A Study on the Impact of Roads on Grassland Degradation in Shangri-La City

Abstract

Shangri-La is located in the eastern part of the Qinghai-Tibet Plateau, which has a fragile ecology. The plateau grassland has suffered from irreversible degradation under the influence of human activities. To address this issue, the Sentinel-2A data obtained is used in this study to calculate the RVI and build an inversion model of grassland degradation grade with GDI data, which was used to obtain the area and proportion of grassland degradation. Landscape indexes were then calculated for different degradation grades of grassland to examine the correlation between roads and degraded grassland in spatial distance and the spatial distribution characteristics of different degradation grades of grassland. The results show that the grassland area in Shangri-La was 2207.94 km², of which the heavily degraded area reaches 824.03 km², exceeding the undegraded grassland area by 172.62 km², indicating that the grassland degradation is severe. From south to north, the proportion of heavily degraded and moderately degraded grassland in townships gradually decreased, while the proportion of lightly degraded and undegraded grassland gradually increased. The townships with high percentages of degraded grassland were predominantly located in the southern area, where there was a dense road network and well-developed transport networks, particularly along National Highway 214, which is the main road in Shangri-La. Conversely, townships with low percentages are generally located in the north with dispersed roads and sparse transport lines. The study’s outcomes are significant in providing a better understanding of the current status of grassland degradation and promoting the sustainable utilization of grassland resources in Shangri-La.

1. Introduction

Grasslands are one of the most widely distributed vegetation types and an essential part of terrestrial ecosystems in the world, playing an important role in global climate change with significant ecological, economic, and social value [1], and are essential for conserving biodiversity and maintaining ecosystem balance. However, according to statistics, about 20% of grassland biomass is declining globally and shows varying degrees of degradation [2], with Asia having the largest area of degraded grassland at 37 million km², accounting for 22% of the total grassland area [3]. In China, grassland resources are significant national resources [4]. Currently, China has over 400 million hectares of grassland, more than 2/3 of which have been degraded to different degrees due to the unreasonable utilization of them by human beings and the invasion of alien species [5]. The degradation of grasslands has seriously restricted the development of animal husbandry, directly affected the production and life of grass-dependent grazing peoples, and threatened the quality of the whole ecological environment and ecological security [6]. Grassland degradation has become a severe ecological problem facing the world today.

As a large province with rich grassland resources in southwest China, the area of natural grassland in Yunnan Province accounts for about 40% of its total land area [7]. Still, the excessive exploitation and utilization of grassland have led to the degradation of a large amount of natural grassland [8]. Heavy degradation in a part of the grassland has emerged [9], among which the Shangri-La region in northwest Yunnan is typical [10]. This region is one of the vital biodiversity conservation areas in Yunnan Province and even nationwide, with a relatively large area of natural grasslands. Moreover, animal husbandry is the dominant industry of local Tibetan residents. Since the northwest Yunnan region is located in the middle of the Hengduan Mountains, the natural conditions are harsh, and the grassland resources are fragile, which is a fragile and sensitive area of the ecosystem. Therefore, a comprehensive understanding of the regions’ status of grassland degradation has become a more urgent need.

With the socio-economic development of Shangri-La and the boom in local tourism, overgrazing, trampling, and other human-related factors have brought about increasingly severe damage to the grassland, which has made the fragile grassland ecological environment suffer more severe threats than before. Since anthropogenic activities are mainly carried out along traffic routes, this paper extracts the distribution data of degraded grassland in Shangri-La with the help of satellite remote sensing data. It analyzes the relationship between road distribution and grassland degradation, which will help to fully reveal the spatial and temporal characteristics of grassland degradation and the mechanism of grassland degradation in the region, and provide scientific reference and practical support for the local formulation of degraded grassland management methods.

2. Materials and Methods

2.1. Overview of the Study Area

Shangri-La, a county-level city in Diqing Tibetan Autonomous Prefecture, Yunnan Province, is situated in the northwestern Yunnan Province, with a geographical location of 26°52′~28°52′ N and 99°13′~100°19′ E, at the southeastern edge of the Qinghai-Tibet Plateau (Figure 1). Shangri-La, located in the heartland of the Three Parallel Rivers of Yunnan Protected Area, has a cold–temperate mountainous monsoon climate and intense solar radiation, with the terrain descending from north to south, its temperatures between day and night varying tremendously [11]. Shangri-La is a Tibetan-inhabited area and supports animal husbandry, one of its pillar industries, with rich grassland resources [12]. With the rocketing progress in recent years, Shangri-La City has formed a transportation network pattern of East–Central–West, three arterial roads through the city [13], of which the Shangri-La Central National Highway 214 (G214), crossing the entire territory of Shangri-La, is the most critical transportation route in the region. The eastern route comprises the East Ring Road and the Xiangwei Line. The Jinjiang Line and other county and township highways dominate the western route. These roads above together connect the townships of Shangri-La City. Convenient transportation has driven the development of the local economy but also impacts grassland degradation.

2.2. Data Introduction

In this study, the 2017 Sentinel-2A series remote sensing images, obtained from Google Earth Engine (https://developers.google.cn/earth-engine/ (accessed on 4 March 2021)), are used as the data source. Land use data are obtained from the 2017 remote sensing monitoring data of the current land use situation in China published by GlobelLand30 (http://www.globallandcover.com/ (accessed on 5 March 2021)) with a spatial resolution of 30 m × 30 m. According to the national land classification standard, the land use of the study area was classified into forest, grassland, arable land, water bodies, artificial surface, etc. The road data are obtained from the national road vector data of the Geographic State Monitoring Cloud Platform (www.dsac.cn (accessed on 6 March 2021)).

2.3. Overall Research Method

In this study, the degradation level and spatial distribution pattern of grassland in Shangri-La are analyzed by pre-processing Sentinel-2A satellite remote sensing data, constructing a degraded grassland monitoring model, and classifying grassland degradation classes based on a survey on the impact of Shangri-La roads on grassland degradation. Then the study area is divided into grids. The landscape indexes are calculated within the grids for different levels of grassland degradation, including the Class Area (CA) and Patch Density (PD). Class area (CA) represents the actual size of the plaque. The higher the CA index is, the larger the area of this type of plaque is, while the lower the index is, the smaller the area of this type of plaque is. Patch Density (PD) represents the number of patches per unit area, and the higher the value of PD, the greater the number of patches per unit area, and vice versa. Then the relationship between landscape indexes of different degraded grades of grassland and road distance is analyzed. Finally, the pattern of the influence of road distribution on the degradation of grassland in Shangri-La city is analyzed based on the results.

2.4. Remote Sensing Monitoring of Grassland Degradation

Remote sensing monitoring of grassland degradation requires the determination of grassland degradation grades and the construction of a model for monitoring degraded grassland. The Grassland Degradation Index (GDI) is a comprehensive evaluation index of grassland degradation and can classify the grassland degradation level. The detailed calculation of GDI is as follows [9]:

$$\mathrm{GDI}=\sum _{i=1}^n{v}_i{w}_i-v_d{w}_d$$

(1)

In this formula, GDI is the grassland degradation evaluation index, n is the number of variables, v_i is the normalized value of the indicator characterizing grassland degradation, w_i is the weight of the degradation representing the variable, v_d is the normalized value of the toxic weed percentage, and w_d is the weight of the toxic weed percentage indicator.

A higher GDI value indicates good growth of grassland, no degradation, or low degradation of grassland; a lower GDI value indicates poor grassland growth and heavy grassland degradation. According to informed studies, GDI can be divided into four classes [9], which are: undegraded grassland (GDI > 0.5162), lightly degraded grassland (0.3757 < GDI < 0.5162), moderately degraded grassland (0.2560 < GDI < 0.3757), and heavily degraded grassland (GDI < 0.2560).

Establishing a remote sensing inversion model of grassland degradation is to invert the degraded grassland in the study area quantitatively. We selected seven vegetation indexes and GDI which are commonly used for correlation analysis, namely the Normalized Difference Vegetation Index (NDVI), Renormalized Difference Vegetation Index (RDVI), Difference Vegetation Index (DVI), Soil-Adjusted Vegetation Index (SAVI), Perpendicular Vegetation Index (PVI), Enhanced Vegetation Index (EVI), and Ratio Vegetation Index (RVI). Among them, RVI has the highest correlation with the Grassland Degradation Index (GDI), and the RVI increases with the increase of GDI, showing a positive correlation. Therefore, we used the RVI for grassland degradation inversion.

Based on the close relationship between the vegetation index and the surface morphology of grassland vegetation, the vegetation index of Sentinel-2A was used as the data source, which is applicable to vegetation monitoring when the vegetation is actively growing and has high-coverage RVI after pre-processing. The ratio vegetation index (RVI), also known as greenness, is the ratio of the near-infrared channel to the red channel [14]; the Sentinel-2A satellite band ratio is the near-infrared band (b8) to red band (b4). The ratio vegetation index (RVI) can be expressed as the following equation:

$$\mathrm{RVI}=\frac{{\rho }_{\mathit{nir}}}{{\rho }_{\mathit{red}}}$$

(2)

The remote sensing monitoring model of grassland degradation, i.e., RVI and GDI linear regression [9], can be expressed as the following equation:

$$\mathrm{GDI}=0.076\times {\mathrm{RVI}}^{1.15}$$

(3)

After extracting the distribution of grassland in Shangri-La, the corresponding RVI classification nodes were deduced at the same time, and then each degradation level of grassland was classified.

To test the accuracy of the fitted values of the constructed GDIrs regression model, 30 grassland quadrats were selected as the experimental group (Figure 2). The Relative Error between the GDIrs obtained by RVI fitting and the real GDIg was 1.61%, and the RMSE value was 0.736.

2.5. Landscape Pattern Analysis Based on Grid

Based on the landscape indexes calculated by the grid, this paper analyzes the spatial distribution pattern of grassland degradation in Shangri-La city. As the largest county-level city in Yunnan Province, the size of the grassland grid is determined to be 500 m × 500 m by considering the size of the city and the resolution of the spatial distribution data of each degradation type. The landscape indexes (Class Area (CA) and Patch Density (PD)) for different degradation levels of grassland in the grid are then calculated in Fragstats4.2 software. Finally, the distances from grassland grids of different degradation grades to roads are calculated in ArcGIS software.

2.6. Quantile Regression

Quantile regression is to regress the explanatory variables at different quartiles between 0 and 1 on the explanatory variables, establish the corresponding quantile equations, and obtain the effect of the explanatory variables on the overall conditional distribution. Quantile regression can effectively control the effect of outliers and is suitable for estimating non-normally distributed data, making a higher robustness of the regression coefficient estimates. The implementation method is to load the package “quantreg” in R language for quantile regression.

In this study, the patch landscape index of each degradation type as explanatory variables is non-normally distributed, so quantile regression is suitable for estimation. Taking the calculated landscape index as the explained variable of the analysis and the Euclidean distance between the degraded grassland patch and the road (EDPR) as the explanatory variable, the impact of roads is analyzed under different quantiles of the landscape indexes. EDPR refers to the Euclidean distance between the grass patch and the road. However, because there are many roads, there are multiple EDPRs for a particular grass patch. Therefore, EDPR refers to the Euclidean distance between the grass and the nearest road (Figure 3). The positive regression coefficient indicates that the EDPR has a positive impact on the landscape index change, and the negative regression coefficient has a negative impact. Based on the regression results of EDPR and multiple landscape indexes, the impact of roads on grassland degradation is analyzed.

3. Results

3.1. Analysis of the Degree of Grassland Degradation

The grassland data of different degradation grades in the whole city of Shangri-La in 2017 is obtained through calculation. From the spatial distribution map of degraded grasslands of different degradation grades (Figure 4), it can be seen that the grassland area of Shangri-La has been heavily degraded. The total grassland area of Shangri-La is 2207.94 km², among which the heavily degraded area reaches 824.03 km², exceeding 172.62 km² more than the undegraded grassland. Although the moderately and lightly degraded areas in Shangri-La are relatively small, the total degraded area reaches 1556.54 km², accounting for 70.5% of the city’s total grassland area (

Table 1. Grassland area of each degradation degree (unit: km²).

No.	Administrative Divisions	Heavily Degraded	Moderately Degraded	Lightly Degraded	Total Degraded	Undegraded	Total
1	Tiger Leaping Gorge	106.46	22.79	13.74	142.99	24.09	167.08
2	Shangjiang	17.30	4.53	2.75	24.58	4.15	28.74
3	Xiao Zhongdian	93.67	26.02	17.45	137.14	40.85	177.99
4	Jiantang	142.11	44.52	31.91	218.54	62.86	281.40
5	Jinjiang	33.30	17.50	8.89	59.69	8.65	68.35
6	Luoji	79.60	38.40	24.62	142.62	31.57	174.19
7	Wujing	16.04	6.66	4.65	27.35	8.28	35.63
8	Sanba Naxi	117.69	64.03	47.68	229.4	45.38	274.78
9	Gezan	142.33	85.43	81.36	309.12	219.75	528.87
10	Nixi	28.62	25.33	30.38	84.33	54.42	138.75
11	Dongwang	46.91	52.67	81.17	180.75	151.41	332.16
12	Shangri-La	824.03	387.90	344.61	1556.54	651.41	2207.94

). In addition, the degraded grassland area in Shangri-La is mainly focused on the central and southern parts and relatively less degraded areas in the northern part. Viewed from the east to west, the eastern and central grasslands are heavily degraded, while the western grasslands are less degraded, and most of the degraded grasslands are concentrated on both sides of the highway (Figure 4).

Because the size of each township is different, divided the grassland area of different degradation levels in each township is divided by the total area to calculate the area ratio for better comparison (Figure 5b). In this manner, the result can be figured that the townships with a high proportion of heavily degraded grasslands are mainly distributed in the southern part of the study area, with the proportion of heavily degraded grasslands in townships gradually decreasing from south to north.

Except for Shangjiang which ranks in the top four in the proportion of heavily degraded grassland area, the other three townships (Tiger Leaping Gorge, Xiaozhongdian, and Jiantang) are all areas where Shangri-La’s main traffic artery—the middle line passes. Among the three townships, Tiger Leaping Gorge, located in the southmost region of Shangri-La, has the highest proportion of heavily degraded grassland area. Then the Xiaozhongdian and Jiantang decrease successively. In contrast to the distribution pattern of heavily degraded grasslands, townships with a relatively high proportion of undegraded grasslands are mainly distributed in the northern part of the study area, and the proportion of lightly degraded grasslands gradually decreases from north to south (Figure 5b). The change in the degraded grassland area in each township also shows a similar pattern. Gezan, Nixi and Dongwang located in the northern part of Shangri-La are townships where the area of undegraded grassland is larger than that of heavily degraded grassland. Contrary to these three townships, the area of heavily degraded grassland in the rest of the townships is larger than that of undegraded grassland (Figure 5a).

The top four townships with a proportion of moderately degraded grassland are Jinjiang, Sanba Naxi, Luoji, and Wujing. Compared with these four townships, the remaining townships in the south and north have a lower proportion of moderately degraded grassland. These four townships and the Shangjiang, which has a relatively high proportion of heavy degradation, are mainly distributed in the eastern region of Shangri-La (including the Sanba Naxi and Luoji) and the south-central part of the western region (including the Jinjiang, Wujing, and Shangjiang). The Jinjiang Line is the main highway that crosses the west, but with a generally small grassland area in the west (Figure 5a). The East Ring Road mainly passes through two towns in the east, with the area of moderately degraded grassland being close.

The top four in the proportion of lightly degraded grassland are the Dongwang, Nixi, Sanba Naxi, and Gezan. These four townships are mainly distributed in the northern part of the study area, and only one township is located in the central region. In other townships, the proportion of lightly degraded grassland gradually decreased from north to south (Figure 5b). The changing trend of the proportion of lightly degraded grassland area in townships is similar to that of undegraded grassland and opposite to that of heavily degraded grassland. More precisely, the proportion of undegraded grassland area gradually decreases, the proportion of heavily degraded grassland area gradually increases, and the proportion of lightly degraded grassland area gradually decreases from north to south.

The top four townships with a proportion of undegraded grassland are Dongwang, Gezan, Nixi, and Wujing. Except for the Wujing, which is located in the middle of Shangri-La, the other three townships are located in the north of Shangri-La. Among these four townships, Dongwang and Gezan, the two townships with the largest undegraded grassland area in Shangri-La, are only crossed by Provincial Highway X219 and a few county highways (Figure 5a). Meanwhile, the proportion of undegraded grassland in townships gradually decreases from north to south and is associated with an increase in the proportion of heavily degraded grassland (Figure 5b).

Combined with the spatial distribution of townships and roads in the study area, the area along G214, which spans the central part of the study area, is the main area where grasslands are heavily degraded; the areas along the Jinjiang Line and Xiangwei Line in the southwest of the study area and the East Ring Line in the east of the study area are the main areas where grasslands are moderately degraded; lightly degraded and undegraded grasslands are mainly distributed in the northern and central areas of the study area where roads are scarce. Therefore, the distribution of roads has led to varying degrees of grassland degradation: the grassland degradation is the most serious in the area where the main traffic artery (G214) of Shangri-La passes through; the grasslands in the areas crossed by the roads in the east and west of Shangri-La are the second most degraded; the area traversed by the road in the northern region of Shangri-La has the lowest degree of grassland degradation.

As a result, a conclusion may be reached that from south to north townships, the proportion of heavily degraded and moderately degraded grassland decreased gradually, while the proportion of lightly degraded and undegraded grassland increased gradually. Townships with a higher proportion of degraded grassland are generally located in the south region of the study area with dense roads and well-established transport networks, while towns with a lower proportion are generally located in the north region of the study area with dispersed roads and sparse traffic lines.

3.2. Analysis of Main Factors of Grassland Degradation

3.2.1. Analysis of Possible Influencing Factors of Grassland Degradation

Shangri-La is located on the southeast of the Qinghai-Tibet Plateau. The alpine grassland distributed here is a natural barrier to maintaining regional ecological security. The short-term climate change is not the main factor leading to grassland degradation [15]. On the Qinghai-Tibet Plateau, the height of the grassland community decreases significantly with the increasing altitude [16]. The northern regions of Shangri-La at a relatively high altitude have a large area of non-degraded and slightly degraded grasslands. The degree of degradation of alpine grasslands in the northern regions of Shangri-La is less severe than in the southern regions. Therefore, altitude is not the most critical factor influencing grassland degradation in Shangri-La.

From 2012 to 2015, the total population of Shangri-La remained at around 176,000 (175,700 in 2013 and 176,600 in 2015), and the total population in 2016 and 2017 reached 178,000 and 179,400, respectively. The total population remained stable during the period [17,18,19,20,21,22]. Moreover, Shangri-La, covering a total area of 11,613 km², is the largest city-level administrative region in Yunnan Province. Therefore, Shangri-La presents the characteristics of “a vast territory with a sparse population”. Among the total population of Shangri-La, most of the population is distributed in urban areas, and only 18.4% of the total population is engaged in animal husbandry [17,18,19,20,21,22]. At the same time, the grasslands in Shangri-La are widely distributed. Therefore, the population has little impact on the grassland degradation in Shangri-La.

The overall degree of grassland degradation in Dongwang and Jiantang is quite different, but they are in the same grade of degraded grassland, the proportion of grassland area of each soil type is similar, and the grassland area of different degradation degrees does not change significantly with different soil types (

Table 2. Grassland proportion of each degradation grade of soil type.

Administrative Divisions	No.	Soil Class	Heavily Degraded	Moderately Degraded	Lightly Degraded	Undegraded	Total
Dongwang	1	Alpine Desert Soil	11.43%	10.15%	18.66%	56.40%	100%
	2	Alpine Meadow Soil	22.32%	18.59%	17.70%	41.39%	100%
	3	Dark Brown Soil	14.40%	17.55%	32.42%	35.63%	100%
	4	Brown Soil	10.14%	15.22%	32.44%	42.21%	100%
Jiantang	1	Artificial Grass Soil	47.66%	15.54%	11.17%	25.63%	100%
	2	Alpine Meadow Soil	34.01%	17.57%	11.82%	36.60%	100%
	3	Dark Brown Soil	54.69%	15.71%	11.85%	17.75%	100%
	4	Brown Soil	49.11%	16.26%	12.47%	22.16%	100%
	5	Bleached Podzolic Soil	56.69%	15.80%	8.79%	18.71%	100%

). Therefore, according to the area and proportion of grasslands with different degradation grades in different soil types, the soil type is not the main factor leading to grassland degradation in Shangri-La.

The main land-use types in Shangri-La from 2013 to 2017 were forest, grassland, cultivated land, water area, and construction land. During this period, the coverage area of grassland decreased from 287,058.23 hm² to 28,6851.31 hm², with the area of change being only 206.92 hm² [23]. According to the change data of land use type area in Shangri-La, it can be concluded that from 2013 to 2017, the conversion area between grassland and other land use types was relatively low, so land use is not a direct factor affecting grassland degradation in Shangri-La.

3.2.2. Grassland Degradation on Both Sides of the Road Caused by Tourism

In Shangri-La, the GDP of primary industry accounted for a relatively low percentage of the city’s GDP, dropping from 5.39% to 3.73% year by year from 2012 to 2017 (Figure 6) [17,18,19,20,21,22]. Shangri-La’s animal husbandry, which relies on rich natural grassland resources, is relatively developed and is the leading primary industry. Following the same trend as the population change, the output value of animal husbandry increased from RMB 182 million in 2012 to RMB 258 million in 2015, and the total output value was smaller than the annual GDP of Shangri-La (Figure 6 and

Table 3. Annual income of animal husbandry and tourism in Shangri-La from 2012 to 2017 (unit: billion RMB).

No.	Industries	2012	2013	2014	2015	2016	2017
1	Animal husbandry	1.82	2.22	2.41	2.41	2.53	2.58
2	Tourism	70.41	86.06	97.02	115.63	140.83	228.69

) [17,18,19,20,21,22]. However, in Shangri-La, the GDP of tertiary industry accounts for a relatively high percentage of the city’s GDP, with a minimum of 51.7% (2013) and a maximum of 61.4% (2015), of which tourism is the primary tertiary industry [17,18,19,20,21,22]. From 2012 to 2017, the GDP of the tourism industry increased year by year from RMB 7.041 billion to RMB 22.869 billion, with the highest annual growth rate at 62.39%. The year with the highest tourism GDP (2017) was 88.6 times the GDP of animal husbandry in 2017 (Table 3) [17,18,19,20,21,22]. Therefore, tourism is the pillar industry of Shangri-La’s economy.

The development of tourism in Shangri-La relies on the construction of traffic routes as highways connect various scenic spots in Shangri-La, enabling tourists to travel around [4]. Between 2015 and 2021, the total road length of Shangri-La reached 105–138.09 km [17,18,19,20,21,22]. The overall road length remains stable, but a few roads [17,18,19,20,21,22] carry a great deal of traffic. Therefore, the degradation of the grassland on both sides of the road is severe.

To sum up, there are various factors that influence grassland degradation in Shangri-La, but the external population has increased in Shangri-La due to the development of tourism, which led to frequent human activities around the roads, thus making the grassland degradation on both sides of the road more severe. Based on the particular process in the study area, we selected roads as the most critical factor to analyze the impact on grassland degradation in Shangri-La.

3.3. Analysis of Spatial Characteristics of Grassland Degradation on Both Sides of the Road

The Euclidean distance between the degraded grassland patch and the road (EDPR) is selected as the road feature and the Class Area (CA) and Patch Density (PD) as the landscape indexes for analysis. By analyzing the correlation between the road feature and the landscape index of grassland patches with different degradation grades, the influence coefficient curve of EDPR on grassland degradation grade patches CA value, and PD value is obtained (Figure 7). The results show that the patch distribution of each degradation level of grassland has prominent spatial characteristics.

3.3.1. Analysis of CA Value Change Characteristics of Grassland Degraded Patches on Both Sides of the Road

In the grade of heavily degraded grassland (Figure 7(a-1)), EDPR has a negative correlation with the CA value of heavily degraded grassland patches, and the quantile regression coefficient has an evident decreasing trend, indicating that the closer to the road, the larger the area of heavily degraded grassland patches. In the grade of moderately degraded grassland (Figure 7(a-2)), with the quantile point 0.68 as the boundary, the grade and the quantile point before the boundary are negatively associated (quantile point 0.2–0.68), while after the boundary is positively associated (quantile point 0.68–0.8). This finding shows that the high CA value of moderately degraded grassland patches is far from the road. In the grade of lightly degraded grassland (Figure 7(a-3)), EDPR is positively correlated with the CA value of lightly degraded grassland patches, indicating that the area of lightly degraded grassland patches gradually increases with the increase of EDPR. Similar to the lightly degraded grassland, EDPR is positively related to the CA value of undegraded grassland patches (Figure 7(a-4)), and the mean value of the quantile regression coefficient is greater than that of lightly degraded grassland, which indicates that a large area of undegraded grassland is distributed in areas away from roads.

Consequently, from the results of quantile regression, the correlation coefficient between EDPR and CA value of each grassland degradation grade varies, indicating that the spatial distribution pattern characteristics of different degraded grades of grassland are significantly different. The spatial distribution of grassland degradation on both sides of the road is as follows: the heavily degraded grassland and moderately degraded grassland patches are distributed in areas with a low EDPR, but heavily degraded grassland patches are distributed in areas closer to roads; with the increase of EDPR, the areas of lightly degraded grassland patches and undegraded grassland patches gradually increases, but the area of undegraded grassland increases more saliently.

3.3.2. Analysis of PD Value Change Characteristics of Grassland Degraded Patches on Both Sides of the Road

According to the quantile regression results of the EDPR and PD values of different grassland degradation grades, EDPR is negatively associated with the PD values of heavily degraded, moderately degraded, and lightly degraded grassland patches. (Figure 7b) However, EDPR has the most pronounced relationship with the PD value of heavily degraded patches of grassland, and then the moderately degraded and lightly degraded grasslands decrease in turn. This result shows that the number of grassland patches of heavily degraded, moderately degraded, and lightly degraded gradually increased with the decrease of EDPR, with the number of heavily degraded grassland patches increasing most significantly, which also indicates that the grassland on both sides of the road is heavily degraded. Furthermore, the correlation coefficient of EDPR to the PD value of undegraded grassland patches is negatively correlated before the quantile point of 0.6, and the minimum value of the correlation coefficient is greater than that of the correlation coefficient of heavily degraded, moderately degraded, and lightly degraded grassland patches; at the same time, the correlation coefficient is positively correlated after the quantile point 0.6, and the correlation coefficient increases gradually with the increase of the quantile point (Figure 7b). Based on this correlation, it can be concluded that the number of undegraded grassland patches is larger and denser, and the degree of grassland degradation is lower in areas farther from the road.

From the analysis of the change characteristics of the PD value of grassland degraded patches on both sides of the road, a conclusion can be reached: the distribution of undegraded grassland is relatively scattered, the number of degraded grassland patches is large, and the distribution of heavily degraded grassland is the densest in the area near to the road; as the distance from the road increases, the number of degraded grassland patches decreases, while the number of undegraded grassland patches increases with the concentrated distribution.

4. Discussion

4.1. Differences in Spatial Distribution Characteristics of Grassland Degradation at Different Degrees

There are differences in the spatial distribution of grassland degraded patches with different degrees. It was found that degraded grasslands are mainly distributed to either side of urban roads, while undegraded grasslands are mainly distributed in areas far from roads. This finding reflects the conclusion that road distribution has a specific impact on grassland degradation. The smaller the distance to the road is, the more frequent human activities are, and the higher the traffic flow on the road is, which results in severe damage to the vegetation on both sides of the road and brings about severe grassland degradation, while the farther away the road is, the less frequently the human activities are or even challenging to reach, which has less impact on grassland degradation.

4.2. Differences in the Impact of Roads on Varying Degrees of Grassland Degradation

Based on the quantile regression analysis of the road distance and each landscape index, it is concluded that the road has different impacts on different grassland degradation grades. It was found that the distribution of roads has the most significant effect on heavily degraded grassland among all grades of grassland degradation, followed by moderately degraded grassland and lightly degraded grassland. Meanwhile, undegraded grasslands were least affected by roads because they were far away from roads. Therefore, the distribution of roads will exacerbate the degradation of grasslands.

The smaller the distance from the road is, the more frequent the human activities are, and the more types of grassland degradation patches are. This trend is mainly manifested as follows: the closer to the road, the more the area of heavily degraded grassland patches increases; the patches are continuously distributed; and the shape of the patches tends to be simpler. Combined with the distribution of land utilization in Shangri-La City, due to the accessibility of traffic, most of the cultivated land and urban and rural human traffic are distributed along the roads, which also leads to grassland degradation [24]. As EDPR gradually ratchets up, the impact of roads on grassland degradation gradually decreases, the area of degraded grassland patches narrows, and the area of undegraded grassland patches enlarges significantly. At the same time, the complexity of the shape of grassland patches with varying degrees of degradation is also affected by EDPR. With the increase of EDPR, the shapes of heavily degraded patches and moderately degraded patches present a complex trend. This trend indicates that the degraded grassland is no longer continuously distributed in a large area after moving away from the road, but gradually discretizes. Conversely, the shapes of lightly degraded and undegraded grassland patches tend to be simpler with increasing road distance. In particular, the shape of undegraded grassland patches tends to be more single, and the patches are distributed continuously over a large area obviously. Therefore, when it comes to road construction, the differences in the impacts of varying degrees of degradation on the grassland should be considered: the route planning should be considered to avoid interfering with the undegraded grassland patches with a large area of continuous distribution.

4.3. The Combined Effect of Road and Other Related Factors

Grassland degradation on both sides of the road is affected by other natural and human factors [25,26]. Factors related to roads include the invasion of poisonous weeds [27], overgrazing [28,29], mining [30,31], changes in soil quality [32], expansion of cultivated land [33,34], random reclamation, indiscriminating digging, rodents, diseases [35,36], etc. The combined influence mechanism and process of road and other related factors are also relatively complex. In order to conduct a more comprehensive study on the influence mechanism of road factors on grassland degradation, the combined effect of road and other related factors should be investigated.

5. Conclusions

Grassland area degradation in Shangri-La is relatively severe. The total grassland area of Shangri-La is 2207.94 km², 824.03 km² of which is severely degraded, exceeding undegraded grassland by more than 172.62 km². Although the moderately and lightly degraded area of Shangri-La is relatively small, the total degraded area reaches 1556.54 km², accounting for 70.5% of the city’s total grassland area. In addition, the degraded grasslands in Shangri-La are mainly concentrated in the central and southern regions, while being relatively less common in the northern part of the city. Viewed from the east to the west, the grasslands in the eastern and central region are heavily degraded. In contrast, the western grasslands are relatively small, and most of the degraded grasslands are concentrated on either side of the highway.

Examining townships in the study area from south to north, the proportion of heavily degraded and moderately degraded grassland gradually decreases. In contrast, the proportion of lightly degraded and undegraded grassland gradually increases. Townships with a high proportion of degraded grassland are generally located in the south of the study area, with dense roads and well-established transport networks, while towns with a low proportion are generally located in the north of the study area, with dispersed roads and sparse traffic lines.

From the analysis of the change characteristics of the CA value of grassland degraded patches on both sides of the road, it can be concluded that heavily degraded grassland and moderately degraded grassland patches are distributed in areas with a low EDPR, while the heavily degraded grassland patches are in areas closer to the road; with the increase of EDPR, the areas of lightly degraded grassland patches and undegraded grassland patches gradually expand, with the area of undegraded grassland increasing more evidently.

From the analysis of the change characteristics of PD value of grassland degraded patches on both sides of the road, it can be deduced that the closer to the road, the more scattered the distribution of undegraded grassland is, the higher the number of degraded grassland patches is, and the denser the distribution of heavily degraded grassland is; as the distance from the road increases, the number of degraded grassland patches decreases, while the number of undegraded grassland patches grows with concentrating distribution.

Based on the basic situation of grassland degradation in Shangri-La, the utilization and conservation of grassland resources in Shangri-La should be improved to alleviate the severe trend of grassland degradation in Shangri-La. For varying grades of grassland degradation, classified governance should be carried out. According to the analysis of the impact of road distance on grassland degradation, human activities carried by roads, such as tourism development, urban and rural construction, transportation, and agricultural activities distributed alongside roads, all lead to the degradation of grasslands and even aggravate grassland degradation. Therefore, when formulating a specific plan to control grassland degradation, on the one hand, the distribution of road construction should be adjusted to alleviate the grassland degradation in Shangri-La. On the other hand, while developing tourism and animal husbandry, attention should be paid to protecting grassland resources, advocating civilized tourism, reducing trampling, promoting rational grazing, and coordinating the development of tourism and animal husbandry with the sustainable utilization of grassland resources.