Differentiation of Vegetation Community Characteristics by Altitude within Urban Parks and Their Service Functions in a Semi-Arid Mountain Valley: A Case Study of Lanzhou City

Abstract

As the primary component of urban green space, parks play an important role in improving local environments and quality of life for city inhabitants. To date, much research on urban park plant community characteristics has been conducted worldwide, but studies focusing on differences across altitudes have been rare. Here, we have investigated fractional vegetation cover (FVC), plant species diversity, plant species composition, and plant life-form types along altitude gradients within four urban parks (Jincheng, Renshoushan, Baitashan, and Lanshan) in Lanzhou city (China) by field sampling and remote sensing. Spatial variations in plant community characteristics at different elevations, as well as their ecological, social, and integrated service levels, were analyzed using combined methods of remote sensing interpretation, mathematical statistics, and service function assessment. The results showed that: (1) spatial variations in plant communities along elevation gradients exhibited a distinct regional differentiation between parks; (2) Lanshan Park had the lowest FVC (43.6%) and plant species diversity (59 species), and these two parameters exhibited the most dramatic spatial change as elevation increased; the opposite phenomenon occurred in Renshoushan Park (FVC of 56.6%; 81 species); (3) the differences in the natural service level of the plant community were smaller for the social and integrated service levels. The highest natural service level occurred in Renshoushan Park, while Jincheng Park had the highest social and integrated service levels; the smallest values occurred in Lanshan Park. We conclude that scientific park management is urgently required to ensure the safety and sustainable development of the local urban ecosystems and to improve the ecological and social service levels of the individual urban parks, especially Lanshan Park. These findings also provide key insights to improve urban park communities and service levels in similar mountainous, semi-arid zones.

1. Introduction

Urbanization is a global phenomenon, and it is predicted that urban areas will account for 60% of land area in 2030 [1]. China, the largest developing country in the world, is also experiencing a rapid urbanization process [2]. Since the second half of the 20th century, especially the 1990s, urbanization in China has not only promoted rapid urban sprawl and socioeconomic development but also exacerbated several urban problems, such as environmental pollution, deterioration of the human environment, and ecological imbalance [3,4]. Sacrificing environment for economic development is also a common phenomenon in developing countries [5,6]. Previous studies report that issues caused by urbanization also have profound and lasting effects on public health, in particular, increased morbidity and mortality in urban areas [3,7]. To meet the needs of its citizens and improve the environmental quality of large population centers, China adopted a governmental framework called ecological civilization construction to ameliorate unbalanced economy–environment urban relationships at the 17th National Congress of the Communist Party of China (CPC) in 2007 [8]. Under its guidance, the local (prefectural) government put forward urban sustainable development goals (SDGs) for the construction of a well-balanced ecological city [9]. Urban parks, as the primary component of green space in towns and cities, play an important role in the construction of such a city.

Urban parks serve the purpose of biodiversity protection, ecological restoration (air purification, microclimate regulation), citizen leisure and recreation, and other services (noise filtration, rainwater regulation, wastewater treatment, etc.) [10,11,12]. Their functions in providing natural and social services are therefore somewhat different from those of traditional natural parks and forests [13]. To promote our understanding of how these service functions can best be performed, much research has been carried out into optimal construction, development, and layout of urban parks [14,15,16]. These studies have evaluated various theories (ecology, landscape ecology, sponge city, etc.) [13,17] and technologies (compounded layers of plant community, ecological revetment, wetland wastewater treatment, ecological regulation, canopy interception, and water conservation) [18], so a relatively mature and robust theoretical and technological framework underpinning urban park construction exists. To improve the ecological service function of these parks, researchers have conducted numerous studies on floral community characteristics, species diversity of woody plants, and ecological function evaluation [19,20,21]. However, large- and small-scale altitude differences in the urban park plant community characteristics have not been assessed; as land space becomes less available, optimizing plant communities at different elevations provides an untapped opportunity for expansion and development.

Lanzhou is a typical semi-arid valley city in northwestern China. In recent decades, urbanization in Lanzhou has achieved considerable development forced by the Great Western Development Strategy (GWDS) [22]. The topography of the valley basin limits the further expansion of the city space, so creating a balance between continued urban development and enhanced environmental quality/ecological protection in this semi-arid climate is an ongoing challenge [23,24]. To ameliorate this dilemma, citizens in Lanzhou have carried out a large-scale reforestation campaign in the north and south mountains since 1954 with the support of a series of projects such as the greening of barren mountains, construction of economic forest, construction planning of urban tourist areas, and so on [25,26]; the barren mountains of the past have been replaced with green vegetation, an effort hailed as “the two lungs of Lanzhou” by many scholars [27]. More importantly, several urban parks have been built in the mountains on both sides of the city, which provide opportunities for city dwellers to build connections with their natural environment, shown to be beneficial to health and wellbeing [28]. Nevertheless, the existing studies show that although these greening projects have greatly promoted the expansion of urban green space and the improvement of urban environmental quality, partial plant community especially cultivated plant community exhibits a certain degree of degradation under the influence of multiple factors, such as regional climate, human disturbance, unreasonable management, and so on [24,27,29]. Therefore, it is of great significance for ensuring sustainable development of the ecosystem and its service functions to conduct regular ecological monitoring.

In this study, we have investigated common vegetation parameters (species diversity, species composition, fractional vegetation cover, and life-form) along altitude gradients of the four urban parks (Lanshan, Baitashan, Renshoushan, and Jincheng) in Lanzhou. The study objectives were to characterize how vegetation indices varied with increased elevation, evaluate differences in plant communities between the four urban parks, and explore the ecological quality based on ecosystem services. Our results provide new insights for urban planners and managers to exploit vertical, as well as horizontal, differences in park ecologies to promote sustainable development of ecosystem health and ecological service functions in parks of semi-arid valley cities.

2. Materials and Methods

2.1. Study Areas

Lanzhou, the capital of Gansu Province, is located in inland, northwestern China (35°53′18″–36°33′56″ N, 103°21′04″–104°00′38″ E) [24,30] (Figure 1). The city adopts an east-west orientation bounded by mountains to the south and north (~35 km × 2–8 km) [31], and the Yellow River flows eastwards through the city. Geologically, Lanzhou lies in the semi-arid loess hilly-gully region in the upper reaches of the Yellow River [26]; climatically, it is located in the transition between monsoon and non-monsoon zones, and is regulated by a moderate temperate continental climate [21]; and the vegetation type is mainly desert steppe. The annual precipitation and average temperature are 327.7 mm and 9.1 °C, respectively, with reasonable sunlight (annual average 2607.6 h) and a long frost-free period (more than 185 days per year) [25]. These topographical and climatic conditions encourage serious soil erosion in steep areas and create a fragile ecological environment [32].

Lanshan, Baitashan, Renshoushan, and Jincheng Parks are representative of several urban parks in Lanzhou city (Figure 1). They are evenly distributed in the southern and northern mountains of the three urban districts including Chengguan, Anning, and Xigu. Among them, Lanshan Park is the largest (~3.47 km²), whereas Jincheng Park is the smallest (~0.5 km²). The parks were built from 1958 to 1987, so their levels of maturity vary. The dominant native trees in the parks are Sophora japonica, Ulmus pumila, Salix matsudana, Elaeagnus angustifolia, Ailanthus altissima, and Platycladus orientalis [25].

2.2. Remote Sensing Data

2.2.1. Landsat Data

This study employed Landsat images to assess variations in fractional vegetation cover (FVC) at pixel scale and across elevation. Based on the data quality and the time of field investigation, we obtained a Landsat 8 image in 2020 from the geospatial data cloud (http://www.gscloud.cn/; accessed on 29 July 2022). The Landsat image has a high quality (cloud-free) and 30 m spatial resolution (

Table 1. Basic information on remote sensing data.

Data Type	Spatial Resolution	Data Quality	Time of Data Acquisition	Data Source
Landsat 8	30 m	High (Cloud-free)	26 July 2020	Geospatial data cloud
Digital elevation model	30 m	High	5 September 2019

), which fully covers the whole Lanzhou city. Before the interpretation of the remote sensing data, we preprocessed remote-sensing images through geometric rectification, atmospheric correction, radiometric calibration, and the transformation of color composites in RGB (Red, Green and Blue) by using ENVI 5.1 software (https://www.harrisgeospatial.com/SoftwareTechnology/ENVI-working; accessed on 23 September 2022).

2.2.2. Digital Elevation Model

To analyze the change patterns of FVC across elevation, we collected the digital elevation model (DEM) from the geospatial data cloud (http://www.gscloud.cn/; accessed on 8 May 2022). The DEM data (30 m spatial resolution) was clipped based on the boundaries of the four urban parks by utilizing ArcGIS 10.2 software (ESRI Inc., Redlands, CA, USA) suitable for the later statistical analysis of FVC at elevational gradients.

2.3. Method

2.3.1. Field Survey

There were some differences in plant community types among the four selected urban parks due to the degree of human intervention in the plant communities. Specifically, Jincheng Park mainly exhibited a cultivated plant community, whereas the other urban parks (Renshoushan, Baitashan, and Lanshan) were mainly covered by natural, semi-natural and cultivated plant communities. During 15–24 May 2022, we selected 28 sampling sites at different altitudes in the four parks (Figure 1) and investigated plant species and dominant plant species. Sites were selected to include almost all vegetation community types to reflect vertical differences in community characteristics. The altitude interval between adjacent sampling sites was chosen in accordance with the topography of the study areas; specifically, the Lanshan Park with steep terrain was set at approximately 50 m above sea level (above sea level) sampling interval, whereas the other parks with relatively flat terrains were set at approximately 25m sampling intervals. In each park, quadrats of 10 m × 10 m were designed as evenly as possible across the elevations, and all plant species within the quadrat were identified and recorded (157 plant species across the 28 quadrats).

2.3.2. Calculation of Vegetation Indices

Four plant parameters (fractional vegetation cover, species richness, species composition, and life-form) were selected to assess the vertical changes in plant community characteristics. FVC, the ratio of the vertical projected area of vegetation to the total ground area, is an important index in the assessment of the ecosystem balance [33] and was obtained from interpreting Landsat remote sensing images. The interpretation process involved extracting the normalized difference vegetation index (NDVI) and calculating FVC based on the NDVI data. NDVI is a common parameter used to estimate FVC by the pixel dichotomy model [34,35] extracted here by ENVI 5.1software. The Equations (1) and (2) of the NDVI and FVC are shown below:

$$NDVI=\frac{NIR-RED}{NIR+RED}$$

(1)

$$\mathrm{FVC}=\left\{\begin{array}{l}0, \mathit{when} \mathit{NDVI}\le {\mathit{NDVI}}_{\mathit{soil}}\\ \frac{NDVI-NDV{I}_{\mathit{soil}}}{NDV{I}_{\mathit{vegetation}}-NDV{I}_{\mathit{soil}}}{, \mathit{when} \mathit{NDVI}}_{soil}{<\mathit{NDVI}<\mathit{NDVI}}_{\mathit{vegetation}}\\ 1, \mathit{when} \mathit{NDVI}\ge {\mathit{NDVI}}_{\mathit{vegetation}}\end{array}\right.$$

(2)

where NIR and RED represent the reflectance in the near-infrared and red bands, respectively; and NDVI_soil and NDVI_vegetation are minimum NDVI values for pure soil and pure vegetation pixels, respectively.

Species richness (N) is a common index that reflects species diversity in a community or biotope, calculated using Equation (3) [36]:

$$N=Species number in unit area$$

(3)

where N is the total number of species occurring at each sampling site.

Plant species composition is used to characterize changes between plant community types and structures, in this study, which was calculated by the Jaccard similarity coefficient (see Section 2.3). Life-form is a taxonomic unit of plant ecology (one of four elements determining plant physiognomy), here divided into herbs, shrubs, and trees.

2.3.3. Calculation of Ecological Indices

The natural service level (NSL) of the plant community were assessed based on FVC, the plant species diversity, and the health of the plant species, whereas the social service level (SSL) was evaluated using FVC, the exotic plant species diversity, the landscape aesthetics, and the accessibility of the parks. The integrated service level (ISL) combines both ecological and social services. Calculations were as follows:

$$NSL=\frac{{\displaystyle \sum _{k=1}^3{v}_{\mathit{km}}}}{3}$$

(4)

$$SSL=\frac{{\displaystyle \sum _{l=1}^4{v}_{\mathit{lm}}}}{4}$$

(5)

$$ISL=\frac{NSL+SSL}{2}$$

(6)

where k and l represent the variable number of the NSL and SSL evaluations, respectively; m represents the four urban parks; the v_km and v_lm represent the variable values (values from 0–1) of the NSL and SSL evaluations, respectively, these variable parameters were described above.

2.4. Statistical Analyses

Based on DEM data and the calculated FVC, we applied ArcGIS 10.2 software to compute the average FVC per 25 m altitude interval within the four parks. To assess the spatial changes in the FVC and species diversity across elevation, linear and nonlinear regression models were used to analyze the relationships of FVC and species diversity with altitude. A Jaccard similarity coefficient was used to assess the similarity in plant species composition across elevation within and between each park [37]:

$$\mathit{Jaccard} \mathit{similarity} \mathit{coefficient}=\frac{C_{\mathit{ij}}}{A_i+B_j-C_{\mathit{ij}}}\times 100\%$$

(7)

where A_i and B_j are the total number of plant species in elevational intervals i and j (or in parks i and j); and C_ij is the number of plant species conserved across elevational intervals/parks i and j.

The regression models and Jaccard similarity coefficient analyses were performed using Origin software (Originlab, Co., Northampton, MA, USA).

3. Results

3.1. Fractional Vegetation Cover (FVC) Change across Altitude

FVC was assessed at the pixel scale and across elevation (Figure 2 and Figure 3). At the pixel scale, although the FVC of all parks varied in the range of 0–100%, the values exhibited a remarkable spatial difference across the four parks. The highest FVC in Jincheng Park occurred in the north and edges; in Renshoushan Park, in the middle zone from northeast to southwest; in Baitashan Park, in the middle and south; and in Lanshan Park, in the northwestern and southwestern edges. The lowest FVC appeared in the middle; the south and northwest; the northwest and the margin; and the northeast and mid-south of the four parks, respectively (Figure 2). Overall, the average park FVC ranged from 44 to 57%, with the highest and lowest values occurring in Renshoushan and Jincheng Parks, respectively (Figure 3).

The FVC of the surveyed parks also exhibited variations across elevation, in discrete patterns for each park. The FVC by altitude followed “U-shaped” (Jincheng Park), linear (Renshoushan Park), inverted “U-shaped” (Baitashan Park), and “N-shaped” (Lanshan Park) patterns. The highest FVC values also occurred at different altitudes in each park: at low altitude in Jincheng (1582–1600 m above sea level; FVC 59%); at mid-to-high altitude (1651–1675 m above sea level) in Renshoushan (FVC 67%) and Baitashan (FVC 66%); and at high altitude in Lanshan (2101–2125 m above sea level; FVC 79%); whereas the lowest values were recorded in the mid-to-high altitude in Jincheng (1651–1675 m above sea level; FVC 31%); at mid-to-low altitude in Renshoushan (1601–1625 m above sea level; FVC 49%); and at low altitude (1536–1550 m above sea level) in Baitashan (FVC 18%) and Lanshan (1551–1575 m above sea level; FVC 11%; Figure 3). Furthermore, based on the calculated FVC at the three elevational intervals (1601–1625, 1626–1650, and 1651–1675 m asl), the FVC in Lanshan and Baitashan first increased and then decreased with altitude; gradually decreased in Jincheng; and gradually increased in Renshoushan Park. The maximum and minimum FVC in these three elevational intervals were 60% (Baitashan) and 31% (Lanshan); 66% (Baitashan) and 34% (Lanshan); and 67% (Renshoushan) and 31% (Jincheng), respectively (Figure 3).

3.2. Plant Species Diversity Change across Altitude

The 157 recorded species in the four parks belong to 122 genera and 58 families. Renshoushan Park had the highest species diversity (81 plant species recorded at nine sampling sites), while Lanshan Park had the lowest species diversity (59 plant species recorded at ten sampling sites). Plant species diversity showed different regional responses to elevation. The species richness in Jincheng and Renshoushan parks “U-shaped” pattern, with an inverted “U-shape” in Baitashan, and an “N-shaped” pattern in Lanshan as altitude increased. Although the index in the Jincheng and Renshoushan Parks exhibited a similar pattern with elevation, the former displayed a more dramatic changing process. The highest species diversity of these two parks occurred in the mid-to-low altitude areas and (43 in Jincheng; 25 in Renshoushan), whereas their lowest species diversity appeared in the mid-to-high altitude areas (17 and 8, respectively). Interestingly, species diversity in Baitashan Park followed the opposite pattern, with its highest diversity in the mid-to-high altitude area (29 plant species at sampling site 3), and the lowest diversity in the mid-altitude area (19 plant species at sampling site 4). Lanshan Park exhibited a different pattern again, with both maximum (24) and minimum (8) species diversity present at sampling sites in the mid-to-high altitude area (Figure 4).

3.3. Plant Species Composition Changes by Altitude

A total of 73 plant species were recorded at four sampling sites in Jincheng Park; 81 species at nine sites in Renshoushan Park; 70 species at five sites in Baitashan Park; and 59 species at ten sites in Lanshan Park. The Jaccard similarity coefficients between different sampling sites in each park ranged from 0.10–0.23 at Jincheng; 0.02–0.45 at Renshoushan; 0.12–0.39 at Baitashan; and 0.06–0.53 at Lanshan Parks, indicating that plant species composition in Jincheng Park showed the most change across altitudes, while Lanshan Park exhibited the highest similarity between altitudes. The highest similarity coefficient (0.53) occurred between sampling sites 8 and 10 in Lanshan Park.

In Jincheng Park, the largest similarity occurred between non-contiguous sampling sites (1 and 4), while the most difference was observed between sites 1 and 3. The dominant species of each sampling site gradually declined as altitude increased. Rosa rugosa (Rosa) was the most common species across all sampling sites. In Renshoushan Park, the similarity coefficients between different sampling sites were highly variable, except for those between sampling site 6 and other sampling sites. The highest similarity was between sites 5 and 8, and the greatest difference was between sites 1 and 9. The types and number of the dominant species among the nine sampling sites were similar except for those at sampling sites 1–2. Ulmus pumila (Elm) and Platycladus orientalis (Oriental arborvitae) were the most common species across the nine sampling sites. In Baitashan Park, species composition differed considerably across elevations, such that the similarity coefficient was very low except between sampling sites 3 and 5. Similar to Renshoushan Park, Baitashan Park also exhibited relatively uniform types and numbers of the dominant species along altitude gradients. In Lanshan Park, although the similarity coefficient of species between most sampling sites was still <0.5, the species composition exhibited a regular variation, i.e., the index between the adjacent sampling sites was relatively high. The dominant species of this park showed an obvious spatial difference as altitude increased (Figure 5 and Figure 6).

The similarity coefficient of species between the four parks was also very low, ranging from 0.22–0.3. Among them, Jincheng and Baitashan Parks had the highest similarity, while the lowest species similarity occurred between the Jincheng and the Lanshan Parks (Figure 5).

3.4. Plant Life-Form Change by Altitude

Herbs, shrubs, and trees were all observed at all sampling sites, but the distribution of plant species among these life-forms varied by altitude. Trees exhibited a highly variable pattern of response to altitude. At Jincheng Park, they displayed a U-shaped relationship with altitude, a general decline at Renshoushan and Baitashan Parks, and an increase in occurrence at Lanshan Park at higher elevations. They were generally the most prevalent life-form at Jincheng and Baitashan Parks, while at Renshoushan Park, trees were generally the second most common life-form. Shrubs also showed a different response to altitude at the four parks, with a U-shaped pattern at Jincheng Park, general increases in Renshoushan and Lanshan Parks, and a general decrease in Baitashan Park at higher altitudes. Herbs behaved the most consistently, with a general decline in predominance at three parks as elevation increased, and an inverted U-shape relationship with altitude at Baitashan Park. Herbs were generally the most prevalent at sampling sites in Lanshan Park (Figure 7).

3.5. Ecological Service Function of the Current Plant Community

The four urban parks comprised natural, semi-natural, and cultivated plant communities, with very different distributions, especially in the Lanshan and Jincheng Parks in the southern mountain. Specifically, the cultivated plant communities within all parks were present at lower elevations except in Lanshan Park; while Jincheng Park contained only very mature (>30 years old) and very young (<10 years old) cultivated plant communities. Furthermore, the natural and semi-natural plant communities in the Renshoushan and Baitashan Parks were mainly distributed in the mid-to-high altitude areas, while those in the Lanshan Park occurred mainly in the lower regions.

From the perspective of the ecological service function of the plant community within each park, the NSL and SSL of the plant community showed a regional difference among the four urban parks. The biggest values of these parameters occurred in Renshoushan and Jincheng Parks, respectively, whereas the smallest values existed in Lanshan Park. Overall, Jincheng and Lanshan Parks had the highest and lowest ISL of the plant communities, respectively (Figure 8).

4. Discussion

4.1. Spatial Difference in Plant Community Characteristics Follow Altitude Gradients

Urban parks are the primary component of green space in cities, and play important roles in improving urban environments, protecting biodiversity, and providing recreational services for local residents [38,39,40]. Vegetation community characteristics reflect urban ecosystem health and the quality of ecological services [13,28]. Here, we assessed spatial variations in plant community characteristics along altitude gradients in Jincheng, Renshoushan, Baitashan, and Lanshan Parks in Lanzhou city. The results showed that the plant community characteristics in these parks exhibited considerable spatial heterogeneity, we speculated that this phenomenon was potentially controlled by the differences in the level of human intervention, park age, irrigation conditions, topographic conditions, and so on [22,35,36,41] in situ investigation. These findings were not entirely consistent with previous studies; for example, Zhao et al. (2018) reported that Lanzhou urban parks had made great progress in the g)reen area, accounting for 34.5% of the total urban green space [42], whereas the community characteristics were mainly controlled by designer aesthetics, the choice of practical organization, and citizen behavior [20].

Of the four surveyed parks, Lanshan and Jincheng Parks, located in the South Mountain, exhibited the most heterogeneous community characteristics. Lanshan Park had the steepest terrain (1535–2125 m above sea level), with plant growth (except at the highest altitude) rarely affected by human intervention such as exotic plant cultivation and artificial irrigation. These physical characteristics have shaped the current plant community characteristics, i.e., the introduction of cultivated vegetation at the top of the mountain ensured the highest species diversity and the largest FVC, while the natural and semi-natural plant communities in other areas had relatively low species diversity and FVC (Figure 3 and Figure 4). In general, the lowest species diversity and FVC occurred in the areas of the steepest terrain and no irrigation conditions (Sampling site 7).

Conversely, Jincheng Park had the flattest terrain and smallest area and was mainly covered by cultivated vegetation. The species diversity and FVC exhibited similar variation trends, namely, first decreasing then increasing with altitude, and likely reflected ongoing park expansion in the mid-to-high altitude area that was currently under development (Figure 3 and Figure 4). Park architecture has also greatly limited vegetation spatial distribution, further resulting in the decrease in the FVC.

The comparison between Renshoushan and Baitashan Parks in the North Mountain yielded interesting results: while these parks occupy similar altitude spans, plant community types, and proportions of natural/semi-natural/cultivated landscapes, there was a great difference in the plant community characteristics (FVC and species diversity) following elevation gradients (Figure 3 and Figure 4). Both parks were covered by more dense vegetation and more abundant species diversity than Jincheng Park. These discrepancies could be explained by differences in the park age, irrigation levels, and the degree of participation of the park managers [38] in situ observation, in line with the previous studies [28,38,43,44]. Moreover, changes in governmental policies at all levels (national and local governments) over different periods also provide an important contribution to differences in plant community characteristics between the four surveyed urban parks. For example, the Jincheng Park extension during 2014–2019 was mainly planted using exotic ornamental species under guidance for urban tourist areas, whereas native tree species have been the focus in urban parks to support ecological and environmental protection and restoration programs since 1998 [26].

4.2. Ecological Service Function of the Vegetation Community in Four Urban Parks

The ecological service function of urban parks mainly includes natural and social services [13]. The former maintains the healthy function of natural ecological processes and the urban ecological environment, whereas the latter improves citizens’ quality of life through landscape management to provide leisure and entertainment facilities [11,13]. In the four surveyed parks, Renshoushan and Baitashan Parks had a relatively healthy and stable vegetation community, and the community types had completed the transition from cultivated to semi-natural and natural vegetation communities, except for the cultivated vegetation at the foot of the northern mountain. In Lanshan Park, the vegetation community in most areas also exhibited natural and semi-natural characteristics, and some gradients were even covered with fragile local vegetation. In comparison, the reconstruction of Jincheng Park since 2013 has greatly altered the previous vegetation community characteristics in the mid-to-high altitude area, where a new vegetation community was forming, resulting in the lowest current natural service value. However, its social service level was the highest, due to its rich landscape plant diversity, even though Renshoushan Park had the highest species diversity. In contrast, the mid-to-low altitude area of Lanshan Park mainly comprised local vegetation, so this park had the lowest social service value.

In general, the four parks tremendously improved the urban ecological environment and expanded outdoor recreation and entertainment space for local inhabitants. Renshoushan and Baitashan Parks in the northern mountain provided the largest ecological service potential, whereas environment construction should be further improved in the ecologically fragile area of Lanshan Park.

4.3. Strengths and Limitations of the Current Work

In exploring plant community characteristics and service functions of urban parks, most previous studies have mainly examined changes in the horizontal dimension; studies in the vertical dimension are very limited [2,20,25,45]. This study focused on investigating and analyzing how altitude affects plant community characteristics and their service functions at a quadrat scale in urban parks of a semi-arid city. The results not only assess the health of plant communities and their service functions but also provide a basis for improving the ecological and social service values of plant communities at different altitudes in other, similar regions.

However, this study could not completely reconcile differences in the spatial resolution of Landsat data (pixel size of 30 m) and quadrat-based sampling data (ground size of 10 m × 10 m). Thus, FVC could not be compared at the same spatial scale as other vegetation parameters. The use of unmanned aerial vehicles (UAV) are a potential solution to resolve this discrepancy in future studies [46].

5. Conclusions

An urban park is an important ecological and social asset for the sustainable development of cities, especially those in semi-arid climates. To maintain the health and service function of vegetation communities within urban parks, we need to constantly monitor and analyze the development dynamic of plant communities in different spatial dimensions. In this study, we have explored the plant community characteristics of four urban parks (Jincheng, Renshoushan, Baitashan, and Lanshan) in Lanzhou city across different altitudes, and have qualitatively assessed the service level of the plant community in each park. The variations in the plant community characteristics following elevational changes within all parks exhibited a spatial heterogeneity, especially in Lanshan Park. Generally, cultivated plant communities had the most abundant species accounting for 100% (Jincheng), 30% (Renshoushan), 40% (Baitashan), and 41% (Lanshan) of the total species, especially those with diverse exotic plants, but not always the highest FVC due to space limitation of the landscape facilities. While the natural and semi-natural plant communities were primarily composed of the local plant species and their community characteristics seemed to be regulated by the topographic conditions, they were also impacted by irrigation, park age, and local climates. Jincheng Park had the highest social service level but the lowest natural service level, whereas Lanshan Park exhibited both low social and natural service levels. In contrast, the highest integrated service levels of the plant communities occurred in Renshoushan and Baitashan Parks. These research results provide an important reference to optimize plant communities in parks in Lanzhou city and, more broadly, in other mountainous cities in semi-arid regions.