Study Site
The study site of this experiment was located in the Xilin River Basin, a temperate steppe (116°14′E, 43°13′N) near the Inner Mongolia Grassland Ecosystem Research Station (IMGERS) in the Inner Mongolia Autonomous Region, China. The experimental area is about 100 ha natural grassland with an elevation up to1250 m, which has been fenced since 1999 to exclude large animal grazing. Based on a long-term observational dataset (1980-2013), the mean annual temperature here was 0.9 °C, with mean monthly temperatures ranging from −21.4 °C in January to 19.7 °C in July, and the mean annual precipitation was about 350.00 mm, approximately 70%-80% of which fell between May and August. The soil at the site is either a Haplic Calcisol or Calcic-Orthic Aridisol, as classified by the FAO and the US soil classification system, respectively. The dominant plant species in the community areStipa grandis and Leymus chinensis , which together account for more than 60% of the total peak aboveground biomass, the other 40% of vegetation are forbs.
Experimental design
Eighty experimental plots were established, with a randomized complete block design in September 2008 (Y. Zhang et al., 2016). Each plot was measured 10 m × 10 m, and there were 2-m walkways between adjacent plots. Each plot was treated with one of five different levels of N addition, i.e., 0, 2, 10, 20 and 50 g N m-2 yr-1 (designated as N-0, N-2, N-10, N-20, and N-50, respectively), at one of two N addition frequencies (twice a year vs. monthly), and with one of two management regimes (un-mowing vs. mowing). Hence, there were 20 treatments (5 × 2 × 2) in total, each with 4 replicates. The form of N added to the plots was purified NH4NO3(> 99%). N additions started on 1 November and were continued on the first day of June of the same year and November of the next year for the low frequency treatments (2 N addition yr-1), or started on 1 November and continued on the first day of each month for the high frequency treatment (monthly, or 12 N addition yr-1). Fertilizer was applied referring to the method of Zhang et al. (2016). Mowing treatments were carried out two times a year, with 10 cm stubble height left after mowing, on 25 June to 30 June and 25 August to 30 August in every year from 2009 to 2012.
Soil and plant sampling and measurements
Plant and soil samples were collected in mid-August 2012. To investigate vegetation biomass and collect plant and soil samples, a 2 m × 0.5 m quadrat was randomly placed in each plot, but at least 50 cm inside the border of each plot to avoid edge effects. The aboveground biomass (AGB; g m-2) was measured by clipping all plants above the soil surface, oven-drying the combined clippings at 65 °C for 48 h, recording the sample’s dry weight, and then finely crushing the sample in a mill to determine the C and N content of plants (Cplant and Nplant; mg kg-1). Three root cores were collected using a 7 cm diameter root auger at 10 cm intervals in the plots to a soil depth of 30 cm. These core samples were then mixed and soaked in water to remove the mineral components of the soil, then treated as described above to measure the belowground biomass (BGB; g m-2). Soil samples were collected with a soil augur (3 cm in diameter), and in each plot three cores were taken with 0-10 cm depth and at least 50 cm apart. Soil samples were then mixed and sieved through 2-mm mesh, stored at 4 °C for laboratory analysis of soil ammonium (NH4+-N; mg kg-1), nitrate (NO3--N; mg kg-1), and moisture (SM %) content, as well as microbial biomass carbon and N (MBC and MBN; mg kg-1). Soil subsamples were air-dried for analysis of soil pH, and soil C and N (Csoil and Nsoil; mg kg-1). The Cplant, Croot, and Csoil were determined by a H2SO4-K2Cr2O7oxidation method (X. T. Lü, Lü, Zhou, Han, & Han, 2012). The Nplant, Nroot, and Nsoilwere determined by using the Kjeldahl acid-digestion method with an Alpkem auto-analyzer (Kjektec System 1026 distilling unit, Sweden). Soil NH4+-N and NO3--N were analysed using a flow injection auto analyzer (FLAstar 5000 Analyzer; Foss Tecator, Hillerød, Denmark), in which these substances were extracted from 10 g of fresh soil sample with 50 ml of 0.5 M K2SO4solution. MBC and MBN were analysed with a fumigation extraction method (Vance, Brookes, & Jenkinson, 1987). Soil pH was determined by measuring pH of a soil-water suspension (soil: water = 1: 2.5) with a pH meter (FE20-FiveEasy).
Gross ammonification and nitrification rates
The 15N pool dilution technique was used to determine the gross ammonification rate (GA; mg kg-1 SDW d-1) and gross nitrification rate (GN; mg kg-1 SDW d-1) in intact soil cores with one from each experimental plot. Soil in the cores were labelled with15N-(NH4)2SO4or 15N-KNO3 (30 atom%15N enrichment and 3 mL 100 g-1 dry soil equivalent) for quantification of gross N ammonification or nitrification. A custom-made multi-injector, which consisted of 10 simultaneously operated syringes with custom-made side-port cannulas, was used to ensure homogenous labelling. Each injection was equivalent to 0.1 mL, 100 g-1 soil, and then was labelled with 3 mL of 15N-enriched solution that corresponded to 2 mg N kg-1 dry soil; 3 replicate labels were made from each core. When soil was sampled, the fresh soil core was labelled immediately with the apparatus above mentioned (3 times at 0.3 cm, 1.5 cm, and 2.5 cm depth). About one hours later (t1), the first soil core was collected and extracted by 0.5 M K2SO4 solution with a soil: solution ratio of 1:2 as described by Dannenmann et al. (2009). Forty hours after labelling (t2), the second soil core was extracted using the same procedure as at t1.
Soil extracts were frozen for further processing. The diffusion technique described by Dannenmann et al. (2006) was used to collect NH4+and NO3- on acid filter traps, which were then analysed for 15N enrichment with an isotope ratio mass spectrometer (Delta Plus XP; Thermo) at the Center of Stable Isotopes of KIT-IMK-IFU. Subsamples of the soil extracts were analysed to determine their NH4+ and NO3- concentrations by a commercial laboratory in order to avoid bias. Gross ammonification and nitrification rates were calculated with the equation provided by Kirkham & Bartholomew (Kirkham & Bartholomew, 1954)(1954).
Statistical analyses
All results are presented in figures as mean ± SE (standard error). Repeated measures analysis of variance (ANOVA) for a randomized complete block design was performed, and Tukey’s HSD post hoc test was used to test for differences in the response variables between different treatments if the ANOVA revealed an overall significant difference. IfP < 0.05, differences were concluded to be statistically significant. Linear regression was also performed to examine the relationships of GA and GN rates with soil pH and NH4+-N + NO3--N. All statistical tests were performed with SAS 9.0, except for regression analyses, which were performed in the SIGMAPLOT software (SIGMAPLOT 12.5 for windows; Systat Software Inc., San Jose, CA, USA) that was also used to produce graphs. Structural equation modelling (SEM) was carried out with the software package IBM SPSS Amos 21.0.