Quantitative Assessment of Urban Expansion Impact on Vegetation in the Lanzhou–Xining Urban Agglomeration

Abstract

The Rapid expansion of the Lanzhou–Xining (Lanxi) urban cluster in China during recent decades poses a threat to the fragile arid environment. Quantitatively assessing the impact of urban expansion on vegetation in the Lanxi urban cluster has profound implications for future sustainable urban planning. This study investigated the urban expansion dynamics of the Lanxi urban cluster and its impacts on regional vegetation between 2001 and 2021 based on time series land cover data and auxiliary remote sensing data, such as digital elevation model (DEM) data, nighttime light data, and administrative boundary data. Thereinto, urban expansion dynamics were evaluated using the annual China Land Cover Dataset (CLCD, 2001–2021). Urban expansion impacts on regional vegetation were assessed via the Vegetation Disturbance Index (VDI), an index capable of quantitatively assessing the positive and negative impacts of urban expansion at the pixel level, which can be obtained by overlaying the Enhanced Vegetation Index (EVI) and rainfall data. The major findings indicate that: (1) Over the past two decades, the Lanxi region has experienced rapid urban expansion, with the built-up area expanding from 183.50 km² to 294.30 km², which is an average annual expansion rate of 2.39%. Notably, Lanzhou, Baiyin, and Xining dominated the expansion. (2) Urban expansion negatively affected approximately 53.50 km² of vegetation, while about 39.56 km² saw positive impacts. The negative effects were mainly due to the loss of cropland and grassland. Therefore, cities in drylands should balance urban development and vegetation conservation by strictly controlling cropland and grassland occupancy and promoting intelligent urban growth.

1. Introduction

Urban expansion refers to the transformation of nonurban land into urban land accompanied by disturbances to the regional environment [1,2]. Although urban expansion contributes to improving human well-being, it might also exert uncertain impacts on vegetation [3,4]. Changes in vegetation further affect regional ecological environments, particularly in arid areas, where they can lead to even more severe consequences. For example, when natural vegetation is replaced by urban development, the variety and quantity of vegetation decrease significantly. This leads to a phenomenon known as the urban heat island effect [5,6], where local temperatures rise higher than in surrounding rural or natural areas. Additionally, vegetation loss exacerbates urban flooding, as plants and trees play a crucial role in absorbing rainwater and stabilizing the soil. Without sufficient vegetation, there is less absorption, and more rainwater becomes surface runoff. This increased runoff can overwhelm drainage systems during heavy rainfall, leading to more frequent and severe flooding in urban areas [7]. Furthermore, the reduction in vegetation likely intensifies soil erosion in the surrounding areas, increasing the risk of landslides and other natural disasters [8,9]. These problems are more serious in the fragile environments of arid regions.

The Lanzhou–Xining (Lanxi) urban cluster in China is a classic arid region located in the western region of China; it serves as a crucial node in the “Belt and Road Initiative” and represents a typical arid region in China [10], therefore, the Lanxi region holds significant strategic importance. With the advancement of the “Belt and Road Initiative”, the economic and population scales of the Lanxi aera have further increased, leading to a notable rise in human activities and accelerating the pace of urban expansion. Due to the low precipitation and high evaporation rates, there is rapid moisture loss and low soil humidity in the Lanxi urban cluster. These environmental conditions are not conducive to forest growth. As a result, the Lanxi area is predominantly covered by grassland supporting livestock farming and ecological balance, while cropland is significant, and forest and shrubland occupy relatively small areas with low tree cover [11,12]. Urban expansion negatively affects vegetation by directly occupying it [13,14,15], whereas urban greening and improved city management have positive effects on vegetation [16,17]. However, there is still a lack of methods for accurately and quantitatively assessing both the positive and negative impacts of urban expansion on vegetation. Effectively assessing the impact of urban expansion on vegetation in arid areas is of paramount importance for promoting sustainable development in these urban regions [18,19,20].

Numerous studies have demonstrated that the impact of urbanization on arid regions can be quantified using various remote sensing indicators, such as the Normalized Difference Vegetation Index (NDVI) and the Enhanced Vegetation Index (EVI) [21,22,23,24,25]. For example, Sun et al. evaluated the impact of urbanization on vegetation cover in 117 major cities in China over the past three decades via the NDVI. The study revealed that urban expansion from 1982 to 2006 reduced vegetation greenness in arid regions of Northwest China, with a significant decrease in the average NDVI. While the NDVI is one of the most commonly used vegetation indices, its susceptibility to atmospheric and soil influences makes it more suitable for assessing vegetation in areas where atmospheric and soil conditions are relatively stable. Jenerette et al. conducted improved research using the EVI and reported that urban expansion in arid lands of the southern United States from 2000 to 2020 led to a decrease in regional vegetation coverage [26]. Compared with the NDVI, the EVI is more resistant to atmospheric interference and is more capable of reflecting vegetation structure and chlorophyll content [27]. In the aforementioned study, they proposed the direct impact of water flux on vegetation, but the EVI cannot be used to differentiate the effects of precipitation. Compared with the NDVI and EVI, the net primary productivity (NPP) has the advantage of directly reflecting vegetation productivity and the ability to monitor changes in vegetation productivity over the long term [28].

Pei et al. investigated the influence of urban expansion on arid land vegetation in China through NPP and reported a correlation between the growth of urban land area and the loss of vegetation NPP from 2000 to 2006 [29]. However, there is a clear correlation between NPP and precipitation, where NPP decreases with decreasing precipitation. Within the isothermal zone, NPP tends to increase with increasing precipitation levels [30,31]. Nevertheless, this conspicuous pattern evidently does not align with the focus of our study; our objective is to investigate the impact of urban expansion on vegetation while controlling for other factors, especially precipitation. While the NDVI, EVI, and NPP indices can reflect vegetation conditions [32,33], there are still general challenges when these indicators are used to assess the impact of urban expansion on arid land vegetation [34,35]. First, arid land vegetation is significantly influenced not only by urban expansion but also by precipitation, a crucial factor that the aforementioned indicators often neglect [36,37]; consequently, these methods may not accurately and quantitatively reflect the actual impact of urban expansion on arid land vegetation. Second, the majority of current studies predominantly focus on the adverse effects of urban expansion on vegetation [38]. Nevertheless, urban expansion can also positively impact vegetation through various measures, such as increasing urban green spaces. Hence, there is a pressing need for a more objective assessment metric to quantitatively determine the positive and negative effects of urban expansion on vegetation in arid regions [39,40].

The Vegetation Disturbance Index (VDI) represents a novel approach for quantitatively assessing the impact of urban expansion on arid land vegetation. Compared with traditional vegetation indices, the VDI offers greater adaptability to changes in precipitation levels. By incorporating precipitation factors, the VDI diminishes the influence of precipitation fluctuations on vegetation, thereby providing a more accurate reflection of the actual impact of urban expansion on vegetation, particularly highlighting its effects. Moreover, the VDI enables a more precise quantitative assessment of both positive and negative impacts on vegetation. While capitalizing on the advantages of vegetation indices, the introduction of the VDI enhances the understanding of the tangible effects of urban expansion on vegetation and provides a more accurate and comprehensive tool for analyzing spatial distribution changes in vegetation [41,42]. This facilitates a more comprehensive examination and response to the complexities of vegetation changes within the urbanization process [43,44,45].

This study aimed to reveal the dynamics of urban expansion and their impacts on the vegetation of the Lanxi urban cluster over the past two decades. Specifically, this research commenced by collecting land cover data spanning from 2001 to 2021, followed by meticulous preprocessing of the dataset. An examination of alterations in urbanized areas revealed the intricate dynamics of urban expansion over the course of approximately two decades. We subsequently amassed nearly two decades of Enhanced Vegetation Index (EVI) and precipitation data, subjecting them to rigorous preprocessing procedures encompassing cloud removal, interpolation, and resampling to ensure homogenous resolution. We performed overlay calculations on the data to obtain pixel-level VDI values. This facilitated a quantitative analysis of the holistic impact of urban expansion on vegetation within the Lanxi urban cluster throughout the 20-year timeframe. This research provides a comprehensive depiction of vegetation changes during the urbanization process in the Lanxi urban cluster. The findings of this research can serve as valuable references and guidance for future sustainable urban planning and ecological conservation in the region.

2. Data and Methods

2.1. Study Area

The Lanxi urban cluster is located at the westernmost end of the Yellow River Basin, with geographic coordinates ranging from 34°26′ N to 37°38′ N and from 98°55′ E to 105°55′ E [46] (Figure 1). The Lanxi urban cluster encompasses 39 counties and districts across 9 cities and prefectures, spanning an area of 97,500 km². The region features diverse and complex terrain, primarily characterized by mountainous and river valley landscapes, with elevations ranging from 1258 to 5255 m and an average elevation exceeding 2000 m. The Lanxi area is one of the areas in China with the least rainfall, with 58% of the total area receiving an average annual rainfall of less than 300 mm [47]. The Hexi Corridor, which has the lowest annual precipitation, has an annual rainfall of only 40–200 mm [48]. In 2021, grassland predominated as the primary vegetation cover type, accounting for more than 74% of the total area, followed by cultivated land and water bodies, accounting for 13.4% and 3.7% of the total area, respectively. Urban areas cover 0.3% of the total area, totaling 294.3 km², but they can have wide impacts on regional vegetation [49]. The population in the area reached 12.19 million, constituting 66.5% of the total permanent population of the two provinces. The eastern part of the Lanxi urban cluster is characterized primarily by a temperate semi-arid climate, whereas the western part is characterized by a continental plateau semi-arid climate. The average annual precipitation is approximately 300 mm. Severe land desertification, desertification, and soil erosion are prevalent in the region, highlighting the critical importance of strengthening ecological environmental protection for the sustainable development of urban clusters [50].

2.2. Data and Preprocessing

The land cover data for the Lanxi urban cluster in 2001 and 2021 were extracted from the 30 m annual land cover dataset (CLCD) provided by the Xin Huang team at Wuhan University [51]. The dataset, with a spatial resolution of 30 m and uses the World Geodetic System 1984 (WGS 1984) [52], achieves an average accuracy of 79.31%, which outperforms the Moderate Resolution Imaging Spectroradiometer (MODIS) land cover product (MCD12Q1) [53], the annual land-cover (LC) maps produced by the European Space Agency Climate Change Initiative (ESA-CCI) project [22] and Fine Resolution Observation and Monitoring of Global Land Cover (FROM_GLC) [54]. Compared with other Landsat-based thematic products, this dataset exhibits superior consistency in impermeable surface and global forest change delineation. This dataset consists of 9 land cover categories (cropland, forest, shrub, grassland, water, built-up area, barren, barren and wetland), which are widely employed in research on land cover type conversions. They play an indispensable role in land use studies in China [55,56,57]. The dataset was masked via the vector boundary of the Lanxi urban cluster in ArcGIS 10.8 software [58] to extract the urban land cover data for 2001 and 2021. The vegetation defined in this study includes cultivated land, forests, shrubs, and grasslands, which means that this study investigates the impact of urban expansion on the aforementioned land types.

The vector boundaries of the Lanxi urban cluster were obtained from the 1:1,000,000 National Basic Geographical Database, a component of the National Geospatial Information Resource Catalog Service System: https://www.webmap.cn/ (accessed on 20 April 2024).

The rainfall data from 2001 to 2021 were sourced from the Climate Hazards Group InfraRed Precipitation with Station data (CHIRPS) high-resolution global rainfall dataset: https://data.chc.ucsb.edu/products/CHIRPS-2.0/ (accessed on 20 April 2024). This global precipitation dataset was developed by the Climate Hazards Group at the University of California, Santa Barbara; it combines satellite infrared observations with ground station observations to provide high-resolution rainfall data worldwide. The CHIRPS dataset spans over 30 years and has a spatial resolution of 0.05 degrees, providing daily precipitation data. The data were processed by cropping and downloading via the Google Earth Engine platform [59].

The EVI data were sourced from the MODIS dataset released by the National Aeronautics and Space Administration (NASA): https://lpdaac.usgs.gov/data/ (accessed on 17 April 2024). The spatial resolution is 250 m, and the temporal resolution is 16 days. The data were extracted and downloaded via the Google Earth Engine. In ArcGIS, the EVI data were subsequently combined with the rainfall data to calculate the VDI.

2.3. Methods

This research adopts the VDI index to assess the influence of urban expansion on vegetation quantitatively. Initially, the land cover area of built-up regions in key cities within the urban cluster is computed. The dynamic expansion of built-up areas within the Lanxi urban cluster from 2001 to 2021 was subsequently quantified by analyzing impervious areas’ changes in and growth rates. This approach offers a tangible portrayal of urban transformations within the Lanxi urban cluster. With the use of GEE platform, we downloaded the maximum annual growing season EVI data from the MODIS database. Using this data, we performed pixel-level overlay calculations with rainfall data of the same resolution in ArcGIS software to obtain the pixel-level VDI data. By evaluating VDI values, this research discerns the varied impacts—both positive and negative—of urban expansion within the Lanxi urban cluster on vegetation. In addition, this study utilized a land use transfer matrix to investigate the evolution of different land cover types over the past two decades and analyzed the primary factors contributing to disturbances to vegetation. The specific process is shown in Figure 2.

The VDI index integrates the EVI with rainfall data, effectively detecting vegetation disturbances while eliminating the impact of rainfall variations. Ma et al. were the first to use this method to quantify the impact of surface coal mining on vegetation disturbances in semi-arid areas, demonstrating its effectiveness in excluding the influence of rainfall in arid regions [60]. In 2024, Qi et al. used VDI to study the dual impact of urban expansion on vegetation in the arid regions of northern China, and made certain suggestions for urban planning. The study found that urban expansion led to a reduction of 7736 km² in vegetation area but also promoted the growth of vegetation over 5011 km². Population growth and land use changes are the main factors of the negative impact, while economic growth is the main driving force of the positive impact [61]. Therefore, this method is highly reliable and applicable for evaluating the impact of urban expansion on vegetation in the Lanxi urban agglomeration. Thus, the research areas of this study encompass the changes in the VDI values of the urban expansion zones of the Lanxi area from 2001 to 2021. The impact of vegetation within the pixels was determined to be positive, negative, or neutral by comparing the magnitude of the VDI values.

It is worth mentioning that the vegetation referred to in this paper includes cropland, grassland, forests, and shrubs in the context of land use. The main impact on the vegetation in the urban expansion areas is attributed to the city’s expansion, such as environmental degradation caused by population growth and the increase in urban buildings at the expense of a reduction in vegetation. In arid regions, vegetation is extremely sensitive to the amount of rainfall; hence, this paper utilizes the VDI value as an indicator to eliminate the influence of rainfall.

2.3.1. Quantifying Urban Expansion Dynamics

This research adopts the quantitative method proposed by Huang et al. to measure urban areas [62], utilizing changes in the area and the urban land growth rate to quantitatively analyze the expansion dynamics of urban areas from 2001 to 2021. The calculation formula for changes in urban areas is as Equation (1):

$$Ur{b}_{\mathsf{\Delta }k}=A_{t_1}-A_{t_2},t_2>t_1$$

(1)

where $Ur{b}_{\mathsf{\Delta }k}$ represents the change in urban land from year $t_1$ to year $t_2$, and $A_{t_1}$ and $A_{t_2}$ denote the urban land areas in years $t_1$ and $t_2$, respectively. We employed this method to conduct a quantitative analysis of urban expansion and its area in the Lanxi urban cluster between the years 2001 and 2021, precisely determining the specific area of urban expansion over these two decades. Given the differences in the initial sizes of various cities, the sheer magnitude of the expansion area does not directly reflect the rate of expansion. Therefore, we used Equation (2) to calculate the expansion rate for each city:

$$R_{\left(t_1,t_2\right)}={\displaystyle \frac{A_{t_2}}{A_{t_1}}},t_2>t_1$$

(2)

where $R_{\left(t_1,t_2\right)}$ represents the ratio of urban land between year $t_2$ and year $t_1$, indicating the urban expansion rate.

This formula calculates the multiple of urban expansion by comparing the sizes of the city’s area before and after the expansion, intuitively reflecting the changes in the city’s scale. Subsequently, we used Equation (3) to further calculate the average annual expansion rate for each city:

$$Y{R}_{\left(t_{1,}{t}_2\right)}=\left({\left({\displaystyle \frac{A_{t_2}}{A_{t_1}}}\right)}^{{\displaystyle \frac{1}{t_2-t_1}}}-1\right)\times 100\%,t_2>t_1$$

(3)

The Urban Expansion Intensity Index (UEI) is calculated by dividing the expanded area by the product of the initial area and the time interval [63]. This method provides a measure of expansion intensity relative to the initial area and time used to assess the intensity of urban area growth over a specific period. The calculation method is Equation (4):

$$UEI={\displaystyle \frac{\left(A_{t_2}-A_{t_1}\right)}{\left(A_t\times \mathsf{\Delta }t\right)}},t_2>t_1$$

(4)

where $A_{t_1}$ and $A_{t_2}$ represent the built-up area between any two consecutive years, $A_t$ represents the total built-up area of the study area, and $\mathsf{\Delta }t$ represents the time interval between $t_1$ and $t_2$.

2.3.2. Calculation of the VDI

Prior to calculating the VDI, this study preprocessed the MODIS 16-day 250 m resolution EVI data and CHIRPS high-resolution global rainfall data. To ensure that the data were at the same 250 m resolution, considering that the spatial resolution of the rainfall data was 0.05°, ArcGIS spatial interpolation functionality was employed for data processing. Specifically, this study utilized the kriging interpolation method to interpolate the rainfall data to a resolution of 250 m. The purpose of this step is to maintain consistent spatial resolution in subsequent calculations, enabling comparison and analysis of $EV{I}_{max}$ and $Preci{p}_{cumlative}$ on the same scale. All data in the experiment were projected onto the Albers equal-area projection, and all data were resampled to a 250 m resolution.

This study employed the VDI developed by Ma et al. in 2018, which integrates MODIS EVI data with precipitation data to detect vegetation disturbances in the Lanxi urban cluster from 2001 to 2021 [60]. The calculation formula is Equation (5):

$$VDI={\displaystyle \frac{EV{I}_{max}/Preci{p}_{cumlative}}{{\overline{EVI}}_{max}/{\overline{Precip}}_{cumlative}}}$$

(5)

The VDI is a pixel-level vegetation visturbance index, where $EV{I}_{max}$ and $Preci{p}_{cumlative}$ denote the maximum EVI and cumulative precipitation, respectively, during the growing season (June to September) of 2021, and ${\overline{EVI}}_{max}$ and ${\overline{Precip}}_{cumlative}$ represent the multiyear averages of the maximum EVI and cumulative precipitation from 2001 to 2021. Notably, this study used EVI data from the growing season to represent annual vegetation activity because EVI data from the winter months may be inaccurate, potentially leading to false vegetation trends, in calculating the VDI, pixels with maximum EVI values less than 0.025 were removed because they typically represent land surfaces associated with water bodies and snow/ice [64]. When vegetation is disturbed, the $EV{I}_{max}$ value for that year significantly increases or decreases, resulting in a substantial deviation from the multiyear average; hence, the $EV{I}_{max}$ value was used to assess vegetation disturbance on the basis of this criterion.

2.3.3. Impacts of Urban Expansion on Vegetation

To assess the impact of urban expansion on vegetation, we used the findings from the previous section to extract the expansion areas of the Lanxi urban cluster and analyzed the VDI values of the pixels within these areas. The impacts of urban expansion on vegetation can be categorized into two types: positive impacts and negative impacts [65,66]. When vegetation is unaffected by urban expansion, the VDI values fluctuate within a certain range, known as the natural fluctuation range. However, when vegetation is affected by urban expansion, the VDI values deviate from this natural fluctuation range, and the larger the deviation from the natural fluctuation range is, the greater the impact [60]. Therefore, this study assesses the impact on vegetation on the basis of changes in VDI values. Specifically, when the VDI value exceeds the natural fluctuation range, it indicates that the vegetation in that area has experienced a positive impact, whereas if it falls below the range, it indicates a negative impact. This study refers to Ma et al. and employs the method of the mean standard deviation classification to determine the natural fluctuation range and classifies the pixels within the Lanxi urban cluster into three types on the basis of their impact [60]. The specific formula is shown in Equation (6):

$$Impac{t}_i=\left\{\begin{array}{l}P{I}_i=0,N{I}_i=1,VDI<\mu -s\\ P{I}_i=0,N{I}_i=0,\mu -s\le VDI\le \mu +s\\ P{I}_i=1,N{I}_i=0,VDI>\mu +s\end{array}\right.$$

(6)

where $Impac{t}_i$ represents the $i$th pixel affected by urban expansion, $P{I}_i=0,N{I}_i=1$ indicates that the pixel has experienced adverse effects, $P{I}_i=1,N{I}_i=0$ indicates that the pixel has experienced positive effects, and $P{I}_i=0,N{I}_i=0$ represents the VDI values within the natural fluctuation range, indicating that the area has not been affected. $\mu$ and $s$ denote the mean and standard deviation, respectively, of all the pixels within the Lanxi urban cluster.

Cui et al., building on Ma et al.’s approach, assessed the effects of urban expansion on vegetation by measuring the affected area [60,65]. We also used this method to evaluate such impacts. Various formulas are applied to compute different types of impacts. The formula for calculating the positive impact is presented in Equation (7).

$$T{A}_P=S\times {\displaystyle \sum _{i=1}^{n}P{I}_i}\times \left(Urba{n}_i^{2021}-Urba{n}_i^{2001}\right)$$

(7)

In the above formula, $T{A}_p$ represents the area of vegetation positively affected by urban expansion; $S$ represents the pixel area, totaling 0.0625 km²; $n$ represents the total number of pixels; and $P{I}_i$ represents the class value indicating whether the $i$th pixel is affected positively by vegetation. A class value of 1 for a pixel indicates that the vegetation in that pixel is positively affected, whereas a class value of 0 indicates no positive impact on the vegetation in that pixel. $Urba{n}_i^{2021}$ and $Urba{n}_i^{2001}$ indicate whether the $i$th pixel was urbanized in 2001 and 2021, respectively. If the pixel is within an urban area, its pixel value is 1; otherwise, it is 0.

Similarly, the formula for calculating the negative impact is as Equation (8):

$$T{A}_N=S\times {\displaystyle \sum _{i=1}^{n}N{I}_i}\times \left(Urba{n}_i^{2021}-Urba{n}_i^{2001}\right)$$

(8)

where $T{A}_N$ represents the area of vegetation positively affected by urban expansion, and $N{I}_i$ is the class value indicating whether the $i$th pixel is negatively affected by vegetation. A pixel with a class value of 1 indicates that the vegetation in that pixel is negatively affected, whereas a class value of 0 indicates no negative impact on the vegetation in that pixel.

3. Results

3.1. Urban Expansion Dynamics in the Lanxi Urban Cluster

3.1.1. Overview of Lanxi from 2001 to 2021

Over the past two decades, the major cities within the Lanxi urban cluster have undergone significant urban expansion. From 2001 to 2021, the built-up area increased from 183.50 km² to 294.30 km², a growth of 1.6 times. This growth trend illustrates the rapid development and sustained growth of the Lanxi urban cluster in the process of urbanization. The average annual growth rate of 2.39% indicates a relatively stable pace of urban land area expansion. Additionally, the expansion intensity index of 0.03 suggests that the rate of urban expansion relative to the total urban area is relatively low, indicating a gradual growth of urban expansion in proportion to the total urban area (Figure 3).

From the perspective of land use change, the Lanxi urban cluster experienced several significant trends in land cover types between 2001 and 2021. First, the cropland area decreased from 14,660.30 km² in 2001 to 13,272.50 km² in 2021. This change is closely linked to urbanization and land development, indicating a shift of agricultural land towards urban and other land types. Meanwhile, forest area showed an increasing trend, rising from 2813.32 km² to 3243.09 km², an increase of about 15.3%. This growth may be attributed to reforestation and natural restoration activities, highlighting the effectiveness of ecological protection policies.

Shrub area decreased from 1084.71 km² to 984.61 km², which may be related to adjustments in vegetation structure and pressures from land development. Grassland area saw a slight increase, from 73,833.90 km² to 74,126.35 km². This stable growth suggests that despite ongoing urban expansion and land development, the grassland ecosystem has been preserved to some extent. Additionally, the significant increase in water body area, from 3332.12 km² to 3716.10 km², may reflect changes in precipitation patterns due to climate change and improvements in water resource management. The reduction in snow/ice cover, from 3.11 km² to 1.96 km², indicates the impact of global warming on mountain ice cover. The barren land area increased from 3429.48 km² to 3700.02 km², likely due to intensified land development activities and natural erosion processes. (Figure 4,

Table 1. Comparison of the land cover situation from 2001 to 2021.

Type	2001 (km2)	2021 (km2)
Cropland	14,660.31	13,272.51
Forest	2813.32	3243.09
Shrub	1084.71	984.61
Grassland	73,833.91	74,126.35
Water	3332.12	3716.11
Snow/Ice	3.11	1.96
Barren	3429.48	3700.02
Urban	183.49	294.31
Wetland	0.59	2.09

).

The arrows in Figure 4 represent the conversion between different land types. For example, the arrow from yellow (cropland) to light green (grassland) indicates the area converted from cropland to grassland between 2001 and 2021. In the counterclockwise direction, arrows pointing between the same land types signify that there was no change in land type during this period. For instance, the arrow from yellow to yellow indicates that these areas of cropland remained unchanged from 2001 to 2021.

Further analysis of land type transitions reveals that the conversion of cropland to grassland was the most prominent, with 4889.49 km² of cropland turning into grassland. This reflects the outcomes of agricultural restructuring and ecological restoration efforts. Simultaneously, the conversion of grassland to cropland was also significant, reaching 6153.24 km², indicating that agricultural expansion and land development trends are still ongoing. The increase in forest area mainly resulted from the conversion of cropland and grassland, with 28.72 km² and 1115.32 km², respectively, being transformed into forest, underscoring the importance of afforestation policies and natural restoration.

Urbanization also played a key role in land transitions. Approximately 87.78 km² of cropland was converted into urban land, highlighting noticeable urban expansion. Additionally, the increase in water body area partly resulted from the conversion of cropland, forest, and grassland with 31.37 km², 0.51 km², and 44.45 km², respectively, indicating the impact of hydraulic infrastructure construction and urban development on water resource distribution. The changes in shrubland and barren land areas were relatively small but still exhibited certain land transition phenomena, such as 160.94 km² of forest converting to shrubland and 2077.5 km² of barren land mainly originating from grassland. Although the wetland area remained generally small, the transition from grassland and water bodies reflects subtle changes in regional biodiversity conservation.

3.1.2. Statistical Expansion Information of the Key Cities

During the study period, Lanzhou, the hub city within the Lanxi urban cluster, experienced the most significant urban expansion, with an area increase of 47.93 km². Despite the large expansion, Lanzhou’s expansion intensity index was relatively low at 0.02, indicating a stable growth pattern. Baiyin followed with an urban expansion of 25.02 km². Interestingly, Baiyin’s expansion intensity index was 0.05 higher than Lanzhou’s, suggesting more concentrated growth over a shorter period. Xining, starting with a smaller initial area, saw the most significant proportional growth, with its urban land area increasing by 2.84 times at an annual growth rate of 5.35%. The total expansion area for Xining was 9.01 km², and its expansion intensity index was notably higher at 0.09, making it the city with the fastest total area growth within the cluster.

The rapid development of these cities can be attributed to their roles as central cities within the Lanxi urban cluster. The development of the Lanxi region mainly relies on these cities, and the influx of population tends to prioritize these cities for work and settlement, leading to faster development and urban expansion.

From the expansion rate perspective, Haidong expanded by 2.59 times, with an annual growth rate of 2.59%. However, due to Haidong’s smaller urban land base, its total expansion area was only 7.39 km². Dingxi also experienced rapid development, with its urban land area increasing from 14.68 km² to 30.07 km² over the past twenty years, an increase of 15.38 km^2, and an annual growth rate of 3.64%. Multiple factors, including economic development, population growth, industrial distribution, and policy differences, influence the varying expansion rates among cities. Larger cities tend to have higher industrial integration and more policy preferences, leading to larger expansion in big cities and smaller expansion in smaller cities. We have summarized the specific data on urban expansion in

Table 2. Urban expansion dynamics in the Lanxi urban cluster from 2001 to 2021.

City	Urban Expansion Dynamics						Impact on Vegetation
	Urban Land Area (km2)		Expanded Area (km2)	UEI	Annual Growth Rate	Expansion Rate		Positive Impacts		Negative Impacts
	2001	2021						Area (km2)	Percentage	Area (km2)	Percentage
Lanzhou	121.80	169.73	47.93	0.02	1.67%	1.39		15.94	33.25%	25.38	52.94%
Baiyin	27.43	52.45	25.02	0.05	3.29%	1.91		11.75	46.96%	9.75	38.96%
Dingxi	14.68	30.07	15.38	0.05	3.65%	2.05		4.88	31.69%	6.44	41.84%
Linxia	11.72	20.41	8.69	0.04	2.81%	1.74		2.50	28.76%	5.56	63.99%
Xining	4.91	13.92	9.01	0.09	5.35%	2.84		2.56	28.43%	4.56	50.59%
Haidong	2.86	7.39	4.53	0.08	4.87%	2.59		1.94	42.75%	1.81	39.99%
Total	2184.40	2314.97	110.56	0.03	2.39%	1.60		39.57	35.71	53.5	48.28

Note: UEI refers to the Urban Expansion Intensity Index.

and Figure 5. Table 2 compiles the expansion information of major cities within the Lanxi urban cluster, while Figure 5 shows the urban expansion areas.

3.2. Expansion Impacts on Vegetation in the Lanxi Urban Cluster

During the study period, the expansion of built-up areas significantly negatively impacted vegetation in the Lanxi urban cluster, far outweighing any positive effects. Figure 5 represents the comparison between the positive and negative impact areas of major cities in the Lanxi urban cluster. Figure 5a,b show the regions and areas of positive impact from urban expansion in the Lanxi area, while Figure 5c,d represent the regions and areas of negative impact. From 2001 to 2021, the expansion of built-up areas resulted in a negative impact on vegetation covering 53.5 km², whereas a positive impact on vegetation covered 39.56 km². The negatively affected vegetation area was 1.35 times larger than the positively affected area.

The assessment of impacts on various cities within the Lanxi urban cluster reveals a clear trend where, with the exception of Baiyin City and Haidong City, the negative impacts outweigh the positive ones. This indicates that most urban expansions have resulted in a loss of vegetation cover exceeding the areas where vegetation has been enhanced or protected. In Lanzhou, the area of vegetation negatively affected by built-up area expansion was 25.38 km², whereas the area of positively affected vegetation was 15.94 km². The area of negatively affected vegetation was approximately 1.5 times greater than the positively affected area. A similar situation was observed in Xining, where the negatively affected vegetation area was 5.56 km², while the positively affected area was 2.59 km². In Linxia Hui Autonomous Prefecture, the negatively affected area due to urban expansion was much larger than the positively affected area, with the former covering a significant portion of the expansion area, while the latter was smaller. Negative impacts are mainly due to the extensive conversion of natural habitats into urban areas, often driven by rapid urbanization and prioritization of development over environmental conservation. The relatively smaller areas of positive impact indicate limited green space preservation and a lack of effective integrated planning in urban environments.

3.3. Impact of Urban Expansion on Vegetation Distribution Dynamics

During the study period, the negative impacts on vegetation caused by urban expansion far outweighed the positive impacts. Therefore, this research further analyzed the relationship between urban expansion and its impacts on vegetation from the perspective of land cover changes. On the basis of land cover data from 2001 to 2021, we conducted a detailed analysis at the pixel level to explore the changes in land cover types during these two decades. Furthermore, we generated a land cover type transition matrix for Lanxi (

Table 3. Land cover type transition table for the Lanxi urban cluster from 2001 to 2021 (km²).

	2001	Barren	Cropland	Forest	Grassland	Impervious	Shrub	Snow/Ice	Water	Wetland	Total (km2)
2021
Barren		2077.50	56.41	0.15	1534.25	5.71	0.05	3.44	6.17	-	3683.68
Cropland		46.47	8225.32	28.72	4889.49	26.08	0.42	-	31.37	0.04	13,247.91
Forest		0.14	80.84	1870.16	1115.32	0.01	160.94	-	0.51	-	3227.91
Grassland		1111.10	6153.24	765.40	65,289.73	44.43	608.53	0.26	44.45	0.22	74,017.36
Impervious		12.30	87.78	-	85.82	103.45	-	-	5.82	-	295.17
Shrub		-	1.51	137.01	535.68	-	315.25	-	-	-	989.45
Snow/Ice		1.21	-	0.02	0.34	-	-	0.18	-	-	1.75
Water		179.25	59.57	1.79	221.61	4.70	0.12	-	3239.12	0.10	3706.26
Wetland		-	0.27	0.02	1.97	-	-	-	0.04	0.32	2.63
Total		3427.96	14,664.95	2803.26	73,674.21	184.38	1085.31	3.87	3327.48	0.69	99,172.12

) and transition diagrams illustrating changes between various land cover types and built-up areas (Figure 6) to clearly demonstrate the changes in land cover types and the areas occupied by urban expansion during this period.

Figure 6 shows the area of other land types occupied by urban expansion in the Lanzhou-Xining urban cluster and the area of urban land converted to other land types from 2001 to 2021. The transformation of urban land in the Lanxi urban cluster to other types of land, although present, is relatively small in scale, especially in terms of ecological land restoration, which shows a lack of sufficiency. Specifically, the area of urban land converted to cropland was 26.08 km², which may be related to the reclamation of urban land or the reintroduction of agricultural functions. However, compared with the cropland occupied by urban expansion (87.78 km²), this area is still negligible, indicating that the loss of agricultural land during the urbanization process far exceeds its restoration.

Furthermore, the area of urban land converted to grassland was 44.43 km², which, to some extent, shows attention to greening and ecological restoration during the urbanization process, but its effect is still limited. More noteworthy is that the conversion of urban land to forests is only 0.01 km², water bodies were 4.7 km², and barren land was 5.71 km². These figures indicate that the restoration and increase of land types with high ecological value during the urban expansion process are very limited and almost negligible.

Table 3 represents the transfer quantities (km²) between different land use types in the Lanxi urban agglomeration. The rows represent the various land categories in 2001, while the columns indicate the land categories in 2021. The intersection of a row and a column represents the transfer quantity from a certain category in 2001 to another category in 2021, signifying the amount of a particular category that has transformed into another category by 2021. For instance, the value of 1115.32 in the fourth column and the third row under “Grassland” indicates that a total of 1115.3 km2 of Grassland had transformed into forest by 2021. It should be noted that “-” in the table signifies that no transfer occurred from one land category to another.

The urban expansion in the Lanxi urban cluster has encroached upon various land types, including cropland, grassland, water bodies, and barren land. Figure 7 illustrates the extent to which urban expansion has affected these land types in different regions. The data reveals significant variability in how different cities have occupied these land cover types during their expansion. Cropland and grassland are the most heavily impacted, indicating that agricultural land and grassland ecosystems have experienced substantial pressure. For instance, urban expansion in Lanzhou and Dingxi has particularly affected cropland, with 45.25 km² and 14.94 km² being converted, respectively. Similarly, the occupation of grassland in Baiyin and Lanzhou was notable, with 23.69 km² and 32.56 km² being affected. These figures suggest that agricultural land and grassland resources are under the greatest pressure from urbanization. These are also the main sources of the negative impacts of the Lanxi area.

Furthermore, while the occupation of water bodies and barren land is relatively smaller, it is still significant in certain cities. For example, urban expansion in Lanzhou has consumed 4.75 km² of water bodies and 4.19 km² of barren land, while in Baiyin, 6.56 km² of barren land has been occupied. This occupation may be related to urban growth into undeveloped areas or increased utilization of water resources. In summary, urban expansion in the region is primarily focused on the occupation of agricultural and grassland resources, but the impact on water bodies and barren land is also becoming increasingly evident.

Figure 7 reveals that the primary source of negative impact on vegetation in the Lanxi urban cluster comes from the loss of grasslands and cropland. In the analysis, we documented the trends in area changes for different land cover types, compared the areas of various types of land surfaces affected by urban expansion, and found that from 2001 to 2021, the urban expansion of the Lanxi urban cluster primarily encroached upon cropland and grassland within the cluster area. The primary land cover type in the Lanzhou urban cluster is grassland, followed by cropland. Although there was an increase in grassland area over the study period, the land types predominantly occupied by the expansion of built-up areas were grassland and cropland. The total area of cropland and grassland occupied by this expansion accounted for more than 90% of the negatively affected vegetation area. This extensive conversion underscores why the negative impact on vegetation has surpassed the positive effects. From 2001 to 2021, the urban expansion of the Lanxi urban cluster occupied an area of 85.72 km² of grassland and 87.78 km² of cropland. However, due to land type conversion, from 2001 to 2021, 44.43 km² and 26.08 km² of urban land were converted to grassland and cropland, respectively.

Among the main cities experiencing urban expansion, Lanzhou exhibited the largest expansion area, with the greatest area of cropland and grassland being encroached upon. Specifically, 45.25 km² of cropland and 32.56 km² of grassland were converted to construction land. The next was Baiyin, whose expansion occupied 13.56 km² of cropland and 23.69 km² of grassland. Dingxi in the Lanxi urban cluster also encroached upon a considerable amount of cropland and grassland, with 14.94 km² and 9 km², respectively.

4. Discussion

4.1. Differences in the Expansion of the Lanxi Urban Cluster

From a land use perspective, the expansion of the Lanzhou–Xining urban agglomeration reflects the typical characteristics of urbanization. During the study period, urban land use increased significantly, with this growth trend becoming more pronounced following the establishment of the Lanzhou–Xining urban cluster. Liu et al. found that from 2000 to 2015, the annual growth rate of regional urban land area reached 7.2%, exceeding the national average [67]. Nie et al. indicated in their study that the land in the Lanzhou-Xining urban agglomeration primarily transitioned from green ecological space and agricultural production space to urban construction land, rural living space, and industrial and mining construction land, which is consistent with our research [68]. Moreover, the increase in urban construction land was mainly concentrated in areas such as Lanzhou [68]. Rapid urbanization drove the transformation of land use and accelerated economic restructuring and social development in the region, positioning the Lanzhou–Xining urban cluster as a key engine of regional development. Yin et al.’s research further indicates that during the process of coordinated urban development within the Lanzhou–Xining urban cluster, the strength of connections between cities in terms of industry, transportation, and information networks has significantly increased [69]. This increasingly close urban linkage has injected strong momentum into the development of the Lanzhou–Xining urban cluster, forming a mutually beneficial regional development network. The improvement of transportation infrastructure, particularly the expansion of highways and railway networks, has facilitated the efficient flow of goods and human resources within the region and enhanced economic interaction and cooperation between cities. Additionally, the construction and upgrading of information networks have further promoted the sharing and exchange of information within the urban cluster, boosting the overall region’s competitiveness and sustainable development capabilities.

During the expansion of the Lanxi urban cluster, several cities, such as Lanzhou, Baiyin, and Xining, expanded significantly more than other cities did. Previous studies have shown that these cities were originally relatively developed, with relatively flat terrain and well-developed infrastructure [70,71,72]. As urban clusters develop rapidly, well-developed infrastructure makes life more convenient, attracting more people [73]. Yang et al. analyzed the cities in the Lanzhou–Xining region from the perspective of urban transportation networks and urban vitality, finding that high urban vitality is mainly concentrated in areas like Lanzhou and Xining [74]. Their investigation revealed that urban vitality is closely linked to urban transportation. Lanzhou, Baiyin, and Xining have relatively advanced transportation networks, including highways, railways, and air routes, which can support large-scale population movement. As the population begins to flow, the city’s economy is driven, and urban expansion follows. These cities also possess abundant educational and medical resources. High-level educational institutions and medical facilities attract more people to settle and work. Furthermore, Lanzhou, as the provincial capital, and Xining, as the gateway city to the Qinghai–Tibet Plateau, have political, economic, and cultural significance that brings more development opportunities and resource allocation. These factors collectively promote the rapid expansion of these cities. Fu et al. confirmed this viewpoint in their paper “Spatio-temporal Difference Analysis of Land Use Efficiency in Lanxi Urban Agglomeration Based on the SBM-Undesirable Model” [75].

In contrast, other cities in the Lanxi urban cluster, such as Linxia and Haidong, do not have these conditions. These cities are mostly mountainous, with rugged terrain and minimal transportation. This makes the flow of external populations relatively inconvenient. As a result, economic development is slower, and the degree of urban expansion is relatively small. Moreover, these cities’ infrastructure and public services are relatively insufficient, making attracting and retaining a large population difficult. Owing to terrain limitations, construction costs are high, and development space is limited, which further hinders the expansion speed of these cities.

Therefore, urban development within the Lanxi urban cluster has experienced significant imbalances. Central cities, such as Lanzhou, Baiyin, and Xining, expanded rapidly, whereas other cities have lagged behind. In the future development of the Lanxi urban agglomeration, policies should be formulated to promote the balanced distribution of resources within the urban cluster. This includes public services such as infrastructure, education, and healthcare, as well as investment and financial support. A regional collaborative development strategy should be established to encourage cooperation and exchange between central cities and surrounding cities, sharing the fruits of development and forming complementary industrial chains and economic structures. The unique advantages of different regions should be fully leveraged to develop the characteristic industries of each city. For the slower-developing mountainous cities within the Lanxi urban agglomeration, the first priority is to strengthen the construction of infrastructure, such as transportation and communication in mountainous areas, enhancing their connectivity with the outside world. Encouraging mountainous cities to actively establish cooperative relationships with other regions, sharing resources, and jointly developing the potential of mountainous areas can promote their economic development. At the same time, urban planning that adapts to the terrain should be formulated to avoid large-scale terrain modification and reduce damage to the natural environment. Addressing the imbalances in the development of the Lanxi urban agglomeration will promote coordinated and sustainable development of the entire region [76].

4.2. Negative Impacts Derived from Cropland and Grassland Loss

One of the primary factors contributing to the negative impact on vegetation within the Lanxi urban cluster is the extensive appropriation of cropland and grassland for urban expansion [77]. As cities continue to expand, there is a growing demand for more residential areas, commercial zones, and infrastructure, leading to significant land acquisition in surrounding areas. Croplands and grasslands are often the most accessible and economical types of land for urban development, as they are typically located in relatively flat and easily developable geographical areas. Consequently, these lands often become the primary choice for urban expansion. The research conducted by Tang et al. indicates that urban expansion has reduced vegetation coverage in the Lanzhou area, leading to the fragmentation of vegetation habitats [78]. Similarly, Yang et al., after studying the impact of urban expansion on vegetation across multiple urban clusters in China, found that the intensity of urbanization is negatively correlated with vegetation coverage; in areas of urban expansion, the NDVI index significantly declined, suggesting that vegetation in these zones often suffers more severe degradation [79]. The reduction and degradation of vegetation caused by urban expansion not only impacts the health of urban ecosystems but also has profound negative effects on regional biodiversity and environmental quality.

Urban expansion also presents opportunities for environmental improvement. Li et al., using examples from Urumqi, Lanzhou, Yinchuan, and Hohhot, studied the positive impact of urban expansion on forest, grassland, and other vegetation types through vegetation indices [80]. The study found that urban expansion in these areas extended the growing season of urban vegetation compared to natural vegetation areas, demonstrating the positive effects of urbanization on the vegetation growth cycle. While the direct occupation of land for construction during urban expansion has a noticeable impact on vegetation, implementing effective greening measures, such as converting non-vegetated land into parks or increasing urban green space, can help mitigate these negative effects. When combined with sound urban planning and management, these strategies can effectively advance environmental protection and achieve ongoing improvements. Lv et al. utilized the Remote Sensing Ecological Index (RSEI) to study changes in the ecological quality of the Lanxi urban agglomeration. They reported that over the past 20 years, the annual average RSEI value for the Lanxi urban agglomeration has been increasing at a rate of 0.0057 per year, indicating that the area’s vegetation coverage is increasing year by year [50]. Additionally, the degree of residents’ involvement in environmental protection and their awareness of conservation are crucial factors. Active participation in environmental activities and advocating for green lifestyles among residents contribute to strengthening the positive impact on the urban environment.

In the future, with continuous technological advancements, substantial development in technological innovation applications will also emerge as a significant pathway for improving the urban environment [81]. By integrating big data analysis, artificial intelligence, cloud computing, and other technologies, intelligent city management systems can be developed to optimize urban management and enhance resource utilization. For example, establishing systems such as intelligent traffic management to reduce traffic congestion and vehicle emissions, thereby decreasing reliance on fossil fuels and reducing greenhouse gas emissions and air pollution. Leverage the advantages of big data analysis and artificial intelligence to optimize urban planning, rationally allocate land resources, and minimize encroachment on cropland and grasslands. Moreover, the widespread utilization of renewable energy can decrease the reliance on traditional energy sources, reduce carbon emissions, and mitigate adverse environmental impacts [62]. Integrating technological innovation with resident engagement can pave the way for a more extensive scope of sustainable urban development and environmental protection.

4.3. Limitations and Implications of This Study

From the perspective of urbanized areas, this study examines the urban expansion of the Lanxi urban cluster and its impact on the environment. However, owing to the wide-ranging nature of urban expansion and ecological security concerns, coupled with the continual evolution of landscape types over time and the frequent occurrence of human activities, several aspects warrant further improvement and exploration. First, the raster imagery utilized in this study has a uniform sampling resolution of 250 m, limiting its ability to detect urban expansion effectively on a larger scale; consequently, in future studies, higher-resolution imagery could be employed to analyze the Lanxi urban cluster more comprehensively. Moreover, the vegetation examined in this study is confined to the area of urban expansion, without considering the vegetation surrounding the built-up areas or at greater distances. To delve deeper into vegetation indirectly affected by urban expansion, future research could incorporate buffer zone analysis or gradient analysis within the study area.

Currently, China’s arid regions are experiencing rapid urbanization, which will continue in the future. He [82] et al. found that compared with the urban land area of 21,000 km² in 2015, by 2050, the urban land in the drylands of northern China will further expand to 31,000 km². However, the ecological environment of drylands is fragile and vulnerable [83]. Such rapid and large-scale urban expansion has led to a series of ecological issues, such as the degradation of natural habitats and the loss of biodiversity, having significant adverse consequences on the well-being and sustainable development of the region [81]. The extensive urban expansion in drylands in the future will also have a severe impact on regional vegetation, threatening the sustainable development of cities in drylands.

Therefore, we suggest considering the loss of cropland and grassland caused by urban expansion in drylands to mitigate the negative impacts on vegetation. On the one hand, improving urban land use efficiency and developing high-density cities is imperative, thus encouraging smart urban growth and reducing urban sprawl [62]. On the other hand, rational urban development planning should be adopted to avoid regions with concentrated biodiversity and reduce the encroachment on cropland and grassland [81]. In addition, it is also necessary to expand urban green spaces through the construction of parks, forests, green roofs, lawns, and community gardens. Such measures can increase the positive impacts on vegetation and further offset the negative impacts [17].

5. Conclusions

Over the study period, the Lanxi urban cluster underwent a significant expansion of its built-up areas. The built-up area increased from 183.5 km² in 2001 to 294.3 km² in 2021. The primary driving force behind this expansion of built-up areas was the major cities within the Lanxi urban cluster, with Lanzhou serving as the representative “main force”. From the perspective of the VDI, the expansion of built-up areas in the Lanxi urban cluster has a significant impact on vegetation. In most cities, the negative impact of built-up areas outweighs the positive impact, with only a few cities experiencing a positive impact greater than the negative impact. Overall, the expansion of built-up areas in the Lanxi urban cluster has a far greater negative impact on surrounding vegetation than positive impact.

Over the past 20 years, urban expansion in the Lanxi urban cluster has resulted in significant negative impacts on local vegetation, primarily manifested in the extensive occupation of grasslands and farmlands, which are crucial vegetation types. A reduction in such vegetation may have adverse effects on the regional ecological environment, posing threats to local biodiversity and soil quality. However, urban expansion has also generated positive ecological impacts by enhancing urban greenery construction. Therefore, in planning and managing the development of the Lanxi urban cluster, it is imperative to carefully balance the benefits and drawbacks of urban expansion and formulate more scientifically rational urban development strategies. It is essential for governments, enterprises, and the public to work together to facilitate the harmonious integration of economic and social development with ecological environment protection in the future development of the Lanxi urban cluster.

While this study has made some progress in understanding urban expansion and its impact on the environment in the Lanxi urban cluster, further exploration and refinement are still needed. It is hoped that future research can make more profound and comprehensive contributions in this field, providing more scientific decision-making support for the sustainable development of the Lanxi urban cluster and similar regions.