Contribution of Sheep Grazing to Plant Diversity in Natural Grasslands

Abstract

Global climate change and overgrazing have led to the degradation of natural grasslands and seriously threaten the diversity of grassland plant species, as plant species richness is very sensitive to natural grassland degradation. Therefore, we conducted a sheep grazing experiment in Zinniquan pasture on the northern slope of Xinjiang Tianshan Mountains to test the effects of grazing on plant species diversity in natural grasslands through the spatial and temporal characteristics of the foraging behavior of grazing sheep and the plant species composition on the grazing trajectories. Data on sheep grazing tracks were collected based on GPS, vegetation composition around the tracks was investigated and seeds were collected from sheep manure to determine germination rates by simulating rumen fistula uptake tests in sheep. We found that sheep grazing can bring seeds from agricultural fields to grow on natural grasslands, causing changes in plant species diversity and community structure. The results of this study provide new insights into natural grassland species restoration and offer a range of new strategies that can be applied to the management of natural grassland plant ecological restoration.

1. Introduction

The Seirphidium semidesert grassland is an important spring and autumn pasture for sheep grazing in the Xinjiang Province in China. It is also an indispensable part of the connection between the mountains and oases. In addition, it is an important grassland type distributed in the semi-agricultural and semi-pastoral areas of Xinjiang [1]. Sheep use free grazing in spring and autumn. The Seirphidium semidesert provides forage for sheep operations, and the grazing livestock serve as seed dispersers for various plants in the autumn because of the maturity of the various plant seeds [2]. In addition to grazing in the Seirphidium semidesert, the sheep also graze in the farmland after the harvest of the crops near the grassland in the autumn. For many years, this grazing behavior has become a fixed habit in the agro-pastoral ecotone. This grazing behavior may from farmland to spread seeds the pasture through the sheep, causing them to germinate and form seedlings in grasslands, which in turn may change the species diversity, community structure, and even landscape characteristics of grasslands, ultimately causing important impacts on grasslands.

The dispersal of plant seeds by herbivores is primarily carried out by epizoochorous (the seeds are swallowed by the animal while feeding on the plant and are excreted in the digestive tract.) [3] and endozoochorous organisms (seeds are dispersed by adhering to animal hairs or other behaviors) [4]. The studies on epizoochory to date have primarily focused on the time of attachment and release of the seeds on animal fur, the size and morphological structure of the seeds, and the relationship between herbivore fur types and seed dispersal [5]. Since the “Foliage is the Fruit” hypothesis was proposed, endozoochory has caused wide concern [6,7,8]. Endozoochory is also considered to be one of the important ways in which long-distance seed dispersal takes place in addition to hydrochory (plants that mature and produce seeds will propagate by wind) and anemochory (seeds of river-grown or aquatic plants fall into the water and spread with the current) [4]. The large-scale spatial dispersal of the seeds through the animal enables the plant population to occupy new habitats, avoid fierce sibling competition, and greatly increase the landscape diversity on a spatial scale [9,10].

In recent years, several models have been derived to describe the long-distance seed dispersal of endozoochory, including classical diffusion and mixed dispersal models [11]. However, most of the models adopt empirical or hypothetical data, especially the lack of data on the actual moving trajectory and distances, retention time, and the defecation rule on the grazing trajectory surface. Therefore, this analysis produces a severe “fat tails” effect, which greatly limits the application of the model in the endozoochory of seeds [12]. The wide application of GPS makes it an excellent tool to document the location and trajectory of moving targets [13]. Therefore, this study combines the Seirphidium semidesert and the grazing sheep as the research objects. First, the real-time data of the sheep-grazing trajectory were acquired by GPS technology in the field. Real-time data defined the regularity of the sheep-grazing trajectory in the grazing area. Through the collection of the total feces from the sheep and germinative seed detection in the feces to determine the plant species and their amount of establishment on the grazing trajectory surface, we determined the germinability of the seeds after simulated ingestion by permanent sheep rumen fistula. The research aimed to answer these two questions: (1) what are the spatiotemporal characteristics of the sheep grazing trajectory? (2) What are the characteristics of the plant community structure on the grazing trajectory? We revealed the relationship between exotic plant seeds and sheep grazing behavior by answering these questions. This research can also provide theoretical support for grazing management and the restoration of this type of grassland that has been degraded.

2. Materials and Methods

2.1. Study Area

The research area (approximately 7.09 km²) is located in the Ziniquan sheep breeding farm on the north slope of the Tianshan Mountains (43°56′–44°03′ N, 85°40′–85°59′ E) (Figure 1) in Xinjiang Province, has a temperate continental climate. The mean annual temperature is 7.2 °C (−18.5 °C in January and 26.6 °C in July) and a mean annual precipitation of 231 mm, of which most precipitation falls from June to August. The community types are Seriphidium borotalense, Stipa spp., Carex spp., and Liliaceae species, with Seriphidium borotalense serving as the dominant species. The sheep used in the experiment included 170 Chinese Merino, and the grazing time was in the spring (from March to May) and the autumn (from September to November) every year.

2.2. Data Acquisition and Processing

2.2.1. GPS Trajectory Data Acquisition

A grazing distribution map was created from the tracking data recorded at 3-min intervals using GPS to evaluate the grazing trajectory’s spatial distribution during the study period. Five grazing sheep were randomly selected to wear GPS tracking collars to collect their tracking data. Spring grazing took place from April to June 2016 and autumn grazing from September to November 2016. The collected trajectory data was uploaded to the geographic information server in real time. The uploaded data primarily included the device ID worn by the sheep, location information, and send time. After acquiring the trajectory data, each track point is superimposed with the Landsat 8 image of grazing land by “creating point” operation in Mapinfo, and the grazing trajectories were formed by connecting the nodes through line tools. Based on the relationship between the length and time of the grazing trajectory, basic data such as the sheep retention time at different grazing positions were calculated.

2.2.2. Grassland Community Characterization Survey

Based on satellite imagery with track data, a survey of community diversity was conducted in the grazing area centered on the area with the most intensive grazing tracks, and then every 50 m, and the survey points were made to cover the track surface formed by sheep grazing. The survey was conducted in May–June 2017, and the vegetation in the non-grazed area was used as a control during the survey, mainly recording quantitative indicators such as plant species, density, and height. At the same time, the surface of the litter, sheep excrement and manure, field measurement of soil humidity, pH value, soil compactness and other indicators, the collection of soil samples, and soil physical and chemical characteristics of the analysis, with the above data as environmental factors indicators. Through the above work, we obtained five environmental indicators for each sample, including soil compaction, soil organic matter, pH value, amount of litter and amount of livestock manure, and formed a two-dimensional data table of sample × environmental elements.

The study area pasture is divided into multiple grids, and the cell grid size is set to 30 m × 30 m according to the actual demand; the trajectory points present in each cell are counted, and then the frequency distribution of the location points. The grazing intensity (GI) of the given time range (spring and autumn grazing time periods) is calculated. The calculation method is as follows.

$$D_{GI}=\frac{C\times N}{S}$$

(1)

where: D_GI is the grazing intensity in terms of the number of sheep units (SU) per square kilometer, SU·km⁻²; C is the total number of location points in the cell; N is the ratio of the total number of sheep in the herd to the total number of GPS devices, the number of sheep units represented by each GPS terminal, and S is the area of the cell in km².

2.2.3. Faeces Collection

Three sheep were randomly selected to have their excrement collected. The feces were collected every hour from the beginning of the sheep grazing on the grassland to their return to the pens. The following day before the start of grazing, the fecal bags were replaced to start the second cycle, which was collected for three consecutive days. The collection process was then repeated every seven days, and the collection was stopped at the end of the autumn grazing. After the collection of feces was completed, all the fecal samples were brought back to the laboratory. The glass bottles were crushed with manure and all seeds were picked out. Seeds were germinated using between paper method: seeds were placed on filter paper and put together in a Petri dish with a certain amount of water, germinated in an incubator, and their number was examined. Finally, the spatial and temporal patterns of seed discharge on the grazing track were calculated.

2.2.4. Seed Collection and Simulated Ingestion Experiment

Mature seeds were collected between August and October 2017 according to the areas with the most intensive grazing trajectories, and mature seeds were collected from more than 100 individual plants in the 15 species surveyed including seven species that were collected in the farm fields. To maintain their vigor, the seeds were taken to the laboratory, air dried, and then stored in brown paper envelopes at −4 °C to maintain their vigor.

A simulated ingestion experiment was conducted so that there were enough seeds to accurately determine the percentage of seed germination [14]. A permanent rumen fistula was created in the three sheep that were not used in the feeding experiment. Most seeds were retained: Heat-sealed nylon bags (11 cm × 7 cm, 40 µm pore size), containing 100 seeds of plant (single species) were introduced through the fistula, and five nylon bags were placed in the rumen of each sheep. Most seeds were retained in the rumen for 22–46 h and in an acidic part of the gut (abomasum and duodenum) for 2–4 h [15]. The bags were incubated for 22, 34, or 46 h inside the sheep rumen to simulate these conditions. After removing them from the rumen, each bag and its contents were rinsed with tap water and placed in a 0.1 N pepsin-hydrochloric acid solution for 2 h at 40 °C. The solution was produced by dissolving 2 g of pepsin (Merck reference 1.07190.1000 with activity 2000 FIP-U/g) in 1 L of 0.1 N HCl. The seeds from the simulated ingestion experiment were used to determine the germination percentage. The incubations were conducted in controlled environment chambers with 16 h light at 25 °C and 8 h of darkness at 15 °C [16].

2.3. Data Analysis

The quadrate data were sorted to form a two-dimensional matrix of “species × quadrat”. All data were preliminarily organized using Excel 2007 software, and community diversity characteristics, community similarity comparisons, and analysis of variance (ANOVA) were performed using data processing system (DPS 7.5) software. The Bray-Curtis similarity coefficients were calculated using the information on the number of species and the number of individuals of each species in each community. The Bray–Curtis similarity coefficients were calculated using the number of species and the number of individuals of each species. According to the species composition, the different community types and plant species distribution characteristics were divided using DCA (detrended correspondence analysis) of the CANCCO 4.5. CCA (canonical correspondence analysis) was used for antecedent selection to screen out the main environmental factors. To assess the overall plant diversity change to grazing, the richness, individual number, Simpson’s index, Shannon Wiener index, and evenness were determined. The Tukey test of a one-way ANOVA was used to verify significant differences in the diversity index, quantitative characteristics of the displacement trajectories, germinated plant seedlings, and the germination behavior of the seeds of the exotic plant species.

$$Simpson\mathrm{Index}(D):D=1-{\displaystyle \sum _{i=1}^S{p}_i{}^2}$$

$$Shannon–Weiner\mathrm{Index}(H):H^{\prime }={\displaystyle \sum _{i=1}^S{p}_{i}Log}{P}_i$$

In the equation: P_i is the proportion of species i in the community.

3. Results

3.1. Sheep Grazing Area and Moving Trajectory in the Spring and Autumn

Flock distribution across the study area was clearly uneven. In general, the intensity and range of the grazing activities in the autumn were significantly larger than those in the spring. The mean sheep grazing intensity by markers in different colors across all 90 sampling days indicated a preference for the northeast corner of the study area (Figure 2A). The sheep were fed from the northwest to the southeast in the spring, which runs through the primary mountain range (a relatively flat area) of the whole pasture. In the autumn, sheep were observed to be dispersed across the study area, they moved along the southeast direction for a period of time and then moved towards the southeast, the central, and the northern parts, respectively. In addition, sheep also grazed in the farmland near the grassland.

The statistical calculation of the quantitative characteristics of the trajectory data in the grazing process shows that the sheep tended to be inactive in the spring pastures (

Table 1. Quantitative characteristics of the displacement trajectories of the grazing sheep in the spring and autumn.

Project	Spring (T = 5 d, N = 5 Sheep)	Autumn (T = 5 d, N = 5 Sheep)
Average trajectory length (km)	6.39 ± 0.442 b	7.58 ± 0.071 a
Average velocity (km/h)	0.61	0.63
Trajectory length of the grassland Area (km)	4.59 ± 0.421 b	6.31 ± 0.052 a
Track length in the non-grassy areas (km)	2.80 ± 0.552 a	1.27 ± 0.015 b
Average length of residence in the grass area (h)	6.02 ± 0.821 b	7.13 ± 0.235 a
Stay time in the non-grassland areas (h)	2.03 ± 0.552 a	1.02 ± 0.102 b
Duration of the stay in the farmland (h)	0.00 b	2.21 ± 0.423 a
Trajectory length in the farmland (km)	0.00 b	1.06 ± 0.058 a

Note: Different lowercase letters in the same line indicate significant difference (p < 0.05).

). The average length of the grazing trajectory, the length of the trajectory in the grassland area, and the length of the trajectory in the non-grassland area are significantly different from those in the autumn pastures (p < 0.05). In addition, affected by the anthropic factor, the flock is prohibited from grazing near spring farmland. In the autumn, the sheep can feed freely in the farmland where the crops were harvested, so that the time of stay and the length of the trajectory in the farmland are all higher than those in the spring pasture (p < 0.05).

3.2. Quantitative Characteristics of the Species and Communities in the Grazing Trajectory Surface

Community Types and Species Distribution in Grazing Trajectory Surface

The DCA ordination was used to divide 45 quadrats into four communities: (I) The Polygonum aviculare + Chenopodium glaucum + Carex liparocarpos community, including 22 species (quadrats 1 to 11). (II) The Carex liparocarpos + Arenaria serpyllifolia community, including 23 species (quadrats 12 to 22 and 36). (III) The Carex liparocarpos + Artemisia annua community, including 18 species (quadrats 23 to 31 and 37 to 46). (IV) The Ceratocephalus orthoceras + Carex liparocarpos + Arenaria serpyllifolia community, including 15 species (quadrats 32 to 35) (Figure 3).

The first axis of the DCA ordination reflects the change in altitude, which gradually increases from left to right. As the elevation increases, the growing area gradually changes from the valley floor or flat land to the mountain top, resulting in large differences in the species and numbers of each community. The second axis reflects the change in soil type and stability. The soil type changes and the soil stability gradually decreased from top to bottom.

Further, DCA sorting analysis of all species showed that the seven exotic species were concentrated in relatively flat terrain (Figure 4), and the DCA of 45 sample plots (Figure 3) corresponded to the DCA of 25 species, and the distribution pattern of the dominant species of each cluster was basically similar to the distribution pattern of the sample plots. For example, community (I) was distributed at the leftmost end of the DCA, and the dominant species, such as the Polygonum aviculare and the Chenopodium album, were also distributed at the leftmost end of the DCA. The distribution pattern of dominant species further reflects the environmental characteristics of the sample site, and the results reflected by the two sorting maps are consistent.

Comparing the species diversity characteristics of the four communities, the results show that the richness, abundance, Simpson’s Index, Shannon–Wiener Index, and evenness in the four communities were significantly different (p < 0.05) (

Table 2. Comparison of species diversity characteristics of four communities.

Community	Richness	Abundance	Simpson’s Index	Shannon–Wiener Index	Evenness
I	16.27 ± 1.103 a	135.27 ± 30.454 b	0.83 ± 0.049 a	3.07 ± 0.243 a	0.76 ± 0.056 a
II	12.67 ± 2.964 b	115.25 ± 83.263 bc	0.74 ± 0.034 a	2.475 ± 0.270 b	0.68 ± 0.059 b
III	8.84 ± 2.340 c	70.00 ± 14.921 c	0.49 ± 0.112 b	1.66 ± 0.398 c	0.53 ± 0.088 c
IV	12.25 ± 0.957 b	302.75 ± 71.400 a	0.60 ± 0.106 b	1.85 ± 0.331 c	0.51 ± 0.078 c

Note: Different lowercase letters in the same column indicate a significant difference (p < 0.05).

). Compared with the other communities, community III’s diversity and ecological advantages were lower. The characteristic of the species diversity of each community corresponded to the species-quadrat DCA sort chart. Except for the abundance, the remaining index significantly decreased as the altitude increased (p < 0.05).

3.3. Environmental Interpretation of Community Formation in Grazed Grassland

Using the Monte Carlo test, marginal and conditional influences on community characteristics were derived. The influence of environmental factors on the characteristics of plant community composition was the marginal influence, and the characteristic values obtained by the forward selection, pre-selection, and linear substitution to remove the relevant variables were the conditional influences. The top-ranking according to the marginal influence of environmental factors was soil organic matter, followed by exposure, soil compactness, humidity, amount of waste, and litter (

Table 3. Influence of environmental factors on the characteristics of plant community composition.

Marginal Effect		Conditional Effect
Environmental Elements	Eigen Value	Environmental Elements	Eigen Value	p-Value	R-Value	F-Value
Soil organic matter	0.48	Soil organic matter	0.48	0.002 **	0.23	8.46
Exposure	0.46	Amount of waste	0.31	0.002 **	0.15	5.89
Soil compactness	0.37	Grazing intensity	0.34	0.002 **	0.16	6.75
Humidity	0.34	Litter	0.29	0.002 **	0.14	5.26
Amount of waste	0.32	Altitude	0.23	0.002 **	0.12	4.41
Litter	0.31	Exposure	0.19	0.002 **	0.09	1.34
Altitude	0.25	Soil compactness	0.06	0.190		1.46
Grazing intensity	0.23	Humidity	0.06	0.134		1.69
pH-Value	0.22	pH-Value	0.07	0.070		1.69

Note: ** Indicates the marginal and conditional effect of environmental elements extremely significant difference (p < 0.01).

). However, after forward selection, it was found that soil organic matter content was still the most important influence on species composition, but the influence values of environmental factors such as the amount of waste, grazing intensity, and litter were significantly higher.

3.4. The Source of Exotic Species within the Community

Statistical analysis showed a significant difference in the number of seeds germinated in manure after spring and autumn grazing (p < 0.05) (

Table 4. Species and quantity of the germinated plant seedlings collected in sheep feces (plant/kg feces).

Plant Species	Autumn	Spring
Chenopodium album	54.67 ± 8.96 a	0 b
Setaria viridis	2.67 ± 2.081 a	1.60 ± 0.58 b
Semen Trigonellae	9.67 ± 2.08 a	1.21 ± 0.12 b
Echinochloacrus galli	3.00 ± 1.00 a	0 b
Polygonum aviculare	2.33 ± 0.58 a	0 b
Medicago lupulina	6.78 ± 1.21 a	0 b
Convolvulus arvensis	1.23 ± 0.59 a	0 b
Amaranthus retroflexus	1.67 ± 1.15 a	0 b
Bassia dasyphylla	5.67 ± 1.53 a	0 b
Carex liparocarpos	1.67 ± 0.58 a	0 b
Artemisia annua	2.32 ± 0.58 a	0 b
Festuca ovina	3.58 ± 0.54 a	0 b
Salsola collina	3.65 ± 0.28 a	0 b
Dysphania botrys	6.33 ± 2.081 a	0.67 ± 0.58 b
Galium verum	2.33 ± 0.58 a	0 b

Note: Different lowercase letters in the same line indicate a significant difference (p < 0.05).

). After feeding in the autumn, there were 15 species of germinating plants in the feces. Chenopodium album germinated the most frequently (54.67), while Convolvulus arvensis germinated the least frequently (1.23). After feeding in the spring, only three species of plants germinated in the feces, such as Setaria viridis, Trigonella arcuata and Dysphania botrys. The germination number of Setaria viridi was the highest (1.60), while that of Dysphania botrys was the lowest (0.67).

The germination test showed that before the rumen digestion of the sheep (

Table 5. Germination behavior of exotic plant seeds after rumen digestion in sheep.

Plant Species	0 h	8 h	24 h	36 h	48 h
Chenopodium album	40.67 ± 2.31 d	80 ± 3.06 ab	80.67 ± 5.03 ab	76 ± 3.11 b	64 ± 3.02 c
Setaria viridis	76.67 ± 3.06 b	92.67 ± 4.00 a	87.33 ± 1.15 a	72.67 ± 2 b	71.33 ± 4.00 b
Convolvulus arvensis	39.21 ± 2..50 d	58.65 ± 3.24 c	89.24 ± 10.26 a	90.21 ± 9.58 a	78.65 ± 14.23 b
Eragrostis pilosa	8.25 ± 1.21 d	25.26 ± 1.34 c	77.01 ± 11.25 a	45.10 ± 10.21 b	25.14 ± 8.53 c
Polygonum aviculare	89.00 ± 12.21 a	56.32 ± 8.45 a	13.50 ± 3.24 c	9.32 ± 3.14 c	8.52 ± 2.45 c
Medicago lupulina	18.21 ± 2.35 d	31.24 ± 4.58 c	56.24 ± 8.25 b	75.85 ± 6.89 a	65.32 ± 8.24 b
Trigonella arcuata	5.20 ± 0.28 c	15.32 ± 1.69 b	18.25 ± 1.25 b	20.32 ± 2.69 a	25.26 ± 2.34 a

Note: Different lowercase letters in the same line indicate a significant difference (p < 0.05).

), the highest seed germination rate was exhibited by Polygonum aviculare (89.00%), while the lowest was that of Trigonella arcuata (5.20%). The germination rates of the Eragrostis pilosa (8.25%) and Trigonella arcuata (5.20%) seeds were lower, and there were obvious signs of dormancy. The germination rate increased with the increase in the time of digestion of the rumen. Among the seven types of seeds, the germination rate of Polygonum aviculare gradually decreased as the digestive time increased. The germination rate of Trigonella arcuata gradually increased as the digestive time increased. The germination rate of the remaining five species of seeds increased first and then decreased with the increase in the digestion time.

4. Discussion

4.1. Effects of the Different Seasons on the Trajectory of the Grazing Sheep

Quantitative characteristics of the displacement trajectories of grazing sheep showed that the intensity of activity, feeding time, and feeding range in the autumn were larger than those in spring (Table 1). In the autumn, the sheep are highly active and graze more intensively. They move to the central and northern slopes after feeding in the flat areas to meet their own growth and development needs. Free-grazing cattle tend to graze in areas with less slope and rock. When the grazing intensity tends to be saturated or the grassland biomass is low, they will graze in areas with steeper terrain [17]. Studies have shown that with the seed setting and withering of the pasture in autumn, the sheep will frequently choose feeding sites to eat enough pasture, and the sheep will also prolong their feeding time [18]. The feeding range of the herds is primarily related to three factors, including the vegetation condition, the terrain condition, and the anthropic factor [19]. In recent years, many studies have used spatial analysis techniques to explore the relationship between grazing livestock and grasslands [20,21,22,23,24]. An analysis of the distribution of the grazing sheep track points indicated that there was a significant difference between spring and autumn. In this study area, due to the single vegetation species caused by long-term grazing, the impact of the condition of the vegetation can be neglected. The terrain condition and the anthropic factor are the primary factors that affect the difference in the grazing trajectories in the spring and autumn. The sheep primarily feed in the southeast direction of the relatively flat terrain in the spring. In the autumn, in addition to feeding toward the southeast direction, the central, northern, and farmland areas are also grazing areas (Figure 2).

4.2. Quantitative Characteristics of Species and Communities in the Grazing Trajectory Surface

Species diversity, as a basic feature of community [25], indicates the structure and composition of the community [26] and reflects the characteristics of the different aspects of the plant community [10,27]. Species diversity differences are formed under the combined influence of environmental [28,29] and biological [30,31] factors. Compared with the vegetation in the non-grazing area, it is found that the possible cause of exotic species is sheep grazing. The DCA ordination of all the species indicates that the six exotic species are concentrated in relatively flat areas which are the areas where grazing sheep eat frequently. The DCA ordination diagram of 25 species (Figure 3) corresponds to the DCA ordination diagram of 45 quadrats (Figure 4), and the pattern of distribution of the dominant species of each cluster is basically similar to the distribution pattern of the quadrats. For example, community I was distributed at the left end of the DCA ordination diagram, and its dominant species, such as Polygonum aviculare and Convolvulus arvensis, were also distributed at the left end of the DCA ordination diagram.

4.3. The Source of Exotic Species within the Community

4.3.1. Species and Quantity Characteristics of Germinating Plant Seedlings in the Sheep Feces

Spring and autumn serve as natural grazing times to provide essential forage for livestock. By counting the species and number of plants that germinated in sheep’s feces, it was found that after foraging in spring pastures, three species of plants sprouted in sheep’s feces, namely Setaria viridis, Trigonella arcuata, and Dysphania botrys. However, after foraging in autumn pastures, up to 15 species of plants germinated in feces. This may be due to the fact that in the spring, when the climate becomes warmer and forage grasses start to germinate and grow, sheep feed on the leaves, stems, and a few roots of various plants, so no new plant seeds enter the digestive tract of sheep, and the seeds that sprout in feces may come from the residues of plants that died in the previous fall and winter or from seeds that are dormant. Sheep are particularly fond of eating field weeds and crop seeds after the fall farm harvest, and most seeds are excreted 24–48 h after feeding in the farm area with a clear excretion peak [16]. According to the strong mobility of sheep, sheep move from farmland to pasture to graze, the seeds eaten by sheep in the farmland area are excreted into the grazing path following the feces, and the undamaged seeds as well as the seeds that are released from dormancy sprout and grow in the environment. This further suggests that the foraging behavior of sheep in the farmland area can bring field plant seeds into the grazing area and that seeds are excreted with sheep feces, a process that provides the potential for seed dispersal. In addition, exotic seeds, namely field Chenopodium album, Setaria viridis, Convolvulus arvensis, Eragrostis pilosa, Polygonum aviculare, Medicago lupulina, and Trigonella arcuata., were detected in seeds that sprouted within sheep manure after grazing in fall pastures, all of which have some forage value and can contribute to natural grassland plant species diversity.

4.3.2. Effects of the Digestive Tract on the Seed Germination Characteristics of the Exotic Plant Species

The successful germination of the plant seeds after the action of the digestive tract is the primary prerequisite for endozoochory [32]. In this study, the seed germination rates of seven exotic plants were determined before and after rumen digestion in the sheep. With the prolongation of the digestion time, the germination rates all tended to increase first and then decrease with the exception of Polygonum aviculare and Trigonella arcuate; this may be related to the characteristics of the seeds. Previous studies revealed that Chenopodium album, Setaria viridis, Convolvulus arvensis, Eragrostis pilosa, Medicago lupulina, and Trigonella arcuata seeds have obvious needs for dormancy [33,34], and some plant seeds cannot imbibe and germinate because the seed coat is impervious to water [2,35]. After the seeds pass through the digestive tract, the dormancy of such seeds is broken to varying degrees [36]. After a long period of rumen digestion, the germination rate of Medicago lupulina, Setaria viridis, Eragrostis pilosa, Convolvulus arvensis, and Chenopodium album seeds did not increase, it decreased. The reason could be that the seeds remained in the digestive tract, and after prolonged exposure to rumen fluid, the seed coat and the embryo were seriously damaged, resulting in a lower germination rate, and possibly causing the inactivation of the seed. Studies have shown that the germination rate of soft-shelled or non-dormant seeds decreases when they are immersed in rumen fluid [8]. In this study, the germination rate of the Polygonum aviculare seed decreased significantly after traversing the digestive tract. The rate probably decreased because the seed coat of Polygonum aviculare was thin and soft. In the digestive tract, its seed coats do not provide effective protection for the seeds, resulting in the inactivation of the seed by the rumen fluid.

The successful germination of the plant seeds after the action of the digestive tract is the primary prerequisite for endozoochory, while the key to the success of endozoochory is that the seeds are placed in natural habitats, and then the seedlings grow. This process represents the effectiveness of endozoochory [37]. Some studies have shown that the excretion rule of sheep conforms to the Gaussian model. A small number of seeds begin to be excreted at 6 h after feeding [1,16], and the peak of excretion is concentrated at 24–48 h. With the exception of Polygonum aviculare, the germination rate of the other seeds at the peak of excretion is high, and this provides a possibility for the secondary dispersal of seeds. In this study, seven species of plants are all exotic species, which still have some rate of germination after passing through the digestive tract. It shows that the seven species of exotic plants can still germinate and grow in the environment through digestive tract excretion after feeding by sheep. It can increase the diversity of the Seirphidium semidesert grassland community structure and the heterogeneity of vegetation landscape. This is also a process that the plant adapts to in the changing environment.

5. Conclusions

In this study, the investigation of sheep grazing trails revealed significant differences between spring and autumn sheep grazing trajectories, which may be related to vegetation and topographic conditions. Therefore, we used the DCA method to classify the grasslands in the study area into four community types and found that the diversity indices of each community differed significantly. The results indicate that herbivores can effectively spread plant seeds over long distances, and their seeds still have different degrees of germination rates after transmission through the sheep endozoochoryt, which provides the possibility of establishing exotic species and thus enriching the diversity of grassland vegetation and improving the utilization of grazed grasslands.