Introduction

Human activities have altered global nutrient cycles and caused nutrient enrichment, especially nitrogen (N) and phosphorus (P), in many ecosystems (Mahowald et al. 2008; Vitousek et al. 1997). Nutrient enrichment can influence composition of biota and processes of the ecosystem. Matson et al. (2002) reviewed the consequences of N deposition for terrestrial ecosystems, and they suggested that not all ecosystems respond to N deposition similarly. In addition, P fertilization can lead to eutrophication of water by P runoff and leaching (Schoumans and Groenendijk 2000), but few studies have examined the ecological effects of P addition on terrestrial ecosystems (Bennett and Adams 2001; Stöcklin et al. 1998). Moreover, the effects of simultaneous N and P additions on terrestrial ecosystems remain uncertain (Elser et al. 2007) because of the limited number of studies. Soil quality is an important indicator in ecosystem management and sustainability (Masto et al. 2008), and thus, changes in soil quality due to nutrient additions need to be evaluated (Malý et al. 2009), so as to develop effective strategies for the management and sustainability of ecosystems under nutrient additions.

Soil microorganisms play a key role in sustaining soil quality, and soil microbial properties have been proposed as sensitive indicators of changes in soil quality (Filip 2002; Jenkinson 1988; Nannipieri et al. 2003). Soil microbial biomass serves as a source and sink of plant-available nutrients and is closely related to soil quality (Kaschuk et al. 2010). The microbial quotient (qMIC), the percentage of microbial biomass C (MBC) to soil organic C (SOC), can provide an effective indication of the improvement or deterioration of soil quality (Wardle 1992). The microbial metabolic quotient (qCO2), which is the amount of CO2-C produced per unit MBC, has been used as an eco-physiological measure of ecosystem succession or disturbance (Wardle and Ghani 1995). Soil microbial biomass, qMIC, and qCO2 have been used as indicators of soil development or degradation and changes in soil quality (Insam et al. 1989; Kaschuk et al. 2010). Potential N mineralization and nitrification have been frequently used to evaluate the N supply by soil (Robertson et al. 1999) and as indicators of ecosystem susceptibility to degradation (Aber et al. 1989).

Grasslands that are characterized by sandy soils constitute a large percentage of natural vegetation in arid and semi-arid regions of northern China. Located at the ecotone of agro-pastoral systems, the Keerqin Sandy Lands are one of the biggest sandy lands in China and are characterized by nutrient-poor soils. Rapid economic development in China has increased nutrient addition to soil throughout the country (Li et al. 2004; Lü and Tian 2007). To date, however, few data on the effects of nutrient addition on soil quality are available for this sandy grassland (Zeng et al. 2010). In this study, we have evaluated changes in soil microbial and chemical properties after 5 years of N and/or P additions in a sandy grassland in Keerqin Sandy Lands.

Materials and methods

Site and treatments

The experimental site was located at Daqinggou Ecological Station (42°58′ N, 122°21′ E, 260 m above sea level) of the Institute of Applied Ecology, Chinese Academy of Sciences, in the southeastern Keerqin Sandy Lands, Northeast China. The mean annual temperature was 6.4°C, the lowest mean monthly temperature (January) was −12.5°C, and the highest mean monthly temperature (July) was 23.8°C. The mean annual precipitation was 450 mm, with more than 60% occurring in June–August and a high potential evaporation of 1,300–1,800 mm per year. The plant community at the experimental site was dominated by Pennisetum flaccidum, Chenopodium acuminatum, Cleistogenes chinensis, Artemisia scoparia, and Lespedeza davurica. The soil type is Typic Ustipsamment, characterized by coarse texture and loose structure, with a texture of 90.9% sand, 5.0% silt, and 4.1% clay. The surface (0–15 cm) soil was slightly acidic (pH = 6.6) and the soil bulk density was ∼1.5 g cm−3.

The experimental site was a flat sandy grassland and was fenced to exclude livestock grazing. A factorial N × P addition experiment in a randomized block design with six replicates was established in 2004 and was based on four treatments: no addition (control), N addition (N), P addition (P), and combined addition of N and P (N + P). Nitrogen was added at the rate of 20 g N m–2 year−1 as urea in 2004–2006 and as NH4NO3 in 2007–2008. Phosphorus was added at the rate of 4.4 g P m−2 year−1 in the form of NaH2PO4 in 2004–2008. Nitrogen + P were added with the same forms and rates as N and P treatments. The fertilizer rate was similar to that normally used by farmers of the region. Each plot (4 × 4 m) was separated by a 2-m buffer strip. The fertilizers were dissolved in 16 L water and applied in early May (30%) and mid-June (70%) each year. Control plots only received 16 L of water without fertilizers.

Soil sampling

Soil samples were collected in April and July of 2009. After removal of the litter layer, eight soil cores (2.5 cm in diameter) were randomly collected from 0 to 15 cm layer in each plot and were mixed to form a composite sample. After removing visible roots, fauna, and organic debris by hand, the soil samples were sieved (<2 mm) and divided into two subsamples. One subsample was air-dried at room temperature (around 20°C) and used for chemical analyses. A portion of these samples was ground with a mill and sieved (<0.25 mm) prior to SOC, total N (TN), and total P (TP) analysis, whereas the other part was used for soil pH measurement. The other subsample was stored in plastic bags at 4°C for 24 h prior to analysis of NH +4 –N and NO −3 –N concentrations, net N mineralization and nitrification rates, basal respiration (BR), MBC, and microbial biomass N (MBN). The samples collected in July 2009 were only analyzed for SOC concentration.

Soil chemical properties

SOC concentration was determined by the K2Cr2O7–H2SO4 wet oxidation method of Walkley and Black after removal of carbonates by acid pretreatment (Nelson and Sommers 1996). Total N and TP concentrations were determined with a continuous flow autoanalyzer (AutoAnalyzer III, Bran + Luebbe GmbH, Germany) after the samples were digested (340°C) with H2SO4 using a mixture of K2SO4 and CuSO4 as catalyst. Soil pH on a 1:2.5 (w/v) mixture of soil and water was measured using a PHS-3C pH meter (Shanghai Lida Instrument Factory, China).

Soil N mineralization potential was measured by a laboratory incubation method described by Menyailo et al. (2003). Briefly, fresh soil (equivalent to 20 g oven-dry soil) was extracted with 2 M KCl solution (1:5 soil/solution ratio) to determine NH +4 –N and NO −3 –N concentrations with the continuous flow autoanalyzer. Additionally, another fresh soil (20 g oven-dry soil) was placed in a plastic flask, moistened to 60% of water-holding capacity, and aerobically incubated at 25°C for 15 days. To avoid anaerobic conditions, the flask was opened every 3 days for 5 min. At the end of the incubation, the soil sample was analyzed for the final inorganic N concentration as described above. Soil net N mineralization rate (per day) was determined by the difference in inorganic N concentrations before and after the incubation. Likewise, net nitrification rate was calculated as the difference in NO −3 –N before and after the incubation.

Soil microbial properties

Soil MBC and MBN were determined by the chloroform fumigation–extraction method (Cabrera and Beare 1993; Vance et al. 1987). Briefly, soil (25 g) was pre-incubated in a humidified, darkened, 25°C incubator for 7 days. Soils with or without chloroform fumigation were extracted with 50 mL 0.5 M K2SO4 for 30 min and then filtered. The organic C in the soil extract was measured by dichromate oxidation (Vance et al. 1987) and TN in the extract by K2S2O8 oxidation method (Cabrera and Beare 1993). Microbial biomass was calculated as the differences in K2SO4-extractable C or N concentration between fumigated and unfumigated soils, divided by efficiency factors for MBC (K C = 0.38; Vance et al. 1987) and MBN (K N = 0.45; Jenkinson 1988), respectively. The unfumigated K2SO4-extractable organic C value was also considered as the soil dissolved organic C (DOC) concentration (Zeglin et al. 2007). The qMIC was calculated as a percentage of MBC to total organic C.

Basal respiration was determined by measuring CO2 evolution (Hu et al. 2006). Briefly, field-moist soil (20 g oven-dry soil) was placed in 500-mL air-tight glass vessels and water was added to reach 60% of the water-holding capacity. The vessels were closed with rubber stopper and pre-incubated at 25°C for 3 days. Then, they were opened and closed again for a 24-h incubation period at the same temperature. The CO2 evolved from the soil was absorbed in 10 mL 0.2 M NaOH, and the remaining base was titrated with 0.1 M HCl after the formed Na2CO3 was precipitated by adding 2 mL 1 M BaCl2. The results were expressed as microgram CO2-C per gram per hour. The qCO2 was calculated by dividing the hourly BR by the corresponding MBC.

Statistical analysis

Changes in soil chemical properties (SOC, TN, TP, and pH) were analyzed by the two-way ANOVA using the General Linear Models procedure to test the main effects of N addition, P addition, and their interactions. Repeated measures ANOVAs were used to examine N addition and P addition effects on NH +4 –N and NO −3 –N concentrations, net N mineralization and nitrification rates, and MBC, MBN, BR, DOC, qMIC, and qCO2 values at the two samplings. Between-subject effects were evaluated as N addition or P addition treatment and within-subject effects were time-of-season. All analyses were performed by the SPSS (version 13.0) and the accepted significance level was α  = 0.05.

Results

Nitrogen addition, P addition, and their interactions did not significantly affect SOC and TN concentrations (Table 1). The P addition significantly (P = 0.002) affected TP concentration, whereas N addition and its interactive effects with P addition were not significant on TP concentration (Table 1). Nitrogen addition markedly (P < 0.001) affected soil pH, while no significant interactive effect of N and P additions on soil pH was detected.

Table 1 Concentrations of organic carbon (SOC), total nitrogen (TN), and total phosphorus (TP) and pH of surface (0–15 cm) soil under four treatments: control, nitrogen (N), phosphorus (P), and both nitrogen and phosphorus (N + P) additions
Full size table

The effects of N addition, P addition, and their interactions were not statistically significant for net N mineralization rate, whereas N addition significantly enhanced net nitrification rate (P = 0.017; Fig. 1a, b). No interactions were observed between sampling season and N and/or P addition on net N mineralization and nitrification rates (Fig. 1a, b).

Fig. 1
figure 1

Changes in a net N mineralization and b nitrification rates, c microbial biomass carbon (MBC) and d nitrogen (MBN), e basal respiration (BR), f extractable dissolved organic C (DOC), g microbial quotient (qMIC), and h microbial metabolic quotient (qCO 2 ) in surface (0–15 cm) soil in the control, nitrogen (N), phosphorus (P), and both nitrogen and phosphorus (N + P)-treated plots. Bars represent standard errors (n = 6). Above the bars, the superscript letters “N” and “S” indicate a significant effect of N fertilization or sampling season, respectively, and “S × N” indicates an interactive effect, followed by the level of statistical significance (based on repeated measures ANOVA, with N and P addition treatments as between-subject effects and sampling season as within-subject effect)

Full size image

The ANOVA analysis showed that N addition markedly decreased MBC (P = 0.005; Fig. 1c), and thus, qMIC (P = 0.004; Fig. 1g), but MBC and qMIC were not significantly affected by P addition (P = 0.858 and P = 0.587, respectively). Moreover, the interaction of P with N addition did not significantly affect MBC and qMIC (P = 0.326 and P = 0.820, respectively; Fig. 1c, g). There was no effect of sampling season on MBC and qMIC (Fig. 1c, g). Nitrogen addition significantly decreased MBN (P = 0.004) and its interaction with sampling season affected MBN (P = 0.016; Fig. 1d). No significant effects of N addition (P = 0.958), P addition (P = 0.121), and their interactions (P = 0.269) on BR were detected (Fig. 1e). Therefore, N addition significantly increased qCO2 (P = 0.001) due to the significant decrease of MBC with N addition, but P addition (P = 0.082) and its interaction with N addition (P = 0.061) did not significantly affect qCO2 (Fig. 1h). In addition, N addition significantly (P = 0.010) increased extractable DOC (Fig. 1f).

Discussion

Our results show that N addition decreased MBC and MBN (Fig. 1c, d), which confirms what was observed in temperate grassland soils (Sarathchandra et al. 2001; Treseder 2008). Sarathchandra et al. (2001) suggested that the decline in MBC by N addition was related to the reduction in SOC, which is the energy source of soil microorganisms. In our study, however, SOC was not affected by N addition over 5 years (Table 1) even though C input was significantly increased due to the increased aboveground biomass (Zeng et al. 2010). The cause of decreased MBC is probably due to the soil acidification by nitrification caused by N addition (Table 1); indeed, accumulation of NO −3 –N was observed (Fig. 1b). Aciego Pietri and Brookes (2009) suggested that soil pH is an important factor controlling biomass, activity, and community structure of soil microorganisms. A recent study also showed marked influence of pH on growth and biomass of soil microorganisms and composition of decomposers (Rousk et al. 2009). Low pH can decrease microbial activities and growth of soil microorganisms (Rousk et al. 2009) and consequently reduce microbial biomass under N addition (Fig. 1c, d).

The lack of effects of N addition on soil respiration with the decline in microbial biomass (Fig. 1c–e) does not agree with what was reported by Treseder (2008). The decrease in microbial biomass was responsible for the increase in qCO2 and the decline in qMIC values (Fig. 1g, h). Probably, despite the significant increase in extractable DOC under N addition (Fig. 1f), there was a decline in the efficiency of converting available C into microbial biomass.

In our investigation, P addition had negligible effects on soil net N mineralization and nitrification rates, MBC, MBN, BR, and extractable DOC, and thus on qMIC and qCO2 values, although it significantly increased soil TP concentration (Table 1). Moreover, P addition had no significant interactions with N addition on these soil properties, probably due to the negligible effects of P addition on plant N uptake associated with negligible effects on plant growth (unpublished data).

In conclusion, N addition over 5 years did not affect SOC, TN, and TP concentrations in semi-arid, sandy grassland. Nitrogen addition decreased soil pH which probably reduced microbial biomass, and thus declined qMIC and increased qCO2 values, while it also increased net nitrification rate and extractable DOC concentration. However, P addition had no effects on soil microbial properties. These results suggest that in the mid-term the high input of N may decrease microbial biomass due to induced pH effects, with a possible decline in soil quality.