The experimental study focuses on Yunwu Mountain (Mt.) Grassland in the Loess Plateau of China (between 106°21' and 106°27' eastern longitude and 36°10' and 36°17' northern latitude, Fig. 1). The elevation of the grassland varies from 1 800 to 2 200 m above sea level. The area of Yunwu Mt. Grassland is 6 660 ha which is composed of three areas: core area (1 700 ha), buffer area (1 400 ha) and external transition area (3 560 ha). The core area of the grassland has been enclosed and ungrazed for over 30 years (since 1982) which results in large amount of litter lying or standing on soil surface. The Yunwu Mt. Grassland has been promoted to the state-level nature reserve in China since 2013.
Climate at Yunwu Mt. Grassland is temperate semi-arid with a mean annual precipitation of 424 mm (mean values from 1980 to 2009). Precipitation shows a marked seasonality with 65%–75% rainfall occurring in the vegetation period. The drought index varies from 1.5 to 2.0. Mean annual temperature is 4–6 ℃ with the warmest month averaging 22–25 ℃. The annual cumulative temperature above 10 ℃ is 2 100–3 200 ℃. The mean annual day-length is approximate 2 500 h. Meteorological data, including air temperature, relative humidity, rainfall, sunshine duration were measured by an automatic weather station. Table 1 shows monthly meteorological data at the growing season in 2013 and 2014.
Year Meteorological data May Jun. Jul. Aug. Sept. Oct. 2013 Rainfall (mm) 61.9 78.7 164.3 109.1 96.9 25.5 Tmean (℃) 15.3 18.5 19.0 19.6 13.8 9.2 Mean wind speed (m/s) 2.6 2.7 2.2 2.3 2.0 1.8 Sunshine duration (h) 7.6 7.5 5.4 9.2 5.6 7.1 2014 Rainfall (mm) 8.4 94.6 78.3 93.0 150.0 46.9 Tmean (℃) 14.7 18.4 20.1 17.9 14.2 9.4 Mean wind speed (m/s) 2.7 2.2 2.1 2.3 2.2 1.8 Sunshine duration (h) 9.8 7.8 8.7 8.0 5.2 6.2
Table 1. Monthly meteorological data in the growing season of Yunwu Mountain Grassland in 2013 and 2014, respectively
Soil is classified as a Cal-Orthic Aridisol according to Chinese Soil Classification and Terminology (Qiu et al., 2012). Soil organic carbon content at a depth of 0–100 cm is around 20.5 kg·m-2 in the grassland (Chang et al., 2017). Texture analysis of root zone soil (0–60 cm) yields a sand content of 31.8%, a silt content of 56.5% and a clay content of 11.7%. Soil is classified as silt loam according to the U.S. textural classification triangle. The average saturated water content and field capacity of the top 100 cm are 0.52±0.02 and 0.43±0.02 cm3·cm-3, respectively. Detailed soil physical properties are shown in Table 2.
Soil depths (cm) Soil particle fraction (%) Soil texture Bulk density (g·cm-3) Saturated wetness (cm3·cm-3) Clay (< 0.002 mm) Silt (0.05–0.002 mm) Sand (> 0.05 mm) 0–10 8.3 49.6 42.1 Loam 1.18 0.52 10–20 11.7 59.5 28.8 Silt loam 1.27 0.48 20–40 10.2 59.1 30.7 Silt loam 1.25 0.5 40–60 9.1 43.9 46.9 Loam 1.25 0.5
Table 2. Physical properties of root zone soil in the litter manipulation site
The preponderant vegetation species is Stipa bungeana. The main growing period is from June to September, which is general consistent with precipitation. Plant roots were mainly concentrated at a depth of 0–30 cm.
Litter manipulations were conducted to assess the effects of litter on soil properties and plant growth. Three litter treatments were carried out including removal of all standing and lying litter, an untreated in-situ control with original litter levels, and a double litter treatment where the removed litter from the first treatment was added to the plot. The added litter was evenly spread over the plot to provide double litter treatments similar to the treatment conducted by Deutsch et al. (2010) and Villalobos-Vega et al. (2011). The plot size of each treatment is 114 m2 (6 m×19 m), and each manipulation has three repetitions.
Soil water in the root zone (0–60 cm) was measured weekly (May–September) in 2013 and 2014 using a neutron probe and a TRIME moisture probe. The neutron and TRIME pipes were set in the center of each plot at 5, 10 and 15 m location along the length. Consecutive five-day soil water content was measured following a relative heavy rain in August to investigate the effects of litter on depletion of soil water. Soil in root zone was sampled monthly to measure total carbon and nitrogen. Soil at depth of 0–100 cm was sampled using an auger, and placed in plastic bags and transported to the laboratory for detailed analysis. Total carbon and nitrogen content in soil were examined by the element analyzer (Elementar vario MACRO cube).
Plant characteristics were measured on the dominant specie of Stipa bungeana. The heights of ten plants (twenty tillers per plant) from the soil surface to the top of the longest leaf were measured in each plot during growing periods in 2013 and 2014. Aboveground biomass was sampled in the size of 0.25 m2 (0.5 m×0.5 m). The sampled plants were first dried at 105 ℃ for one hour to de-enzyme and then dried at 60 ℃ to constant weight. Root biomass was sampled using a root-core with volume 785 cm3 (10 cm diameter by 10 cm depth) to a depth of 50 cm with interval of 10 cm. The roots were separated from the soil by washing over a sieve of 0.5 mm mesh and were dried at 60 ℃ to constant weight.
Data were statistically analyzed by a one-way analysis of variance (ANOVA) using the software of SPSS 21.0 software (SPSS Inc., Chicago, IL, USA). Statistical differences between treatments were determined by the Duncan test with significance at P=0.05.
1.1. Study Site Description
1.2. Litter Manipulation Experiments
Soil moisture was affected by the manipulation of plant litter in the semi-arid grassland. Table 3 shows the average soil water content at a depth of 0–10 cm under different litter manipulations. Monthly average soil water content in the superficial layer varied in the range of 0.20–0.35 cm3·cm-3 during the growing periods. In the earlier (May) and late (September) growing period, soil water content in plots without litter was significantly (P < 0.05) lower than that in plots in the presence of litter. But there was no significant difference (P > 0.05) for the plots of in-situ and double litter manipulations. In contrast, litter manipulation has no significant (P > 0.05) effect on monthly average soil water content in high growing periods in both experimental years.
Year Month Removal In-situ Double 2013 Jun. 0.22±0.03 a 0.28±0.05 b 0.29±0.02 b Jul. 0.30±0.05 a 0.31±0.04 a 0.32±0.04 a Aug. 0.26±0.08 a 0.27±0.07 a 0.29±0.06 a Sept. 0.20±0.04 a 0.18±0.05 b 0.19±0.04 b 2014 May 0.23±0.07 a 0.26±0.04 b 0.29±0.05 b Jun. 0.22±0.07 a 0.25±0.06 b 0.26±0.05 b Jul. 0.23±0.03 a 0.30±0.13 b 0.29±0.06 b Aug. 0.24±0.09 a 0.24±0.07 a 0.24±0.08 a Sept. 0.26±0.08 a 0.27±0.07 ab 0.29±0.06 b
Table 3. Mean monthly soil water content at a depth of 0–10 cm for the different litter manipulations during growing season (values within columns followed by different letters are statistically significant at the 0.05 level)
Figure 2 illustrates the depletion patterns of soil water up to 5 days post rainfall under different litter manipulations. The depletion pattern of soil water varied with soil depth. Soil water content in plots of litter removal subsequent to a rain was evidently higher than that in plots in the presence of litter for the upper root zone, which suggested that litter has an impact on rainfall interception. The fastest soil water depletion took place in the plots of litter removal which demonstrated the barrier effects of litter on soil evaporation. In contrast, there was no significant difference (P > 0.05) of soil water content in the deeper root zone (20–40 cm), so that the depletion patterns were almost identical for the different litter treatments.
Figure 2. Depletion of soil water up to 5 days post rainfall for different litter manipulations (a) 0–10 cm, (b) 10–20 cm, (c) 20–30 cm, and (d) 30–40 cm.
Soil carbons, nitrogen as well as their ratio at a depth of 0–50 cm are presented in Fig. 3 for different litter treatments. Soil total carbon monotonically decreases from surface downward to deeper root zone at different growing periods. The total carbon of soil at a depth of 0–10 cm was around 40 g·kg-1, and varied around 26 g·kg-1 in deep root zone (40–60 cm). The variation of soil carbon in different layers was not significant (P > 0.05) for different plant litter treatments.
Figure 3. Mean (±SD) soil total carbon and nitrogen distribution profiles as well as carbon-to-nitrogen ratio for different litter manipulations (a) 2013-6, (b) 2013-8 and (c) 2013-9.
The distribution of soil nitrogen in root zone was similar to that of soil carbon (Fig. 3). Total nitrogen was around 3.0 g·kg-1 for different growing periods at a depth of 0–10 cm and gradually reduced to around 1.7 g·kg-1 in deep root zone (40–60 cm). Variations of total soil nitrogen in different depth were not significant (P > 0.05) for litter treatment as well. Two-year consecutive litter manipulation did not show evidently change of the total soil nitrogen.
The carbon-to-nitrogen ratio shows the degradation rate of organic matter. The carbon-to-nitrogen ratio was approximately 13 for superficial soil at different growing periods. In the vertical profile, the ratio of carbon to nitrogen slightly increased with soil depth.
Plant height at different growing periods for different litter treatments is presented in Fig. 4. In general, the dominant species of Stipa bungeana in plots with litter removed was significantly shorter (P < 0.05) than those in plots of in-situ and double litter treatments. In contrast, the height of plants in plots with in-situ and double litter treatments varied in different growing periods. In the first experimental year, there was no significant difference (P > 0.05) between the double and in-situ litter treatments. In mid and high growing periods (June to August) of the second year, plants in double litter plots were significantly (P < 0.05) shorter than those in in-situ plots. However, in earlier and late of growing periods, there was no significant difference for plant in double and in-situ litter plots.
Figure 5 shows the dry biomass of tillers in plots with different litter treatments. The effect of plant litter on tiller weight varied with different growing periods. In the late of first and earlier of second experimental year, addition and removal of litter did not significantly influence (P > 0.05) the tiller weight. However, the tiller weight was significantly (P < 0.05) influenced by litter in the high growing periods of second experimental year. The presence of litter significantly (P < 0.05) promoted the tiller weight.
Root weight for different litter treatments is presented in Fig. 6. More than 70% of root weight concentrated in the upper root zone (0–20 cm) for all litter treatments. Root at a depth of 0–10 cm account for approximate or more than 50% of total root mass. The effects of litter manipulation on root weight varied with growing period. In the first experimental year, litter manipulation did not show significant difference for the root mass and distribution in the vertical profile (Fig. 6a). In the early growing period of second experimental year, litter addition or removal increased root biomass. However, the highest root biomass obtained in plots with in-situ litter level in the high growing periods. The large standard deviation of root weight may be attributed to spatial heterogeneity of grass growth in the plots.