Evolution and Ecological Implications of Land Development and Conservation Patterns on the Qinghai-Tibet Plateau

Abstract

The Qinghai-Tibet Plateau serves as an important ecological security shelter in China and across Asia. During the past two decades, the patterns of land development and conservation on the Qinghai-Tibet Plateau have undergone significant changes under the impacts of global climate change and human expansion. This paper analyzes the evolution of land development and conservation patterns and potential ecological implications on the Tibetan Plateau from 2000 to 2020 based on urbanization, agricultural and pastoral patterns with multi-source data, such as long-term time series land use data, ecological indices, environmental pollution, and population and economics. It was found that: (1) Coinciding with the agglomeration of population and economy, the urban development pattern on the Qinghai-Tibet Plateau has spatial-temporal differentiation. Urban development in the 2010s was more significant than that in the 2000s, with the intensity increased by 63.31%, and the development pattern varies spatially, which can be seen from the finding that central Tibet (the Yarlung Tsangpo, Lhasa river, and Nyangchu basins) and Huangshui River Basin in Qinghai Province were developed in a planar pattern, while other node cities and border towns in a dotted pattern. (2) The agricultural production pattern is relatively stable, the grain yields have increased by 34.68% in the past 20 years, but the total amount of arable land is shrinking, and the degree of fragmentation has increased. The pattern of pastoralism has changed greatly, showing a migration trend from west to east spatially, and there is a serious problem of pasture overload, with an increase of 18.4% in livestock products. Regardless of the pattern of urbanization or agricultural and livestock development, the intensity of human activities on the Qinghai-Tibet Plateau has shown an intensified trend in the past. (3) It shows that Lhasa City area and the middle and lower reaches of Huangshui show a trend of diffusion of atmospheric and water environmental pollution. The western part of the Ali region and the northern foothills of the Himalayas and other regions, under the combined impact of climate change and human activities, have undergone significant ecological degradation. Accordingly, this paper proposes policy recommendations for optimizing production and living space, advancing the establishment of national park clusters and nature reserve systems, and the integrated recovery of mountain, water, forest, lake, grass, sand, and ice with ecological value achievement.

1. Introduction

Formed from plate collision, the Qinghai-Tibet Plateau is known as the “Roof of the World” and the “Water Tower of Asia”, which is not only the birthplace of the Yangtze, Yellow, Lancang, Yarlung Zangbo, Mekong, and other rivers, but also an important barrier to maintain the stability of the natural environment in East Asia and beyond [1,2,3]. Covered by glaciers, permafrost, and other frozen ground, the Qinghai-Tibet Plateau has a unique climate system categorized by low air temperatures, high daily temperature differences, low annual temperature differences, and strong solar radiation [4]. The high altitude and cold climate of the Qinghai-Tibet Plateau give birth to a rich variety of rare and endemic plants and animals on the plateau, making it a key area for global biodiversity protection. However, at the same time, the Qinghai-Tibet Plateau is also the most typical ecologically fragile area. The special climatic conditions and topography have meant that the ecosystem of the Qinghai-Tibet Plateau has been in a fragile and unstable state for a long time. It is easily affected by global climate change and the disturbance of human activities [5,6,7]. Some studies have shown that under the background of continuous warming and wetting of the global climate, the area of frozen soil, marshes, and wetlands in the Qinghai-Tibet Plateau will be greatly reduced [8,9,10]. The resulting ecosystem changes will endanger the sustainable development of social economy in the region. In addition, with the intensification of human activities, the rapid development of the population and social economy has also caused great pressure on the fragile ecological environment, such as landscape fragmentation caused by urban expansion [11,12,13], degradation of grassland by overgrazing [14,15,16], etc.

Ecologically fragile areas are much more sensitive to climate change and human activities than other areas and they are typical cases for studying the interaction between natural systems and human social systems. At present, many scholars have used the evolution of land use patterns and functions to reveal the change in the human–land relationship in ecologically fragile areas. From the research perspective, they can be divided into three categories. Firstly, the land cover data obtained from satellite remote sensing images can be used to describe the spatial organization rules of the area, proportion, and interaction of different land use functions, including spatial agglomeration, spatial differentiation, etc. [17,18,19]. Secondly, the evolution process and mechanism of the regional ecosystem function pattern can be analyzed from the perspective of landscape ecology [20,21,22]. Most of these studies establish the relationship between ecosystem services and land use patterns through various indices, thus reflecting the environmental and ecological effects brought by human activities. The third is research regarding the regional spatial structure, which is represented by regional economics and economic geography. This kind of research focuses on the analysis of the characteristics and the mechanism of the evolution of social and economic patterns, including the spatial structure of the population [23], the urban system [24], the tourism spatial structure [25], etc. Most of the existing studies select fixed time intervals or time nodes with significant changes from multi-temporal and cross-sectional perspectives for analysis. These studies lack temporally continuous analysis based on long-time-series data that can accurately reflect the evolution of spatial patterns over a period of time.

In addition, the change in the human–land relationship will also lead to a series of spatial effects on various systems, such as the ecological environment and the socioeconomic system. The impact of these spatial effects on ecologically fragile areas is likely to be long-term, tending to form path dependence, such as economic clustering, population agglomeration, environmental pollution, ecological degradation, etc. [26,27,28]. Many scholars across the world have made the exploration in quantification and simulation of spatial effects. Among them, the study of socioeconomic effects mainly focuses on the driving effects of various land use changes on the socioeconomic system. For example, the intensification of cultivated land promotes the agriculture and pastoral economy [29], the positive effect of land policy on the social economy [30,31], and so on. Most of these studies use econometric models to quantify spatial effects. The research on ecological and environmental effects mainly relies on resource and environment carrying capacity or environmental and ecological assessment, emphasizing the impact on water resources and ecology, including water pollution [32], water use efficiency [33], and changes in ecosystem quality and ecological service function [34,35]. However, ecological and environmental effects are often studied separately, and few scholars have integrated them and conducted coupling analyses between environmental and ecological effects and social economy.

As a typical ecologically fragile area, the evolution of the spatial structure of land use in the Qinghai-Tibet Plateau involves several subsystems, such as environment, resources, ecology, and society, which have a strong synthesis. Therefore, it is necessary to reveal the evolution process and mechanism from multiple perspectives, such as landscape ecological pattern and socioeconomic spatial organization. At present, the relevant research in this region mainly focuses on climate change and human activities. Based on the perspectives of different disciplines, the evolution of the spatial pattern in the Qinghai-Tibet Plateau has been studied, including land use/cover [23,36], carbon storage [37,38], landscape ecology [39,40,41], etc. However, most of the existing studies on the Qinghai-Tibet Plateau only focus on its fragile ecological pattern or socioeconomic development but fail to combine the perspectives of regional economic spatial structure, landscape ecological function, and ecological spatial structure. In terms of research dimensions, they often focus on changes in large-scale functions and ignore the characterization of small-scale features such as structure and elements [42,43]. In addition, the data sources selected by most studies are relatively simple, which makes it difficult to meet the requirements of multi-scale analysis and induction. There are few types of research on the spatial effects of land use patterns from the aspect which combines perspectives of social economy, environment, and ecology, using multi-source data, such as land cover, environmental monitoring data, and social economic statistics data.

This paper explored the theory regarding the evolution process of land use structure and its spatial effects in ecologically fragile areas and divided the urbanization pattern, agricultural pattern, and pastoral pattern on the Qinghai-Tibet Plateau, based on the provincial major function-oriented zoning planning issued by Tibet Autonomous Region and Qinghai Province (Figure 1). Using multi-source data such as land use data interpreted from remote-sensing images, night light data, ecological environment remote sensing indexes, and socioeconomic statistics, we measured the characteristics of the spatial evolution of urbanization development, agriculture, and pastoral development on the Qinghai-Tibet Plateau in the past 20 years. Through the coupling analysis of the evolution of social economy, ecological environment, and the land use development and protection pattern, the possible spatial effects were analyzed. Accordingly, we put forward policy suggestions for optimizing the land use development and protection pattern.

2. Data and Methods

2.1. Data

The data sources of this study mainly include four categories: First, the classification data of different types of spatial structures on a district and county scale, which were obtained from provincial major function-oriented zoning planning issued by Tibet Autonomous Region and Qinghai Province. The second are the long-time-series land use data produced from multi-source remote sensing and Landsat image data. The third are CO, SO₂, water quality, and other environmental data interpolated from the ecological environment monitoring stations or the ecological environment bulletin. The fourth are the socioeconomic statistics, including resident population, GDP, grain output, livestock output, etc. The data cover the period 2000 to 2020, including at least 3 years in 2000, 2010, and 2020. Some long-time-series socioeconomic data have a 5-year or annual interval.

Land use remote sensing classification products included the China Land Cover Dataset (CLCD) [44] and the China Land Use Monitoring Dataset (CLUMD) (https://www.resdc.cn, accessed on 8 May 2022). The former is an interannual product from 1985 to 2020, while the latter is a multi-temporal product with a 5-year interval. Environmental data include three district (or county)-level pollutant monitoring datasets: the surface CO concentration data, surface SO₂ mass density data, and river water quality level data. The surface CO concentration data were obtained from the China National Environmental Monitoring Centre (CNEMC), which is based on the environmental monitoring stations distributed across the country. It is available on the real-time urban air quality open platform (https://air.cnemc.cn:18007, accessed on 15 June 2022), with a time resolution of one day. The surface SO₂ mass density data were derived from NASA’s MERRA-2 (the second Modern-Era Retrospective analysis for Research and Applications) project, and the data we used were M2TMNXAER [45]. The river water quality data were obtained from hydrological observation stations distributed on the Qinghai-Tibet Plateau. According to “the Standards for Surface Water Environmental Quality” (GB3838-2002), the data were divided into five grades, from I to V. The socioeconomic statistics mainly came from the public data of government departments in Qinghai and Tibet provinces. The years of data collection were 2000, 2005, 2010, 2015, and 2020.

2.2. Methods

The method of this paper mainly includes three parts: (1) Fusion of multi-source land use data and spatiotemporal consistency analysis, aimed at dealing with spatiotemporal inconsistency in land use classification products in this region. (2) Spatiotemporal cube segmentation and clustering, to realize the classification and identification of different temporal and spatial evolution patterns. (3) The coupling analysis of environmental and ecological effects: through coupling different spatial and temporal evolution modes of land use with environmental pollution and ecological quality evolution, we analyzed the environmental and ecological effects possibly caused by the evolution of different patterns (Figure 2).

(1)

Fusion of multi-source land use data and spatiotemporal consistency analysis

Due to the large scope of the Qinghai-Tibet Plateau, the poor overall image quality, and since the bare land is easily confused with other ground features, the overall accuracy of land use classification on the Qinghai-Tibet Plateau is not satisfactory. Existing remote sensing image interpretation products generally have problems, such as the inconsistency in time and space, and the identification accuracy of urban land is seriously low, which seriously affects the accuracy of the analysis of the evolution and effects of the land development and protection pattern on the Qinghai-Tibet Plateau. To integrate the common advantages of the long-time-series CLCD and the high-accuracy CLUMD, this paper used a bidirectional consistency detection algorithm to fuse and correct the consistency by constructing a sliding window, w_m, and a seed window, w_s. The main processing steps are:

① Set the bidirectional sliding window, w_m: its plane length and width are the sizes of the spatial neighborhood window of n pixels, and the height is the time series length x, where n and x are preset values: n is a positive odd number that is not 1, and x is a natural number greater than 0.

② Place the sliding window, w_m, immediately outside the seed window, w_s.

③ Calculate the dominant land type, f_m, in the sliding window, w_m, where the calculation follows this method: if the occurrence frequency of land use type α in the sliding window, w_m, is higher than r, then call α the dominant land type, f_m, of the sliding window, w_m; if no land use type meets the standard, then confirm no dominant type in the sliding window, and the value of r is usually set from 0.6 to 0.8.

④ Determine whether the feature type f_s in the seed window, w_s, is consistent with the dominant land type, f_m, in the immediate sliding detection window, w_m. If they are consistent, set all pixels between the first and the last dominant land use type pixels in the sliding detection window, w_m, to f_s, and move the position of the seed window to the grid position, where the dominant class f_s finally appears in the sliding window, and go to step ③ until the two sliding detection windows meet. If they are different, move the seed window one pixel inwards, then go to step ③ until the two sliding detection windows meet.

⑤ Perform the above steps on the CLCD and CLUMD datasets in turn and replace the CLCD data on corresponding time nodes with the processed CLUMD dataset, then perform consistency processing again.

(2)

Space–time cube segmentation and clustering

In this paper, a grid made up of 10 km space–time cube units is constructed, and the above-mentioned integrated land use interpretation products are segmented with the grid, and then the K-medoids clustering method is used to realize the clustering of the space–time cube unit to reveal the spatial patterns of the space–time cubes. Proceed as follows:

First, take the proportions of land use types as functional parameters, and count the proportions of urban land, cultivated land, and grassland in each cube unit year-by-year:

$$P_t^{m,q}=\frac{s_t^{m,q}}{s}$$

(1)

where $s_t^{m,q}$ represents the area occupied by the land use type q in the year t in the cube unit m, and s is the area of the cube unit.

Secondly, according to the characteristics of the long-time-series data of functional parameters, take the Euclidean distance as the similarity index, and use the K-medoids clustering method to automatically cluster the long-time-series data of the functional parameters in all cube units. The similarity index is calculated by the following formula:

$$dist\left(i,j\right)=\sqrt{{{\displaystyle \sum }}_{t=1}^n{(P_t^i-P_t^j)}^2}$$

(2)

where $dist\left(i,j\right)$ represents the similarity between cube i and cube j, n represents the length of the data, t represents time node t, and $P_t^i$ and $P_t^j$ represent the functional parameters of the cube i and cube j on time node t.

Finally, extract the values of different categories and temporal characteristic curves and merge the categories with high similarity until the differences between the categories are significant. According to the overall evolution characteristics on the Qinghai-Tibet Plateau, the urbanization pattern is divided into four categories: stable low value, micro-growth, medium–low-intensity development and continuous growth, and high-intensity development and continuous growth. The agricultural and pastoral pattern is divided into stable low value, substantial reduction, slight reduction, and slight increase, and the livestock pattern is divided into stable low value, slight reduction, slight increase, and stable high value.

(3)

Coupling analysis of environmental and ecological effects

We coupled the evolution of the urbanization pattern, agricultural and pastoral pattern, and livestock pattern with the results of environment quality assessment and ecology quality assessment, to analyze the environmental and ecological effects of the evolution of the land use development and protection pattern. Among them, the environmental quality assessment mainly uses the long-term monitoring data of two types of environmental pollutants, CO and SO₂, and the latest water quality data. In this paper, the inverse distance weighted (IDW) method was used to generate CO concentration raster data on the Qinghai-Tibet Plateau with a spatial resolution of 500 m year-by-year from the surface CO concentration monitoring data. At the same time, this paper used the bilinear interpolation method to interpolate the M2TMNXAER dataset year-by-year to generate raster data of SO₂ mass concentration on the Qinghai-Tibet Plateau with a spatial resolution of 1.1 km. Due to the strong diffusion effect of air environmental pollution, it is difficult to measure the environmental and ecological effects in a small scale. Therefore, this study used the districts and counties in the Qinghai-Tibet region as the unit to divide the CO and SO₂ raster data and count the mean pollutant concentration in each district/county. Based on the “Ambient Air Quality Standard” (GB 3095-2012), the concentrations of CO and SO₂ pollutants in districts and counties were grfiaded year-by-year. According to the year-by-year grading results and the water quality of the districts and counties, the environmental assessment parameter, $E_c$, is calculated as follows:

$$E_c=\frac{0.4}{m}{\displaystyle \sum }_{i=1}^m{R}_{CO,i}+\frac{0.4}{n}{\displaystyle \sum }_{j=1}^n{R}_{SO2,j}+0.2\ast R_{water}$$

(3)

where $R_{CO,i}$ represents the grade of CO in year $i$, $R_{SO2,j}$ represents the grade of SO₂ in year j, and $R_{water}$ is the water quality data.

Ecological effects were characterized by a vegetation index, KNDVI (kernel normalized difference vegetation index). KNDVI is a type of vegetation index proposed by Gustau Camps-Valls in 2021. It optimizes the saturation of NDVI in high-value areas and can linearly characterize the surface vegetation biomass [46]. The formula for calculating KNDVI is:

$$KNDVI=\frac{k\left(nir,nir\right)-k\left(nir,r\right)}{k\left(nir,nir\right)+k\left(nir,r\right)}$$

(4)

where nir is the reflectivity of the near-infrared band, r is the reflectivity of the red band, and $k\left(a,b\right)$ is the kernel function between a and b. The radial basis function (RBF kernel) is generally used:

$$k\left(a,b\right)=exp\left(-\frac{\left(a-b\right)²}{2\sigma ²}\right)$$

(5)

This paper used the Google Earth Engine (GEE) to obtain KNDVI long-time-series data. First, the Landsat images were processed through cloud filtering, cloud removal, and other steps. The images from May to August with cloud cover less than 15% were selected, and the mean fusion was performed pixel-by-pixel to generate annual Landsat fusion images of the Qinghai-Tibet Plateau. Based on the near-infrared (NIR) and red light (R) bands of the fusion image, annual KNDVI raster data with a spatial resolution of 30 m were generated. Using the 10 km cube unit, the KNDVI raster data were segmented year-by-year, and the average KNDVI value in each unit was calculated. Based on the geometric series method, the units were graded year-by-year, the long-time-series analysis of the graded results was carried out, and the cube units were divided into four evolution modes: stable low value, reduction, increase, and stable high value.

3. Results

3.1. Space–Time Expansion of Urbanization Development

The urbanization development pattern of Qinghai-Tibet Plateau is an urbanization strategy pattern that is mainly based on central Tibet (Yarlung Tsangpo, Lhasa River, and Nyangchu basins) and Huangshui valley, together with other node towns and border towns, including “one axis and two groups (districts)” in Qinghai Province and “one circle, two wings, and three points” in Tibet Autonomous Region. Overall, the urbanization scale of the Qinghai-Tibet Plateau shows an obvious growth in the whole region from 2000 to 2020. The construction land area increased from 533.33 km² in 2000 to 1121.82 km² in 2020. Among them, the development intensity of the three types, which include the grids of high-intensity development and continuous growth, micro-growth, and medium–low-intensity development and continuous growth, increased from 7.66%, 0.16%, and 2.22%, to 11.28%, 0.38%, and 5.21%, respectively. New construction land was mainly distributed in central Tibet (Yarlung Tsangpo, Lhasa River, and Nyangchu basins) and Huangshui River Basin, with the growth rate of construction land in the two regions reaching 195.41% and 70.12%, respectively. Located in river valleys with flat terrain and rich water resources, these two areas have been the core areas of Qinghai Province and Tibet Autonomous Region, with a long history of development and a better level of urbanization and socioeconomic development than other areas of the Tibetan Plateau. From Figure 3, it can be seen that the urban agglomerations of central Tibet and Huangshui River Basin have different degrees of outward expansion, and they both show a polygon expansion of nearly concentric circles. Take Xining-Haidong urban agglomeration as an example: the intensity of construction land development in its central urban area remained at a high level (more than 28%), while the periphery of the city gradually transitioned from a fast growth area to a slow growth area.

Except for central Tibet and the Huangshui River Basin, which originally developed rapidly, a number of new node towns emerged in the Tibetan plateau from 2000 to 2020, such as Mangye City, Golmud City, and Delingha City in Haixi Prefecture, and Republican County, Gonghe County, and Naqu City in Hainan Prefecture in Qinghai (Figure 3). The growth rate of the medium–low-intensity development and continuous growth grid in these areas reached more than 11.54%. In addition, the towns in the southeast of Yushu prefecture, Qinghai Province, at the border with Tibet and in the border area of southern Tibet, have also shown a trend of slow expansion in the past two decades. For example, in Nangqian County, Yushu City, Shona County, and Motuo County, the development intensity of the micro-growth grid ranges from 0.24% to 0.33%. From a phased perspective, the urban development in the Qinghai-Tibet Plateau from 2000 to 2010 mainly concentrated in Qinghai Province, including Xining-Haidong urban agglomeration, Golmud-Delingha urban agglomeration, and other areas with good urban construction foundations. Among them, the expansion of construction land in Xining and Golmud-Delingha urban agglomeration reached 110.86 and 126.76 km², respectively. From 2010 to 2020, except for the above-mentioned areas maintaining a stable expansion, other areas also showed an obvious expansion of urban construction land, such as downtown Lhasa, Nagqu, and Republican County of Hainan Prefecture. Compared with the first stage, the urban development in the Tibetan region is more obvious in the second stage, especially in the area of central Tibet with Lhasa-Zedang urban agglomeration as the core. The new construction land in this area reached 207.23 km², accounting for 90.09% of the total expansion scale.

3.2. Space–Time Changes in Agricultural and Livestock Patterns

The development pattern of agricultural and pastoral areas on the Qinghai-Tibet Plateau consists of two types of development patterns: an agriculture pattern and a livestock pattern. Among them, the agricultural pattern includes the eastern part of Hehuang basin, Qaidam Oasis, the periphery of Qinghai Lake, the middle and upper reaches of Yarlung Zangbo River area, the middle reaches of Yarlung Tsangpo River—Lhasa River area, the middle and lower reaches of Niyang River, the southeast of Tibet, and other areas. The livestock pattern includes the livestock belt around Qinghai Lake, the ecological livestock area in south Qinghai, the northwest of Tibet, the south of Qiangtang Plateau, northeast Tibet, and other livestock areas. In the past 20 years, the agricultural pattern in the Qinghai Tibet region has been relatively concentrated, and the substantial reduction grid (cultivated land proportion accounting for more than 45%) is mainly distributed in three major agricultural regions in Qinghai Province: the Eastern Agricultural Region, the Agricultural Region around Qinghai Lake, and the Qaidam Agricultural Region. The number of slight increase grids is 260, and the distribution is relatively scattered, mainly distributed in Gonghe County and Guide County around Qinghai Lake, the middle reaches of Yarlung Zangbo River—Lhasa River agricultural area, and the southern Tibet agricultural area. Besides, there is a partial decay of cultivated land area in Gyantse County, Sanzhuz District, and Lhazhi County of Rikaze City, and Duilongdeqing District and Mozhugongka County of Lhasa City. From the changes of the proportion of cultivated land in the grid (Figure 4), the total amount of cultivated land in the Tibetan Plateau region showed a shrinking trend from 2000 to 2020, in which the average proportion of cultivated land in the substantial reduction grid decreased from 33.79% to 28.43%, and the proportion of cultivated land in the slight increase grid only increased from 2.92% to 3.68%.

In the Tibetan Plateau region, grassland is widely distributed. Except for the Qaidam Basin, Linzhi City, and Cona County of Tibet, the proportions of grasslands in other areas are at a constant high value (stable high grid), and the average proportion of grasslands in the grid reached 91.85% in 2020. The changes of the livestock pattern in the Tibetan Plateau region are mainly reflected in the spatial layout in the past two decades, showing a trend of migration from west to east. In terms of the total amount, the area of grassland in the Tibetan Plateau region has remained basically unchanged, with the decreasing rate only reaching 0.08%. In the spatial pattern, the changes of grassland area from 2000 to 2020 mainly occurred in the peripheral areas of the livestock pattern (Figure 5). In the western and southern areas, such as the northern part of Ali region and Nagqu city of Tibet, the central part of Shannan city, and the border of Linzhi City, grassland areas showed an obvious declining trend. The slight increase grids are distributed near Haixi Prefecture in the northeast and Changdu City in the southeast. In addition, there are scattered slight increase grids in the northwest of Ali region. The proportion of grassland in these grids has increased from 1.91% to 52.52% in the last two decades.

3.3. Analysis of Environmental and Ecological Effects

The environmental quality evaluation results show that the overall environmental quality of the Qinghai-Tibet Plateau is good, and 61.3% of the districts and counties are at the excellent level. However, there is also local pollution, and 16.8% of the districts and counties are at the level of serious pollution, mainly distributed in Lhasa and Shannan areas of Tibet, and Xining Haidong and Golmud Delingha regions of Qinghai Province (Figure 6). Among them, the partial pollution in the Tibetan area and Golmud-Delingha in Qinghai is mainly contributed by SO₂, while in Xining-Haidong area in Qinghai province it is due to the combined effect of CO and SO₂. Overall, 70% of the seriously polluted areas have also seen a significant growth in urbanization. In total, there are four districts and counties in Tibetan region with serious environmental pollution. They are Sangzhuizi district in Rikaze, Dulongdeqing and Chengguan districts in Lhasa, and Naidong district in Shannan. All of them are the areas with the most significant urbanization expansion in Tibet. In addition, some important node cities in Tibet, such as Kargil County and Saini District in Nagqu City, also have light environmental pollution. Compared with Tibet, the environmental pollution problem in Qinghai is more serious, and the main pollution is distributed in the districts and counties to the north of Kunlun Mountain, which is also the main concentration area of urban agglomeration. The environmental evaluation results of 17 districts and counties, including the main urban areas of Xining and Haidong City, Mangye City, and Duran County, are severely polluted. These areas are high-value areas for urbanization development and agricultural and livestock production.

Through the aggregation analysis of the long-term KNDVI, we found that the overall ecology of the Qinghai Tibet Plateau showed a positive trend (Figure 7). Except for the desert areas in the Qaidam Basin and the water area of Qinghai Lake, the KNDVIs in most areas are in the growth stage or maintained at a high level. Among them, the KNDVI of the Qiangtang Plateau desert area in the northwest and southwest of Tibet, the Yarlung-Zangbo River basin, the west of the Three Rivers, and the southern foothills of the Qilian Mountains all show an increasing trend. Even in the relatively developed urban agglomerations such as Xining-Haidong and Lhasa-Zedang, the KNDVI has maintained high values or increased faster. This indicates that the ecological restoration projects that have been continuously promoted in the Tibetan Plateau region over the past 20 years, such as returning cultivated land to forests and grasses, sand prevention and control projects, and urban greening projects, have achieved significant results.

However, the ecology of the Tibetan Plateau also has local degradation. One type of area includes the extension area of the expansion of towns such as central Tibet (Yarlung Tsangpo, Lhasa River, and Nyangchu basins) and Qinghai Xining-Haidong. Other types of areas include the western Ali region, the northern foothills of the Himalayas, Selincuo, Namucuo of Naqu City, and other regions. The former is due to the increase in ecological pressure caused by the expansion of human activities, and the latter is mostly affected by the expansion of water and local drought caused by climate change. The decreasing range of KNDVI in these regions ranges from 1.45% to 58.24%.

4. Discussion

The land use pattern is a complex structure in which human production and living activities act on the natural ecosystem, and the spatial effects caused by its changes have been widely discussed [47,48,49]. As a typical ecotone, the Qinghai-Tibet Plateau can withstand only a very limited intensity of human activities. When the scope and intensity of human activities exceed natural thresholds, micro-steady-state transitions will occur. Then, it leads to macro-functional and structural changes through adaptive cycles, positive and negative feedback, and scale-spanning mechanisms [50,51,52]. Related studies are often limited by the validity of data and methods while revealing this issue [53,54].

The highest resolution of publicly available land use data on the Tibetan Plateau is 30 m, mainly from the Landsat series of satellites, whose important basis is the quality of Landsat images. The aging of Landsat 5 satellite sensors and the severe banding of Landsat 7 images existing in the Landsat series at around 2010 often led to the lack of usable images of the Tibetan Plateau in specific years, which was an important reason for the low accuracy of remote sensing interpretation data. Consistency correction of multi-source data fusion and long-time-series space–time data can improve land use classification accuracy to some extent, and it is especially necessary for the lower development intensity of ecologically fragile regions such as the Qinghai-Tibet Plateau itself, as it has been confirmed in some related studies [55,56]. When studying the coupling relationship between the land development pattern and environmental and ecological effects, the size of the assessment unit is also a key issue worth exploring. On the one hand, despite the rapid growth of built-up land in the last two decades, the share remains particularly small. It is necessary to choose an optimal spatial scale to bring out the characteristics of this kind of change [57,58,59]. On the other hand, the pollutant surface concentration data are generated by interpolation of meteorological monitoring station data, and the accuracy is slightly lower in western Tibet, where the distribution of meteorological monitoring stations is sparse. In addition, considering the diffusion effect of pollutants, district and county units were used as the assessment objects for environmental effect assessment.

We coupled the results of the land use development pattern, population economy, and environmental and ecological effects on the Tibetan Plateau to further explore the possible spatial effects that may be triggered by the evolution of the development pattern. In the past 20 years, the urbanization development of the Qinghai-Tibet Plateau shows an obvious trend of expansion, where both population and economy have further concentrated in urbanized areas. The population urbanization rates of Qinghai and Tibet increased from 32.33% and 19.30% to 61.02% and 35.73%, respectively. From 2010 to 2020, the average annual growth rate of GDP of Tibetan districts and counties reached 64.56%. Among them, central Tibet (Yarlung Tsangpo, Lhasa River, and Nyangchu basins) was as high as 68.17%, while the GDP of Huangshui River Basin (Xining-Haidong urban agglomeration) in Qinghai province also increased from 13.094 billion yuan in 2000 to 191.356 billion yuan in 2020. Urbanization development not only further concentrates the population and economy, but also brings serious negative effects on the environment. Combining Figure 3 and Figure 6, it is not difficult to find that the environmental quality assessment results are highly consistent with the urbanization development pattern. The two areas with the most concentrated urbanization, central Tibet and Hehuang Valley, are the most seriously polluted areas, which is roughly the same as the results of Yin’s and Tian’s studies [60,61]. With the expansion of urbanization, there is also a risk of spillover of ecological and environmental stresses, especially the trend of expansion to agricultural and pastoral development areas. In the environmental evaluation results, the peripheral areas of these urban agglomerations show different degrees of pollution, and this finding is consistent with many previous studies [62,63,64].

From the previous analysis, it can be seen that the spatial pattern of agriculture and pastoral areas on the Tibetan plateau has changed to some extent. The pastoral (grassland) area is declining in western Tibet and central Qinghai and expanding in eastern Tibet, which is consistent with the results of existing studies [65,66], while the patterns of agricultural (cropland) areas are declining in one river basin and expanding in eastern Tibet, which has been verified in Wang’s and Luo’s studies [23,43]. In terms of area change, the proportion of pastoral (grassland) area remains unchanged, while the area of agriculture (cropland) has gradually decreased. In the past two decades, agricultural and pastoral production has been showing an upward trend, and the grain production in the main agricultural areas has increased by 34.68%, from 1.373 million tons in 2000 to 1.849 million tons in 2020, while the total meat production in the main pastoral areas has increased from 362,000 tons to 429,000 tons, an increase of 18.4%. The increase in agricultural and pastoral production due to technological advances cannot be ruled out here, but on the other hand, it also increased the ecological pressure on the Tibetan Plateau region. For example, the increase in agricultural production will increase the pressure on water use and may lead to a shortage of surface water resources [67]. Studies have shown that the actual livestock carrying capacity of the Tibetan Plateau has been more than 1.6 times the theoretical capacity over the past 50 years [68]. In Figure 5, it can be seen that the KNDVI in the northwest Tibetan pastoral area shows a decreasing trend, which is partly due to the effects of changes in ecological physical processes caused by climate change, resulting in a steady-state shift of the originally fragile ecosystem across the ecological threshold [66], and partly due to the degradation of grasslands brought about by pasture overload [14].

From the previous analysis and the conclusions of existing studies, it can be seen that the ecology of the Qinghai Tibet Plateau shows a positive development, and the overall vegetation biomass shows a trend of growth [69]. On the one hand, this is partly due to the positive ecological effect of regional ecological projects [70,71], and on the other hand, since the permafrost zone is concentrated in central Tibet and southwest Qinghai, with a high spatial coupling to the area with growth in vegetation biomass, this trend may also be caused by the improvement of hydrothermal conditions due to the melting of permafrost under the influence of climate warming [72,73,74]. Although the ecology of the Qinghai Tibet Plateau is generally developing well, the environmental pollution caused by the acceleration of urbanization is also very prominent, especially in central Tibet (Yarlung Tsangpo, Lhasa River, and Nyangchu basins) and the Huangshui River Basin. Therefore, coordinating the relationship between urbanization, agriculture and animal husbandry development, tourism development, national park group construction, and ecological barrier construction is the key to the sustainable development of the Tibetan Plateau region. On the one hand, the scale of agriculture and animal husbandry in the Tibetan Plateau region should be properly controlled, especially in ecologically fragile areas such as deserts, to reduce the pressure on the ecological environment brought by human activities [14]. On the other hand, even in a suitable construction area, the development intensity should be strictly limited within the carrying capacity of resources and the environment. In addition, through the construction of various types of nature reserves, national parks, major ecological projects, and other construction projects, a nature reserve system with national park groups as the main body will be established. We also need to strengthen the research on the impact of global climate change on the Qinghai-Tibet Plateau ecosystem, improve the adaptability of ecologically fragile and sensitive areas to climate change, and promote the integrated protection and restoration of mountains, water, forests, fields, lakes, grasses, sand, and ice by regions and steps. It is necessary to establish a real-time monitoring platform for ecological conditions and expansion of human activities in the Qinghai-Tibet Plateau region that allows for prompt regulation and to form a long-term mechanism that is standardized, normalized and institutionalized.

5. Conclusions

In the past 20 years, urban development on the Qinghai-Tibet Plateau has been significant and differentiated, which has resulted in the agglomeration of the population and economy as well as increased pollution. Urban land expanded by 110.34% over the period, mainly in central Tibet (the Yarlung Tsangpo, Lhasa River, and Nyangchu basins), the Huangshui River Basin in Qinghai Province, and the Qaidam Basin. In the last decade, the development intensity was more significant than in the previous decade, with a 63.31% growth. Tibet lagged behind Qinghai in terms of time and scope, and other important node cities and border towns have also recently shown a dotted development trend. The development of urbanization promotes the further concentration of the population and economy. The population growth rate, urbanization rate, and economic growth rate are significantly higher than the national average during the same period, but it has also caused significant environmental and ecological effects and ecological degradation. The environmental and ecological effects are mainly manifested in the tendency of air pollutants to spread from the towns to the overall watershed and basin areas, and the quality of the downstream water environment tends to deteriorate, which coincides with the time sequence of urbanization development and expansion.

The agricultural and pastoral pattern is relatively stable, but the agricultural and pastoral production show noticeable growth. Cropland is mainly distributed in Huangshui River Basin, the area around Qinghai Lake, Qaidam Basin, Mid-Yarlung-Tsangpo River and Lhasa River agricultural areas, and the southern Tibetan agricultural area. Grain production has increased by 34.68% in the past 20 years, but the area of cultivated land is shrinking, and the degree of fragmentation has increased. The livestock pattern has changed greatly, with the overall trend showing movement from west to east, and the western and southern grassland has significantly reduced. However, the rapid growth of the population and urbanization has led to a significant decline in pastureland, which has further increased the demand for cultivated land, agricultural, and livestock products, with increases of 110.34%, 34.68%, and 18.4%, respectively. These have led to a series of problems such as grass degradation, environmental pollution, etc.

Although the urbanization pattern and the changes in agricultural and pastoral patterns have made some impacts on the ecological environment of the Tibetan Plateau, the overall ecology has shown a positive trend, with the KNDVI in most areas increasing or remaining at high values. This could be the result of nearly 20 years of continuous ecological restoration projects, such as the nearly decade-long project of returning farmland to forest and grass, the sand prevention and control project, the soil erosion control project, and the urban greening project. However, due to the fragile natural conditions of the Tibetan Plateau, human activities, such as urbanization and grazing, need to be controlled within a certain range and intensity, so that systematic ecological degradation can be avoided. The phenomenon of grassland degradation due to overgrazing also exists in the pastoral areas of northwest Tibet. In addition, we need to strengthen the adaptive management of climate change in the future by promoting the integrated protection and restoration of mountains, water, forests, fields, lakes, grasses, sand, and ice on a regional scale and in a step-by-step manner to further optimize the spatial pattern/structure of human activities and the ecological environment on the Tibetan Plateau.