Sampling was performed in August in 2012-2013 along 100-m transects within every sampling site that each contained eight evenly distributed 50 cm × 50 cm quadrats. The plant community was divided into five functional groups: Gramineae, Cyperaceae, forbs, Leguminosae, and shrubs. In each quadrat, we recorded the coverage and dry aboveground biomass of plant functional groups, the thickness of the mattic epipedon, and the slope aspect, position, and angle. We also noted the number of livestock grazing at each sampling site (Lin, 2017).
1.2 Sample collection and preparation
The plant samples were placed in an envelope after collection and subsequently baked at 60–80°C for 48 to 72 h to a constant weight, which was then recorded.
An unstructured questionnaire was used to determine the area of the pasture, the type and quantity of grazing livestock, and the grazing time. According to the conversion ratio of 4 sheep equal to 1 yak, the stocking capacity of the pasture was calculated (Xu, 1994), and the stocking capacity was divided into 5 grades. The average grazing intensities per grazing season were 0.00 sheep/ha for more than 1 year, less than 3.65 sheep/ha, 3.65 to 7.50 sheep/ha, 8.25 to 13.00 sheep/ha, and more than 13.00 sheep/ha. In most of the time when the alpine meadow degradation into Black-soil beach meadow would become to no grazing in the pasture, so we investigated the possible grazing intensity used in the past as the average grazing intensity of the pasture.
1.3 Data analysis
The plant functional group characteristics of mini-patches were determined by calculating the average coverage, dry aboveground biomass, and importance value of each plant functional group in each mini-patch (Lin, 2017). The data collected from every mini-patch were also used to calculate the niche breadths of plant functional groups based on the Levins index (Lin, 2017), and mean and standard error on the biomass aboveground, the importance and the niche breadth were calculated by performing a one-way analysis of variance in SPSS Statistics Software (19.0). The probability of each mini-patch reaching a certain multi-steady stage was calculated as the ratio of the environmental characteristics in a single quadrat to the total in all quadrats. The coupling between the plant functional groups of all mini-patch and each environmental factor was calculated via a redundancy analysis (RDA) in R3.6.1 software.
2 Results