and k EC = 0.45 (Wu et al. 1990). Total N in the extracts was measured using a Dima-N chemoluminescence detector (Dimatec). Microbial biomass N was calculated as E N/k EN, where E N = (total N extracted from fumigated soil) − (total N extracted from non-fumigated soil) and k EN = 0.54 (Brookes et al. 1985; Joergensen and Mueller 1996). Fumigated and non-fumigated 5-g samples were extracted with 25 ml of 1 M NH4NO3 to measure microbial biomass S (Khan et al. 2009). Microbial biomass S was calculated as E S/k ES, where E S = (total S extracted from fumigated soil) − (total S extracted from non-fumigated soil) and k ES = 0.35 (Saggar et al. 1981; Wu et al. 1994).

The fungal cell membrane component ergosterol was extracted from 2 g of moist soil with 100 ml ethanol (Djajakirana et al. 1996). Then, ergosterol was determined by reverse-phase HPLC with 100 % methanol as the mobile phase and detected at a wavelength of 282 nm. The amino sugars (muramic acid, mannosamine, glucosamine and galactosamine) were determined according to Appuhn et al. (2004), as described by Indorf et al. (2011). Moist samples of 0.5 g soil were weighed into 20 ml test tubes, mixed with 10 ml 6 M HCl and hydrolysed for 6 h at 105 °C. After HCl removal from the filtered hydrolysates in a vacuum rotary evaporator at 40 °C and centrifugation (10 min at 5,000×g), the samples were transferred to vials and stored at −18 °C until the HPLC measurement. Chromatographic separations were performed on a Phenomenex (Aschaffenburg, Germany) Hyperclone C18 column (125 mm length × 4 mm diameter), protected by a Phenomenex C18 security guard cartridge (4 mm length × 2 mm diameter) at 35 °C. The HPLC system consisted of a Dionex (Germering, Germany) P 580 gradient pump, a Dionex Ultimate WPS 3000TSL analytical auto sampler with in-line split-loop injection and thermostat and a Dionex RF 2000 fluorescence detector set at 445-nm emission and 330-nm excitation wavelengths. For the automated pre-column derivatisation, 50 ml ortho-phthaldialdehyde and 30 ml sample were mixed in the preparation vial, and after 120 s reaction time, 15 ml of the indole derivatives was injected. The mobile phase consisted of two eluents and was delivered at a flow rate of 1.5 ml min−1. Eluent A was a 97.8/0.7/1.5 (v/v/v) mixture of an aqueous phase, methanol and tetrahydrofuran (THF). The aqueous phase contained 52 mmol sodium citrate and 4 mmol sodium acetate, adjusted to pH 5.3 with HCl. Then, methanol and THF were added. Eluent B consisted of 50 % water and 50 % methanol (v/v).

Calculation and statistical analysis

The stocks of the soil nutrients and microbial indices at different depths were calculated on a volume basis by taking the bulk density of the respective soil layer into account. Bulk density was calculated from core dry weight divided by volume. Fungal glucosamine was calculated by subtracting bacterial glucosamine from total glucosamine as an index for fungal residues, assuming that muramic acid and glucosamine occur at a 1 to 2 ratio in bacteria (Engelking et al. 2007): fungal C (μg g-1 dry weight) = (mmol glucosamine − 2 × mmol muramic acid) × 179.2 gmol−1 × 9, where 179.2 is the molecular weight of glucosamine and 9 is the conversion value of fungal glucosamine to fungal C (Appuhn and Joergensen 2006). Bacterial C (microgrammes per gramme of dry weight) was calculated as an index for bacterial residues by multiplying the concentration of muramic acid in microgrammes per gramme of dry weight by 45 (Appuhn and Joergensen 2006). The values of three sampling points from three blocks per site were used to calculate mean concentrations and stocks for every site (n = 9).

The results presented in the tables are arithmetic means of the stocks and are given on an oven dry basis (105 °C, 24 h). The blocks in each treatment are considered as pseudo-replicates. The mean concentrations per site of each parameter served as field replicates. The concentrations and ratios presented in Figs. 2a, b and 3a–f are the arithmetic means of the tillage treatment in the respective depth from four sites as field replications (n = 4). The mean value of the four field replications was used for a statistical evaluation of the tillage treatment effect (n = 4). Data were checked for normal distribution by Chi-square test. When necessary, the data were ln-transformed. The significance of tillage treatment and site effects on stock parameters was analysed by a MANOVA, with depth as repeated measures and blocks as replicates. All statistical analyses were carried out using JMP 7.0 (SAS Inst. Inc.).

Results

Tillage effects at different depths

There were no significant effects of blocks in the studied soil properties (Table 3). At 0–30 cm depth, MBT led to a homogeneous bulk density (Fig. 1) and a homogeneous distribution of soil organic C (SOC, Fig. 2a), total N, total P and total S (Supplementary Table 1, a and b), microbial biomass C (Fig. 2b), N and S (Supplementary Table 2, a and b) as well as microbial amino sugars (Supplementary Table 3, a and b). In contrast, GRT and NT led to a significant 20 % increase in bulk density from the 0–5 cm layer down to the 20–30 cm layer (Fig. 1). The mean soil bulk density did not differ between the tillage treatments at 0–40 cm depth, but it was significantly higher in the MBT treatment at 0–5 cm depth and significantly lower at 20–30 cm depth in comparison with the GRT and NT treatments. The depth-specific differences between the GRT and NT treatments were in most cases not significant for bulk density, soil chemical properties (Fig. 2a, Supplementary Table 1, a and b) as well as for microbial biomass indices (Fig. 2b, Supplementary Table 2, a and b) and microbial amino sugars (Supplementary Table 3, a and b). At 30–40 cm, the concentrations of soil chemical and microbiological parameters were generally lower than at 20–30 cm depth (Fig. 2, Supplementary Tables 1 to 2) and did not reveal clear tillage effects, although the bulk density of the GRT and NT treatments significantly exceeded that of the MBT treatment (Fig. 1).

Table 3 The effects of different tillage treatments on mean stocks of soil organic C (SOC), total N, total P and total S at varying soil depths and investigation sites
Full size table
Fig. 1
figure 1

Effect of soil tillage on mean bulk density at different soil depths (0–40 cm), means of four sites; *P < 0.05, indicates a depth-specific significant difference between the three tillage treatments (ANOVA repeated measurement); bars represent standard errors of the mean (n = 4)

Full size image
Fig. 2
figure 2

Mean contents over four sites (n = 54 per site per depth) of a soil organic C and b soil microbial biomass C in soils from three tillage treatments at different depths; *P < 0.05, indicates a depth-specific significant difference between the three tillage treatments (ANOVA repeated measurement); bars represent standard errors of the mean (n = 4)

Full size image

The concentrations not only of SOC (Fig. 2a), but also those of total N, total P and total S (Supplementary Table 1, a and b) as well as those of microbial amino sugars (Supplementary Table 3, a and b), declined by 50 % from the 0–5 cm down to the 20–30 cm layer in the GRT and NT treatments, whereas no depth gradient was observed in the MBT treatment. The SOC concentration was 40 % higher in the GRT and NT treatments in the 0–5 cm layer and 10 % lower at 20–30 cm depth in comparison with the MBT treatment (Fig. 2a). Microbial biomass C was 150 % higher in the GRT and NT treatments in the 0–5 cm layer and 30 % lower at 20–30 cm depth in comparison with the MBT treatment (Fig. 2b). Microbial biomass C, N and S (Supplementary Table 2, a and b) declined roughly by 70 % from the 0–5 cm down to the 20–30 cm layer in the GRT and NT treatments, whereas again no depth gradient was observed in the MBT treatment.

At 0–30 cm depth, the long-term GRT and NT treatments did not affect the soil C/N ratio (overall mean, 10.2; Fig. 3a), the amino sugar-based microbial residual C to SOC ratio (overall mean, 0.75; Fig. 3b) and the fungal C-to-bacterial C ratio (overall mean 2.2). The microbial biomass C to SOC ratio and the ergosterol to microbial biomass C ratio decreased by roughly 40 % (Fig. 3c) and 60 % (Fig. 3f), respectively, from the 0–5 cm down to the 20–30 cm layer in the GRT and NT treatments. In contrast, the metabolic quotient qCO2 was the only ratio (Fig. 3d) that significantly increased, by 50 % from 0–5 cm to 20–30 cm depth (Fig. 3d). The metabolic quotient qCO2 did not show strong depth effects at 0–30 cm in the GRT and NT treatments (Fig. 3d). However, the tillage-specific mean was significantly higher in the MBT treatment in comparison with the GRT and NT treatments. The microbial biomass C to SOC ratio was 40 % higher in the GRT and NT treatments in the 0–5 cm layer and 15 % lower at 20–30 cm depth in comparison with the MBT treatment (Fig. 3c). Similarly, the ergosterol to microbial biomass C ratio was 40 % higher in the GRT and NT treatments in the 0–5 cm layer and 40 % lower at 20–30 cm depth in comparison with the MBT treatment (Fig. 3f). At 30–40 cm, the soil C/N ratio (Fig. 3a), the metabolic quotient qCO2 (Fig. 3d) and the fungal C to bacterial C ratio (Fig. 3e) significantly increased, whereas the ergosterol to microbial biomass C (Fig. 3f) and the microbial biomass C to SOC ratio (Fig. 3c) decreased in the GRT and NT treatment compared to the MBT treatment.

Fig. 3
figure 3

Mean ratios over four sites (n = 54 per site per depth) of a soil C to N ratio, b microbial C to soil organic C ratio, c microbial biomass C to soil organic C (percent), d the mean metabolic quotient qCO2 (mg CO2–C g microbial biomass C−1 day−1), e fungal C-to-bacterial C ratio and f ergosterol to microbial biomass C (percent) in soils with three tillage treatments at different depths; *P < 0.05, indicates a depth-specific significant difference between the three tillage treatments (ANOVA repeated measurement); bars represent standard errors of the mean (n = 4)

Full size image

The fungal biomass indicator ergosterol showed significant positive linear relationships with the microbial biomass S (Fig. 4a) and amino sugar-based fungal C (Fig. 4b). A significant positive linear relationship was observed for the ergosterol to microbial biomass C ratio and the SOC concentration (Fig. 5a), whereas a significant negative relationship was found for the fungal C to bacterial C and the SOC concentration (Fig. 5b).

Fig. 4
figure 4

The relationships between a soil microbial biomass S and ergosterol and b ergosterol and fungal C; data from four sites, three tillage treatments and all field replicates were combined (n = 648)

Full size image
Fig. 5
figure 5

The relationships between a soil organic C and ergosterol to microbial biomass C (percent) and b soil organic C and fungal C to bacterial C ratio; data from four sites, three tillage treatments and all field replicates were combined (n = 648)

Full size image

Tillage and site effects on stocks

At 0–30 cm depth, the stocks of SOC, total nutrient elements (N, P, and S; Table 3), soil microbial and fungal biomass indices (Table 4), as well as amino sugars and related indices (Table 5), were significantly higher in the GRT and NT treatments in comparison with the MBT treatment. The positive GRT and NT effects on soil chemical properties declined in the order total N = total S (+14 %) > SOC (7 %) > total P (+5 %) (Table 3). The GRT and NT treatments increased soil microbial biomass C and N as well as ergosterol by 20 % and muramic acid, galactosamine and glucosamine by 8 % but had no effect on microbial biomass S and mannosamine in comparison with the MBT treatment (Tables 4 and 5). The differences between NT and MBT were generally smaller than those between GRT and MBT and not always significant, i.e. for total P, total S (Table 3) and microbial biomass S (Table 4). The stock of mannosamine was significantly lowest in the NT treatment (Table 5).

Table 4 The effects of different tillage treatments on mean stocks of microbial biomass C, N and S, and ergosterol at varying soil depths and investigation sites
Full size table
Table 5 The effect of different tillage treatments on the distribution of mean soil amino sugar stocks at varying soil depths and investigation sites
Full size table

In contrast to the top soil (0–30 cm), in the bottom soil layer at 30–40 cm, the stocks of SOC, total N and total P (Table 3) as well as the stocks of amino sugars and related indices (Table 5) were higher in the MBT and GRT treatments in comparison with the NT treatment. The stock of total S was the highest in the GRT treatment (Table 3). The stock of soil microbial biomass C was significantly higher in the MBT treatment followed by the GRT treatment, while the lowest microbial biomass C stock was observed in the NT treatment (Table 4). The MBT treatment led to a significantly (P < 0.05) higher microbial biomass N and ergosterol stocks than the GRT and NT treatments. The stock of microbial biomass S was not significantly affected by the tillage treatments.

At the site Friemar, the highest stocks were observed for SOC and total N (Table 3); microbial biomass C, N and S; and ergosterol (Table 4) as well as for amino sugars and related residue indices (Table 5). At the site Grombach, the highest stocks were found for total P and total S (Table 3). At the sites Lüttewitz and Zschortau, the lowest stocks were observed for SOC, total N, and total S (Table 3), and ergosterol and microbial biomass N and S (Table 4) as well as for muramic acid, mannosamine and galactosamine (Table 5). The lowest stocks of microbial biomass C (Table 4) and glucosamine (Table 5) occurred at the site Lüttewitz, whereas the lowest total P stock was obtained at the site Zschortau (Table 3).

Discussion

After 15 years, a reduction in tillage intensity by the GRT and NT treatments increased stocks of SOM and especially microbial biomass at 0–40 cm depth in comparison with the MBT treatment. Our results indicate significant (P < 0.05) tillage effects in loess-derived silt loams suitable for rain-fed sugar beet production and verify our first hypothesis. Similar effects of a reduction in tillage intensity on SOM and microbial biomass stocks have been observed by Stockfisch et al. (1999) and by Heinze et al. (2010a), after 20 and 30 experimental years, respectively, with the rotary cultivator and mouldboard plough treatments. Effects of a reduction in tillage depth on SOC stocks at 0–30 or 0–40 cm depth (Ahl et al. 1998; Stockfisch et al. 1999) might be due to differences in aboveground crop yield (Koch et al. 2009), but especially in belowground root development (Qin et al. 2006). Differences in C input are probably the main reason for the small but more positive effects of the GRT treatment on SOC and microbial biomass C stocks in comparison with the NT treatment, which has the lowest crop yields, especially for sugar beet (Table 2). The lowest sugar beet yields in the NT treatment were caused by low population density (Koch et al. 2009), presumably in combination with an unequal distribution of wheat straw by the combine harvester, which intensifies unfavourable conditions by direct seed drilling for embedding of seeds, field emergence and plant establishment (Koch et al. 2009).

Seriously compacted layers were not observed in any of the current tillage treatments, not even below 30 cm depth in the MBT treatment, which has been repeatedly observed in silt loams (Tessier et al. 1997; Ahl et al. 1998; Heinze et al. 2010a). Consequently, root penetration resistance did not differ between the tillage treatments at the same experimental sites (Koch et al. 2009). Another reason for the increased SOC stocks in the current GRT and NT treatments might be an increased formation of macro-aggregates at 0–25 cm depth in comparison with the MBT treatment (Andruschkewitsch et al. 2013), which physically protect SOC against microbial decomposition and mechanical breakdown (Jacobs et al. 2009). However, this occlusion in macro-aggregates does not explain the increased stocks of soil microbial biomass in the GRT and NT treatments, due to the lower yields, especially in the NT treatment, and consequently the lower inputs by harvest and root residues compared to the MBT treatment (Table 2; Koch et al. 2009). Most likely reasons for the tillage-induced increase in stocks of microbial biomass are interactions between the microbial turnover, substrate availability, water content, temperature and bulk density (Taylor et al. 2004; Poll et al. 2008; Jacobs et al. 2011; Chaparro et al. 2012). The higher SOC concentration in the GRT and NT treatments certainly improves the conditions for microbial turnover in dry periods of the year. However, a higher water content in combination with higher bulk density and reduced pore space might reduce the microbial turnover due to lower soil temperatures and a reduced O2 transfer into deeper parts of the A horizon.

A tillage-induced reduction in microbial turnover in the GRT and NT treatments was indicated by the inverse relationship between the metabolic quotient qCO2 and the microbial biomass C to SOC ratio, repeatedly observed in tillage experiments (Hungria et al. 2009; Heinze et al. 2010a; Murugan et al. 2013). The metabolic quotient qCO2 describes the amount of substrate used to build up microbial biomass (Anderson and Domsch 1990), and the microbial biomass C to SOC ratio is an important index for the SOC availability to soil microorganisms (Anderson and Domsch 1989). For this reason, the decreased input of fresh plant material and an increased recalcitrance of SOC usually lead to an inverse relationship between qCO2 values and microbial biomass C to SOC ratios (Lavahun et al. 1996; Meyer et al. 1996; Goberna et al. 2006).

The differences in microbial biomass stocks may also be caused by the different development in the microbial community structure, which is reflected by the ergosterol to microbial biomass C ratio. The cell-membrane component ergosterol is an important indicator for saprotrophic fungi (Joergensen and Wichern 2008) but does not occur in AMF (Olsson et al. 2003). A positive correlation between ergosterol and fungal C (Fig. 4b) has been observed in silt loam soils that forms a gradient in SOM (Appuhn et al. 2006). We observed a positive correlation between ergosterol and microbial biomass S (Fig. 4a) which was not found in a long-term fertilization experiment in a tropical rice field (Murugan and Kumar 2013), two tillage experiments on silt loams (Heinze et al. 2010a) and our one-season tillage experiment in a grassland soil (Murugan et al. 2013). However, all three indices for the presence of fungi differed in their response to the tillage treatment, although they were generally interrelated.

The ergosterol to microbial biomass C ratio showed a significant positive linear relationship with SOC (Fig. 5a), indicating an increased supply of fresh root and harvest residues, which promotes saprotrophic fungi. Especially the regular incorporation of straw is known to have strong positive effects on the ergosterol concentration in soil (Scheller and Joergensen 2008; Heinze et al. 2010b; Rousk et al. 2010). The higher contribution of fungi to the microbial community at higher SOC concentrations is in line with the view repeatedly stated by others (Bailey et al. 2002; Jastrow et al. 2007). The significantly higher ergosterol to microbial biomass C ratio and the SOC concentration suggests that more saprotrophic fungi occurred under the GRT and NT treatments at 0–10 cm depth in comparison with the MBT treatment. This also means that the presence of more saprotrophic fungi is combined with a lower turnover of the microbial biomass based on the inverse relationship between the microbial biomass C to SOC ratio and the qCO2. This relationship has been observed by Sakamoto and Oba (1994) for soil fungi using direct microscopy. However, their measurements of fungal biomass included biotrophic AMF, which have a low saprotrophic capability, which leads to low basal respiration rates and consequently low qCO2 values. This view is supported by the negative relationship between the ergosterol to microbial biomass C ratio and the fungal C to bacterial C ratio, which suggest that the residues of saprotrophic fungi have a stronger turnover than those of AMF.

In contrast to the ergosterol to microbial biomass C ratio, the ratio of fungal C to bacterial C showed a significant negative linear relationship with SOC (Fig. 5b; Murugan et al. 2013), despite the general positive interrelationships between the three fungal indices. This is mainly due to the strong increase in the fungal C to bacterial C ratio at 30–40 cm depth in the GRT and NT treatments and the relatively low values of this ratio at 0–5 cm depth in all treatments, which contrasts our second hypothesis. At 30–40 cm depth, the increase in the fungal C to bacterial C ratio is combined with a strong increase in the soil C/N ratio, which indicates accumulation of recalcitrant SOC in comparison with the soil at 0–30 cm depth (John et al. 2005; Yamashita et al. 2006; Lauber et al. 2008; Chaparro et al. 2012). The high metabolic quotient qCO2 and the lower microbial residual C to SOC ratio are additional indications for the presence of recalcitrant organic matter at 30–40 cm depth in the GRT and NT treatments (Meyer et al. 1996). At 30–40 cm depth, the higher fungal C to bacterial C ratio and lower ergosterol to microbial biomass ratio in the GRT and NT treatments are likely caused by a shift within the fungal community structure towards AMF, which confirms our third hypothesis. In contrast, the MBT treatment with straw incorporation led to increased saprotrophic fungi at 30–40 cm depth. However, this negative relationship between these two ratios contrasts with the results of our one-season tillage experiment (Murugan et al. 2013), probably due to the high contribution of AMF to the fungal biomass in grassland soil, as indicated by very low ergosterol to microbial biomass C ratios (Djajakirana et al. 1996). All soil microbial indices clearly demonstrate the absence of a plough pan in the GRT and NT treatment, whereas no significant differences in the bulk density were observed between 20–30 and 30–40 cm depth, both being considerably higher than that in the MBT treatment.

It is more difficult to explain the low fungal C to bacterial C ratio at 0–5 cm depth in the presence of the highest ergosterol to microbial biomass C ratios. The reverse could be easily explained by the difference in the fungal community structure, as AMF do not contain ergosterol but glucosamine (Olsson et al. 2003; Joergensen and Wichern 2008). One explanation might be that, in the strong presence of saprotrophic fungi, a disproportionate increase in the turnover of fungal biomass may have led to a lower accumulation of fungal residues in comparison with bacterial residues. Differences in the turnover between fungal and bacterial cell-wall residues have been repeatedly claimed in the literature (Amelung 2001; Six et al. 2006; Strickland and Rousk 2010; Liang et al. 2011), although not really proven on the basis of bacterial and fungal cells added to soil (Jenkinson 1976). However, cultured bacteria and fungi might behave differently in soil than autochthonous soil microorganisms. One way to obtain more accurate information is to use amino-sugar specific δ13C analysis (Amelung et al. 2008; Bodé et al. 2009) in the range of natural abundance (Indorf et al. 2012). The fractionation of δ13C sometimes observed between different fungal tissue, organic substrates and SOM (Hobbie et al. 2004; Zeller et al. 2007) might give information on differences in the turnover of fungal and bacterial cell wall residues.

Nearly all elemental and physiological ratios obtained in the present experiment are in the range obtained in the respective literature, such as the microbial biomass C/N ratio (Joergensen and Emmerling 2006), the microbial biomass C/S ratio (Heinze et al. 2010a, b), the ergosterol to microbial biomass C ratio (Djajakirana et al. 1996; Joergensen and Wichern 2008), the microbial biomass C to SOC ratio (Müller and Höper 2004; Heinze et al. 2010a, b; Murugan et al. 2013) and the metabolic quotient qCO2 (Müller and Höper 2004) as well as the fungal C to bacterial C ratio (Joergensen and Wichern 2008). An exception is to some extent the microbial residual C to SOC ratio, which reached on average 75 % and was considerably higher than the mean value of 55 % obtained by Appuhn et al. (2006). Solomon et al. (2001) obtained a mean of 32 % in Chromic Luvisols under different land-use systems and van Groenigen et al. (2010) only 22 % in Irish arable Haplic Luvisols, comparing MBT and GRT treatments. As the C/N ratio of most soils was very similar, methodological constraints of amino sugar hydrolysis or determination method cannot be fully excluded. Also, the conversion values proposed by Appuhn and Joergensen (2006) and Engelking et al. (2007) might be questionable for different soils. However, the question of the conditions under which microbial residues or, in reverse, plant residues are sequestered as SOM apparently remains unsolved.

The site effects on soil chemical and microbiological properties often exceeded those of the tillage treatments, with numerous significant interactions, which suggest that the tillage effects varied often in a site-specific way. The strong site effects of the current research are in agreement with Heinze et al. (2010a). However, in contrast to this experiment, the maximum site effects at the site Friemar can be explained by a combination of the following four site factors: (1) highest clay content, (2) lowest mean temperature due to the highest altitude, (3) lowest annual mean precipitation and (4) the highest soil pH in comparison to the other three sites. The highest sand content in combination with a high groundwater table at the site Zschortau apparently had only minor effects on the soil chemical and microbiological properties of the silt dominated sites. This is especially true in comparison with the site Lüttewitz, which had soil properties very similar to those of the site Zschortau. Consequently, it is not possible to determine the dominant factor on the basis of the current data pool. However, the absence of small-scale block effects for most of the properties analysed suggests that climatic effects might be more important, especially in combination with differences in management by the different farmers. A considerably larger number of sites would be necessary to assess the relative importance of one specific factor.

Conclusions

A reduction in tillage intensity by the GRT and NT treatments increased stocks of SOM and especially microbial biomass in comparison with the MBT treatment. Our results indicate significant tillage effects in loess-derived silt loams suitable for rain-fed sugar beet production, although strong site-specific differences occurred for most of the soil chemical and microbiological properties analysed. A higher C input is the main reason for the small but more positive effects of the GRT treatment on SOM and microbial biomass stocks in comparison with the NT treatment. In the top soil layer at 0–5 cm depth, the increased stocks of SOC are not caused by the preferential accumulation of fungal residues, although the ergosterol to microbial biomass C ratio, an index for the presence of saprotrophic fungi, was significantly increased in the GRT and NT treatments in comparison with the MBT treatment. A disproportionate increase in the turnover of fungal biomass may have led to a lower accumulation of fungal residues in comparison with bacterial residues. In contrast, the GRT and NT treatments promoted AMF at the expense of saprotrophic fungi in the bottom soil layer at 30–40 cm depth. The negative relationship between the ergosterol to microbial biomass C ratio and the fungal C to bacterial C ratio points to the importance of the relationship between saprotrophic fungi and biotrophic AMF for tillage-induced changes in microbial turnover of SOC.