Relationship between Plant Habitat Types and Butterfly Diversity in Urban Mountain Parks

Abstract

Butterflies serve as valuable indicators of urban ecosystem quality. Due to their accessibility, they also provide urban residents with essential opportunities to connect with nature, fulfilling social functions such as education and recreation, which significantly contribute to city dwellers’ physical and mental well-being. Urban mountain parks are critical habitats for butterflies; analyzing their spatial and temporal distribution and the impact of plant elements is crucial for enhancing plant landscape quality and butterfly diversity. The main results were as follows: (1) A monthly butterfly survey was carried out over the course of a year in the seven urban mountain parks of Fuzhou City. This survey recorded 46 species of butterflies from 36 genera across 7 families, totaling 2506 butterflies. (2) Among the seven habitat types analyzed, TS-, T-, and SG-habitats exhibited elevated levels of butterfly diversity, richness, abundance, and evenness. There were variations in butterfly evenness, diversity, richness, and abundance observed between these habitats. With the exception of N-habitat, there was a consistent seasonal pattern in butterfly diversity across different habitat types. (3) Butterfly diversity and abundance were significantly correlated with vegetation habitat factors across the tree, shrub, and herb layers. Multiple regression modeling using the Akaike information criterion revealed that arbor layer vegetation factors were present in the top four models for butterfly diversity, richness, abundance, and evenness. (4) The quality assessment of different habitat types ranked habitats as follows: TS-habitat > SG-habitat > TSG-habitat > T-habitat > TG-habitat > G-habitat = N-habitat.

1. Introduction

Biodiversity is crucial for maintaining the health of the planet and human well-being. Urban biodiversity is essential for sustaining the stability of urban ecosystems [1,2]. With the development of urbanization, the demand for urban environmental modification has increased dramatically. Consequently, the rapid decline in biodiversity due to these environmental changes has become a serious global ecological problem, negatively impacting wildlife and human survival [3,4,5]. As vital components of natural ecosystems, butterflies are common pollinators in natural communities, highly sensitive to environmental changes, and serve as important environmental indicators [6]. Butterflies respond to changes in environmental patterns and vegetation conditions in urban green spaces to varying degrees and are important indicators of wildlife diversity changes [7]. Therefore, enhancing the habitat environment for butterfly communities in urban green spaces plays a crucial role in creating urban green space landscapes.

Mountain parks in urban settings play a fundamental role in the urban green space network. They are crucial for enhancing urban residents’ quality of life and outdoor recreation satisfaction [8,9,10,11]. Previous research has established a robust association between species diversity and altitude [12,13]. Moreover, urban mountain parks, characterized by their elevated and undulating terrain within urban areas, serve as vital habitats for butterflies. Numerous studies have demonstrated the correlation between butterfly community structure and factors such as topography, plant spatial hierarchy, and plant community structure. Koh et al. (2004) evaluated the efficacy of forest reserves (including safeguarded old secondary and primary forests), forest fragments (consisting of scattered, sparse vegetation), and urban parks (characterized as artificially revegetated habitats) in preserving butterfly biodiversity within the highly urbanized tropics of Singapore. Their study aimed to test the hypothesis that forest reserves exhibit the greatest richness of butterfly species compared with other types of habitats [14]. Jonason et al. (2010) conducted a comparative analysis of butterfly distribution across various habitats at eight randomly selected study sites (each measuring 750 m × 750 m) in southeastern Sweden. Their findings indicated that the most species-rich habitats included other grasslands (such as abandoned fields and fallow lands), open spaces, semi-natural grasslands, and marshes, with each habitat exhibiting unique butterfly species compositions [15]. Lien (2015) studied butterfly community diversity within limestone rainforests across four distinct habitat types (natural forests, secondary forests, shrubs and grasslands, and inhabited areas) in Xuan Son National Park, northern Vietnam. The study found that butterfly community species composition and abundance were similar between natural forests and secondary forests, as well as between shrubs and grasslands and residential areas [16]. Koneri et al. (2016) randomly sampled butterflies from three habitats (primary forest, riverside within forest, and farmland) along a 1000 m transect line in North Sulawesi, Indonesia. The analysis of butterfly community similarity across different habitat types revealed the highest similarity index of 80% between virgin forest and riverside habitats within forests [17]. Peiris et al. (2020) conducted a butterfly survey to evaluate three distinct habitat types: botanical gardens, undisturbed forest habitats within forest reserves, and buffer zones between botanical gardens and forest reserves in Sri Lanka. Their findings indicated that butterfly species richness and diversity were highest in the buffer zone, followed by grassland habitats. Furthermore, significant differences in habitat parameters were observed between the various habitat types [18].

Research on butterfly community structure in different habitat types began relatively late in China. In the 21st century, with rapid urbanization, increased attention has been directed towards the ecological environment of urban green spaces and biodiversity in cities. Consequently, numerous scholars have investigated butterfly diversity in urban and peri-urban green space systems. Yan Hua et al. (2006) conducted a butterfly sampling survey in five cross-sections within Chongqing Municipality. Their findings established a significant positive link between butterfly species and diversity indices with the abundance of vegetation species, vegetation cover, and sunlight in the habitat [19]. Lin Fangmiao et al. (2012) investigated and analyzed butterfly diversity across various habitat types in the primary urban area of Chongqing Municipality. Their findings revealed that anthropogenic factors significantly impact environmental quality, disrupting butterfly colony habitats [20]. Sun Guangfang (2019) surveyed the diversity of butterflies across different habitat types in Jiangyangfan Ecological Park. The results indicated that various farmland habitat types exhibited higher performance in terms of richness, evenness, and diversity index for butterfly communities compared to the non-flowering period [21]. Cheng Fan et al. (2020) analyzed butterfly diversity across various habitats in the Republican Basin of Qinghai. Their study revealed a close relationship between butterfly diversity and habitat type, indicating that the richness of butterfly diversity is positively correlated with habitat complexity [22]. Yang Jingru (2021) investigated the community structure of butterflies across various habitats in typical mountainous areas in southwestern Shanxi. The study also revealed a negative correlation between the species diversity and dominance indexes in butterfly communities [23]. Most relevant studies concentrate on individual parks or limited areas, resulting in relatively restricted and singular study sites. More research on various types of parks and green spaces concerning butterfly community changes and diversity is needed. This gap hinders understanding the interconnectedness of butterfly diversity among urban mountain parks. By leveraging butterflies as ecological indicators, novel perspectives and ideas are introduced for the planning and design of urban mountain parks. This approach facilitates the transition of urban mountain parks from singular functions to multidimensional ecological service providers.

This paper investigates butterfly communities in seven urban mountain parks in Fuzhou City, analyzing the main characteristics and environmental factors influencing them. It aims to explore the relationship between butterfly community landscapes and vegetation habitats within urban mountain parks. Additionally, it seeks to elucidate landscape planning and construction methods for butterfly communities in these parks, promoting harmonious coexistence among people, butterflies, and green spaces while prioritizing protecting butterfly diversity. The research addresses the following questions: (1) Are there differences in butterfly diversity among different vegetation habitats in Fuzhou Urban Mountain Park? (2) Are there seasonal variations in butterfly diversity among different vegetation habitats in Fuzhou Urban Mountain Park? (3) Which vegetation habitat factors influence butterfly community diversity in Fuzhou Urban Mountain Park? (4) How does the environmental quality of butterfly habitats vary across different habitat types? This study offers guidance for landscape planning and design in Fuzhou Mountain Park, particularly regarding butterfly community landscapes and plant habitat creation. It aimed to assist park managers and designers in prioritizing biodiversity conservation while planning and constructing future urban mountain parks.

2. Materials and Methods

2.1. Study Area

Fuzhou City in Fujian Province is positioned along the coast, situated in the eastern part of China, with geographic coordinates ranging from 25°15′ to 26°39′ in northern latitude and 118°08′ to 120°31′ in eastern longitude. The region, spanning 12,251 square kilometers, is known for its favorable climate, featuring mild temperatures, abundant rainfall, and minimal frost or snowfall. Forests cover about 7032 square kilometers, constituting 57.4% of the total area. With an urbanization rate of 71.60%, the city maintains a green coverage of 44.92% in built-up areas and a green space rate of 41.69%. Surrounded by 58 mountain ranges, Fuzhou’s landscape is renowned for its natural beauty [24]. The Fuzhou Municipal Government has established several urban mountain parks throughout the city, leveraging the natural mountainous terrain and abundant natural and semi-natural vegetation resources. Consequently, Fuzhou City is ideal for studying butterfly diversity and green spaces within urban mountain parks.

Based on the urban ring road plan and vegetation characteristics of Fuzhou City [25], the selection criteria for mountain parks within the city limits of Fuzhou were as follows: (1) Different and representative regional locations; (2) park area greater than 8 hectares; (3) presence of typical habitats, diverse vegetation landscape types, and abundant butterfly resources. Following these requirements, seven urban mountain parks were selected as study sites, namely Yushan Park (YSP), Wushan Park (WSP), Pingshan Park (PSP), Meifeng Mountain Park (MMP), Feifeng Mountain Park (FMP), Fushan Country Park (FCP), Niugang Mountain Park (NMP). The geographic distribution of the seven selected urban mountain parks is illustrated in Figure 1.

2.2. Sample Point and Line Setting

In order to enhance the correlation between plant survey data and butterfly survey data, field surveys were conducted using the “nested sample line” method [26]. The specific implementation methods were as follows: Initially, a 100 m sample line was established as the basic unit for butterfly surveys [27]. A 20 m × 20 m standard sample plot was designated for botanical surveys within each sample line. Plant samples were randomly placed along the roadside towards the community’s interior. If field conditions were unsuitable, the sample shapes were adjusted to ensure comparability with standard samples. Determining the number of sample butterfly survey lines in each park followed specific principles: Firstly, based on the park’s size, the number of sample lines was delineated. Secondly, a sample line was allocated for every 2 hectares of land. Thirdly, depending on the park’s layout, the number of sample lines could be adjusted accordingly. Additionally, the number of sample lines per park was capped at 15. Based on park size, this approach aimed to maintain sampling consistency and ensure relatively uniform sampling intensity across green spaces within each unit area. For larger mountain parks, sample line numbers were restricted to prevent excessive repetition of habitat types (Figure 2). Following the principles above, the number of delineated sample lines in the seven mountain parks surveyed in this study is as follows: 4 in Yushan Park, 6 in Wushan Park, 5 in Pingshan Park, 8 in Meifeng Mountain Park, and 15 each in Feifeng Mountain Park, Fushan Country Park, and Niugang Mountain Park. The sample line distribution is depicted in Figure 3.

2.3. Butterfly Survey

The study period spans from July 2022 to June 2023, during which investigators conducted a 12-month butterfly survey, encompassing the changing seasons and butterfly growing and breeding periods in Fuzhou City. Butterfly surveys were conducted under clear, windless weather conditions between 9:00 and 17:00, avoiding the scorching midday hours in summer. Surveys utilized sample line and netting methods, with each sample site surveyed monthly at intervals of 20 days or more. During the survey, investigators moved along the sample line at a constant speed (1–1.5 km/h), observing and recording all butterfly species, numbers, and behaviors within a 20 m radius on both sides of the sample line. Observations were not restricted by height to ensure comprehensive coverage within 2.5 m [27]. Netting and on-site identification were conducted with the naked eye before release. Unidentified species were placed in triangular paper bags and brought to the laboratory for further identification and characterization, referring to resources such as Classification and Identification of Butterflies in China [28] and Monograph of Chinese Butterflies [29]. The butterfly recording is illustrated in Figure 4.

Furthermore, subsequent analyses will adhere to the seasonal timeframes as delineated by the Fujian Meteorological Bureau (http://fj.cma.gov.cn/fzsqxj/, accessed on 30 January 2022). According to their classification, spring spans from March to May, summer from June to August, fall from September to November, and winter from December to February.

2.4. Botanical Survey

The “five-point sampling method” was employed to investigate the vegetation within 68 plant sample plots [30], with the specific sampling process illustrated in Figure 5. Within each 20 m × 20 m sample plot, five 2 m × 2 m medium sample plots and five 1 m × 1 m small sample plots were established in the four corners and the central area of the sample plot. The survey encompassed the following data collection process: Within the tree layer of the standard sample plot, data were collected on tree height, diameter at breast height, crown spread, height below branches, number, and species name of trees; within the center sample area, data were collected on the number, species, and cover of shrubs; within the small sample area, data were collected on the cover and species of herbs. The vegetation stratification structure was determined based on the height of plants above ground level, dividing them into the tree layer (H > 2.5 m), shrub layer (0.5 m~2.5 m), and herb layer (H < 0.5 m).

Following the principle of covering as many vascular plants as possible and referring to the Flora of China, various types of green space information in Fuzhou City, and relevant literature, the analysis focused on the following contents within different habitat types and survey samples [31]:

(1)

Species Important Value of Tree Layer (SIt) refers to the proportion of importance of trees within a habitat or a plant community, with the following formula:

$${\omega }_i={(N}_i÷S)\times 100\%$$

(1)

ω_i represents the proportion of tree species i relative to the overall number of species of plants within the community, where i ranges from 1 to S and S denotes to the total count of plant species.

(2)

Species Important Value of Shrub Layer (SIs) refers to the proportion of importance of shrub plants within a habitat or a plant community, calculated in the same way as species important value of tree layer.

(3)

Species Important Value of Herbaceous Layer (SIg) refers to the proportion of importance of herbaceous plants within a habitat or a plant community, calculated in the same way as species important value of tree layer.

The vegetation habitat factors investigated were screened to produce 13 vegetation variables for information collection as follows:

(1)

Species Diversity of Tree Layer (SDt) is the diversity index for species within the tree layer of a standard sample plot is ascertained through the application of the Shannon–Wiener index, which is calculated using a specific formula:

$$\mathrm{S}\mathrm{D}\mathrm{t}=-\sum P_i\ln P_i$$

(2)

P_i represents the proportion of individuals from plant species i relative to the overall number of individuals in the community, where i ranges from 1 to S and S denotes the total count of plant species.

(2)

Species Diversity of Shrub Layer (SDs) is the diversity index of plants in the shrub layer within the mesocosm, calculated in the same way as species diversity of tree layer.

(3)

Species Diversity of Herbaceous Layer (SDg) is the diversity index of plants in the herbaceous layer of the small sample plot, calculated in the same way as the species diversity of tree layer.

(4)

Richness of Tree Layer (RICt) refers to the number of arboreal plant species in the standard sample plot.

(5)

Richness of Shrub Layer (RICs) refers to the number of shrub plant species in the midsample.

(6)

Richness of Herbaceous Layer (RICg) refers to the number of herbaceous plant species in a small sample.

(7)

Vertical Coverage of Tree Layer (VCt), arboricultural depression, the proportion of the vertical projected area of the arborvitae layer to the total area within each standard sample plot.

(8)

Vertical Coverage of Shrub Layer (VCs), shrub cover, the proportion of the vertically projected area of the shrub layer to the total area within each mesocosm.

(9)

Vertical Coverage of Herbaceous Layer (VCg), herbaceous plant cover, the proportion of herbaceous plant area to total area within each small sample plot.

(10)

Abundance of Trees (ABUt) refers to the total number of individuals of all trees in a standardized sample plot.

(11)

Average Branch Height under Trees (BHt) refers to the average under-branch height (m) of all trees in the standard sample plot.

(12)

Average Height of Tree (AHt) refers to the average height (m) of all arborvitae in the standard sample plot.

(13)

Average Heights of Shrub (AHs) is mean height (m) of all plants in the shrub layer in the fingerling sample plot.

The 13 vegetation factors were categorized according to the tree layer factor, the shrub layer factor, and the herb layer factor. Tree layer factors were available: species diversity of tree layer, richness of tree layer, vertical coverage of tree layer, abundance of trees, average branch height under trees, average height of tree. Shrub layer factors included: species diversity of shrub layer, richness of shrub layer, vertical coverage of shrub layer, average heights of shrub. Herbaceous layer factors included: species diversity of herbaceous layer, richness of herbaceous layer, vertical coverage of herbaceous layer.

2.5. Classification of Plant Habitats

Cluster analysis was performed on all plant samples and the results are shown in Figure 6. Based on the clustering results, the 68 plant samples were classified into habitat types, resulting in six vegetation habitat categories (

Table 1. Typical plant habitats in 7 urban mountain parks in Fuzhou City.

No.	Habitat Type	Characteristic
I	T-habitat	It is dominated by the tree layer, which is characterized by high species abundance and richness, and the shrub and herb layers, which are more homogeneous in terms of plant structural composition.
II	TSG-habitat	At the same time by the influence of trees, shrubs and grasses, the vegetation structure within the habitat is diversified, and the abundance and richness of species in the tree, shrub and herb layers are at a high level.
III	TG-habitat	It is dominated by the tree and herb layers, with high species abundance and richness in the tree and herb layers, and monostructured vegetation in the shrub layer.
IV	G-habitat	The herbaceous layer is dominated by the herbaceous layer, which is characterized by a wide variety of species and a high volume of plants, while the tree and shrub layers are more homogeneous.
V	SG-habitat	It is dominated by the shrub layer and herb layer, with high abundance and richness of plant species in the shrub layer and herb layer, and a monoculture vegetation structure in the tree layer.
VI	TS-habitat	It is dominated by the tree and shrub layers, with high species abundance and richness in the tree and shrub layers, and a monoculture in the herbaceous layer.
VII	N-habitat	The composition of plant species in the tree, shrub, and herb layers is relatively homogeneous, and the plant growth in each vegetation layer is relatively undeveloped.

): (1) T-habitat: vegetated habitat dominated by the tree layer. (2) TSG-habitat: habitat with a rich vegetation structure that is influenced by the triad of tree, shrub, and herb layers. (3) TG-habitat: vegetated habitats co-dominated by the tree and herbaceous layers. (4) G-habitat: Vegetation habitat dominated by the herbaceous layer. (5) SG-habitat: vegetated habitats co-dominated by the shrub layer and herb layer. (6) TS-habitats: dominated by the tree and shrub layers, with high species abundance and richness in the tree and shrub layers and a monoculture in the herb layer. (7) N-habitat: habitats with underdeveloped vegetative growth, virtually unaffected by trees, shrubs, or grasses of any kind.

2.6. Data Analysis

2.6.1. Butterfly Diversity Index

Based on the characteristics of butterfly communities and related data sampling, the following four indices were used to analyze butterfly diversity in this study:

(1)

Richness of Butterfly (RICb) indicates the number of butterfly species surveyed in each urban mountain park or each habitat type.

(2)

Abundance of Butterfly (ABUb) refers to the total number of individual butterflies surveyed in each urban mountain park or each habitat type.

(3)

Species Diversity of Butterfly (SDb) is a composite metric reflecting butterfly abundance and evenness was calculated using Shannon diversity,

$$\mathrm{S}\mathrm{D}\mathrm{b}=-\sum P_i\ln P_i$$

(3)

P_i represents the proportion of individuals from butterfly species i relative to the overall number of individuals in the community, where i ranges from 1 to S and S denotes to the total count of butterfly species.

(4)

Evenness of Butterfly (EVE) indicates how evenly the number of individual butterflies in each urban mountain park or each habitat type is distributed among the butterfly community, calculated using the Pielou metric.

$$\mathrm{E}\mathrm{V}\mathrm{E}=\mathrm{S}\mathrm{D}\mathrm{b}/\ln S$$

(4)

SDb is the butterfly diversity index, S is the total number of butterfly species.

2.6.2. Relationship between Butterfly Communities and Vegetation Habitats

Based on habitat data obtained from field surveys and combined with quantitative analysis, this study utilized cluster analysis to examine 13 plant variables in 68 plant samples from seven urban mountain parks in Fuzhou City. Dimensionality reduction techniques were employed, and the 68 plant sample habitats were classified based on characteristics identified through cluster analysis. We employed the “iNEXT” package [32] to generate the species accumulation curves. These curves are used to assess and predict potential changes in community species richness as the number of samples increases. This methodology, commonly employed in the field of biodiversity research, helps evaluate whether the sample size is sufficient and determines the diversity of species within communities [32,33]. Mann–Whitney–Wilcoxon tests were employed to explore differences in butterfly diversity between different urban mountain parks and between different vegetation habitats. Moreover, differences in butterfly community composition between vegetation habitats were analyzed using non-metric multidimensional scaling (NMDS) [34].

A Spearman correlation test was employed to explore whether a significant positive or negative correlation existed between the butterfly diversity index and the vegetation habitat factor. The Spearman correlation coefficient does not necessitate that the data adhere to a normal distribution. Additionally, to further elucidate the extent of influence of each vegetation habitat factor on butterfly diversity, a stepwise regression method under the Akaike information criterion (AIC) was utilized [35]. Adherence to the principle of ΔAICc ≤ 2 during model selection ensured that the model with the smallest AIC value was selected as the best model. The statistical analyses were performed utilizing R 4.2.3.

2.6.3. Butterfly Habitat Assessment

According to the habitat quality assessment methodology, nine vegetation factors, Species Diversity of Vegetation (SDv), Richness of Vegetation (RICv), Evenness of Vegetation (EVEv), Vertical Coverage of Tree Layer (VCt), Vertical Coverage of Shrub Layer (VCs), Vertical Coverage of Herbaceous Layer (VCg), Average Height of Tree (AHt), Average Heights of Shrub (Ahs), Average Branch Height under Trees (BHt), were selected as indicators to be evaluated [36,37]. Each habitat type was ranked from lowest to highest. According to the correlation results, positive correlation indicators were assigned scores from 1 to 7 in ascending order, while negative correlation indicators were assigned scores from 7 to 1 in descending order. The quality of each habitat type was assessed based on the total score obtained by summing all the indicator scores together.

3. Results

3.1. Overview

A total of 46 species of butterflies from 7 families and 36 genera were recorded in six vegetation habitats across seven urban mountain parks in Fuzhou City between July 2022 and June 2023, with a specific number of 2506 butterflies per year. According to Monograph of Chinese Butterflies, there are 1222 species of butterflies in 11 families and 369 genera in China [29]. Among them, Papilionidae had the highest number of species, with 5 genera and 12 species, constituting 26.09% of the total number of butterfly species. Following closely, Nymphalidae had 11 butterfly species across 9 genera, accounting for 23.91%. Lycaenidae were represented by seven genera and seven species, making up 15.22%. Satyridae comprised five genera and five species, accounting for 10.87%. Pieridae and Danaidae had four genera and five species, amounting to 10.87% of the total. Danaidae had two genera and two species, contributing 4.35% of the total butterfly species. Regarding individual butterfly counts, Pieridae exhibited the highest count at 1021 per survey, followed by Lycaenidae with 863 individuals and Papilionidae with 411 individuals. Satyridae, Danaidae, and Hesperiidae had notably lower counts, with Satyridae and Danaidae recording 18 and 10 individuals per survey, respectively. Hesperiidae had the lowest count, with only seven individuals per survey. Detailed information on specific butterfly families, genera, species names, and Latin names is provided in Table S1. Throughout the study, no species on the national conservation registry were detected [38].

The counts of butterflies within different habitat types, following classification, are depicted in Figure 7. Butterfly community structure appeared relatively rich across the remaining five habitat types, except for only one sample plot in the N-habitat and three in the TS-habitat. As illustrated by the cumulative species curve in Figure 8, the butterfly species observed in various surveyed habitat types tended to plateau gradually with the increase in individuals. This suggested that additional butterfly individuals did not substantially contribute to the discovery of new species, indicating that the recorded butterfly species were gradually saturated and that the collected data adequately fulfilled the requirements for further analysis.

The butterfly community composition across various vegetation habitats in Fuzhou City Mountain Park was analyzed using NMDS, as illustrated in Figure 9. The stress value was 0.198 (<0.2), which indicated discernible differences in butterfly community composition among different habitats. The confidence ellipse ranges for SG-habitat and T-habitat exhibit significant overlap, suggesting a higher degree of similarity in butterfly community composition between these two habitats. Similarly, the confidence ellipse ranges for G-habitat and TG-habitat also overlapped considerably, indicating a similarity in butterfly community composition between them. The data points for TS-habitat were relatively centralized, whereas those for TSG-habitat were more decentralized.

3.2. Distributional Characteristics of Butterfly Diversity in Different Habitat Types

According to Figure 10, regarding butterfly abundance, T-habitats exhibited the highest values (mean ± SD = 10.71 ± 2.89), followed by TS-habitats (mean ± SD = 10.00 ± 1.00). Butterfly abundance in G-habitat was relatively low at 5.75 ± 2.66. Additionally, a pairwise comparison of butterfly richness indices between habitats using the Wilcoxon test revealed significant differences between SG-habitat and G-habitat (p < 0.05), between T-habitats and TG-habitats (p < 0.05), and highly significant differences between G-habitat and TSG-habitat (p < 0.01). Concerning the number of butterfly individuals, the highest levels were observed in the TS-habitat (mean ± SD =51.67 ± 6.11), while G-habitat exhibited relatively lower levels (mean ± SD = 25.25 ± 6.92). Furthermore, a pairwise comparison of the number of individual butterflies among different habitats using the Wilcoxon method indicated highly significant differences between SG-habitat and G-habitat (p < 0.001) and significant differences between G-habitat and TS-habitat (p < 0.05). Differences in butterfly evenness and diversity across habitats were insignificant (p > 0.05). These findings provided valuable insights into the impact of various habitats on butterfly communities and could aid in further ecological research and conservation efforts.

3.3. Seasonal Distribution Characteristics of Butterfly Diversity in Different Habitat Types

Figure 11 shows that butterfly abundance across different habitat types in the seven urban mountain parks in Fuzhou City exhibited similar seasonal trends. Peak abundance occurred in the summer for all habitat types except G-habitat, which peaked in the fall. Conversely, winter marked the lowest point of abundance for all habitats. The number of individual butterflies varied seasonally across different habitat types, with the highest numbers observed in the spring and the lowest in the summer, fall, and winter for N-habitat, mirroring a similar trend in G-habitat. Seasonal variations in the number of individual butterflies were generally consistent across habitat types, except for N-habitat. Butterfly evenness across habitat types displayed seasonal similarities, with N-habitat maintaining consistent evenness in spring and summer, as well as fall and winter. In contrast, the other six habitat types exhibited peak evenness in the fall, which then gradually declined to its lowest point in winter. Butterfly diversity in different habitat types exhibited seasonal similarities, with N-habitat maintaining consistent diversity in spring and summer, as well as fall and winter. Meanwhile, diversity in other habitat types increased gradually from spring to summer, peaked in summer, and decreased gradually from summer to winter.

3.4. Correlation Analysis between Vegetation Habitat Factors and Butterfly Community Index Characteristics

The results of Spearman’s correlation analysis between butterfly community index characteristics and vegetation habitat factors are depicted in Figure 12. Species diversity of butterflies exhibited a positive correlation with vegetation habitat factors such as species diversity of the shrub layer and richness of the shrub layer while showing a negative correlation with factors such as vertical coverage of the herbaceous layer and average height of trees. Butterfly richness positively correlated with vegetation habitat factors such as diversity and richness of the shrub layer. It exhibited a negative correlation with factors such as vertical coverage of the herbaceous layer and average tree height. The evenness of butterflies displayed a positive correlation with vegetation habitat factors such as shrub layer diversity and average shrub height. Butterfly abundance showed a positive correlation with vegetation habitat factors such as richness of the tree layer and abundance of trees while exhibiting a negative correlation with factors such as average tree height and branch height under trees.

Based on the Akaike information criteria (AICc) modeling criteria (

Table 2. Butterfly diversity and vegetation habitat factor model selection results based on the Akaike information criterion (ΔAICc < 2).

No.	Model	df	R2	Adjusted R2	AICc	ΔAICc
1	SDt + SDg + RICt + RICs + RICg + VCt + AHt	58	0.4796	0.4168	−129.52	0.00
2	SDt + SDg + RICt + RICs + RICg + VCt + VCg + AHt	57	0.4874	0.4155	−128.52	1.00

), the optimal modeling formula for predicting the expected value of butterfly diversity in urban mountain parks in Fuzhou City was: SDb = −0.297SDt + 0.386SDg + 0.141RICt + 0.078RICs − 0.164RICg + 0.005VCt − 0.096AHt. The optimal model had ΔAICc = 0.00, R² = 0.4796, adjusted R² = 0.4168, F-statistic = 7.636, p < 0.001. Further analysis of the information on significant parameters of the model of averaged valuation and standard errors within the best model (

Table 3. Significant parameter values within an optimal model of butterfly diversity and vegetation habitat factors based on the Red Pool information criterion (* = p < 0.05, ** = p < 0.01, *** = p < 0.001).

No.	Vegetation Habitat Factors	Estimate	Std. Error	p-Value
1	RICt	0.141	0.049	0.006 **
2	RICs	0.078	0.025	0.002 **
3	RICg	−0.164	0.066	0.016 *
4	VCt	0.005	0.002	0.014 *
5	AHt	−0.096	0.021	<0.001 ***

) showed that RICt, RICs, and VCt had a significant positive effect in the model, while RICg and AHt had a significant negative correlation with SDb in the model.

Based on the Akaike information criteria (AICc) modeling criteria (

Table 4. Butterfly richness and vegetation habitat factor model selection results based on the Akaike information criterion (ΔAICc < 2).

No.	Model	df	R2	Adjusted R2	AICc	ΔAICc
1	SDt + SDg + RICt + RICs + RICg + VCt + AHt	58	0.4739	0.4104	139.51	0.00
2	SDt + SDg + RICt + RICs + RICg + VCt + VCs + AHt	57	0.4806	0.4077	140.66	1.15

), the optimal modeling formula for predicting the expected value of butterfly abundance in urban mountain parks in Fuzhou City was: RICb = −4.291SDt + 3.097SDg + 1.604RICt + 0.413RICs − 1.211RICg + 0.039VCt − 0.585AHt. The optimal model has ΔAICc = 0.00, R² = 0.4739, adjusted R² = 0.4104, F-statistic = 7.462, p < 0.001. Further analysis of significant parameters within the best model of averaging valuation, standard error, and other information (

Table 5. Significant parameter values within an optimal model of butterfly richness and vegetation habitat factors based on the Red Pool information criterion (* = p < 0.05, ** = p < 0.01, *** = p < 0.001).

No.	Vegetation Habitat Factors	Estimate	Std. Error	p-Value
1	SDt	−4.291	1.258	0.001 **
2	SDg	3.097	1.504	0.044 *
3	RICt	1.604	0.374	<0.001 ***
4	RICs	0.413	0.188	0.032 *
5	RICg	−1.211	0.507	0.020 *
6	VCt	0.039	0.017	0.025 *
7	AHt	−0.585	0.163	<0.001 ***

), showed that in the model SDg, RICs, VCt had a significant positive effect, RICt had a highly significant positive effect, RICg had a significant negative correlation with RICb in the model, and SDt, AHt had a highly significant negative correlation with RICb.

Based on the Akaike information criterion (AICc) modeling criterion (

Table 6. Butterfly abundance and vegetation habitat factor model selection results based on the Akaike information criterion (ΔAICc < 2).

No.	Model	df	R2	Adjusted R2	AICc	ΔAICc
1	SDt + SDg + RICt + RICg + VCt + BHt + AHt	58	0.4765	0.4133	392.31	0.00
2	SDt + SDs + SDg + RICt + RICg + VCt + BHt + AHt	57	0.4844	0.4120	393.31	1.00

), the optimal modeling formula for predicting the expected value of the number of butterfly individuals in urban mountain parks in Fuzhou City was: ABUb = −30.102SDt + 20.987SDg + 13.001RICt − 6.733RICg + 0.219VCt − 5.594BHt − 0.096AHt. The optimal model has ΔAICc = 0.00, R² = 0.4765, adjusted R² = 0.4133, F-statistic = 7.542, p < 0.001. Further analysis of significant parameters within the best of model averaging valuation, standard error, and other information (

Table 7. Significant parameter values within an optimal model of butterfly abundance and vegetation habitat factors based on the Red Pool information criterion (* = p < 0.05, *** = p < 0.001).

No.	Vegetation Habitat Factors	Estimate	Std. Error	p-Value
1	SDt	−30.102	8.515	<0.001 ***
2	SDg	20.987	9.506	0.031 *
3	RICt	13.001	2.529	<0.001 ***
4	RICg	−6.733	3.354	0.049 *
5	AHt	−2.613	1.179	0.031 *

), showed that there was a significant positive effect of SDg in the model, RICt had a highly significant positive effect, RICg, AHt had a significant negative correlation with ABUb in the model, and SDt had a highly significant negative correlation with ABUb.

Based on the Akaike information criterion (AICc) modeling criterion (

Table 8. Butterfly evenness and vegetation habitat factor model selection results based on the Akaike information criterion (ΔAICc < 2).

No.	Model	df	R2	Adjusted R2	AICc	ΔAICc
1	SDt + ABUt + AHt + AHs	61	0.2514	0.2023	−323.77	0.00
2	SDt + ABUt + BHt + AHt + AHs	60	0.2721	0.2115	−323.62	0.15
3	SDt + RICs + ABUt + BHt + AHt + AHs	59	0.2883	0.2159	−323.10	0.67

), the optimal modeling formula for predicting the expected value of butterfly uniformity in urban mountain parks in Fuzhou City was: EVE = 0.040SDt − 0.002ABUt − 0.016Aht + 0.040AHs. The optimal model had ΔAICc = 0.00, R² = 0.2514, adjusted R² = 0.2023, F-statistic = 5.122, p < 0.01. Further analysis of the information of significant parameters of the model of averaged valuation and standard error within the best model (

Table 9. Significant parameter values within an optimal model of butterfly evenness and vegetation habitat factors based on the Red Pool information criterion (* = p < 0.05, *** = p < 0.001).

No.	Vegetation Habitat Factors	Estimate	Std. Error	p-Value
1	AHt	−0.016	0.004	<0.001 ***
2	AHs	0.040	0.018	0.031 *

) showed that AHs had a significant positive effect in the model and AHt had a highly significant negative correlation with EVE in the model.

3.5. Environmental Quality Assessment of Butterfly Habitat Habitat in Urban Mountain Parks

According to

Table 10. Environmental quality assessment table for butterfly habitats in different habitats in Fuzhou Urban Mountain Park.

Habitat Type	SDv	RICv	EVEv	VCt	VCs	VCg	AHt	AHs	BHt	Sum
G-habitat	2	2	6	2	2	1	6	1	5	27
N-habitat	1	1	1	1	1	7	7	2	6	27
SG-habitat	6	6	5	3	5	3	5	7	7	47
T-habitat	4	4	2	6	4	5	2	6	2	35
TG-habitat	3	3	4	5	3	2	3	3	3	29
TS-habitat	7	7	7	4	7	6	4	4	4	50
TSG-habitat	5	5	3	7	6	4	1	5	1	37

SDv: Species Diversity of Vegetation; RICv: Richness of Vegetation; EVEv: Evenness of Vegetation; VCt: Vertical Coverage of Tree Layer; VCs: Vertical Coverage of Shrub Layer; VCg: Vertical Coverage of Herbaceous Layer; AHt: Average Height of Tree; AHs: Average Heights of Shrub.

, the environmental quality of butterfly habitats in urban mountain parks in Fuzhou City ranked as follows in descending order: TS-habitat > SG-habitat > TSG-habitat > T-habitat > TG-habitat > G-habitat = N-habitat. N-habitat, which lacked dominance by any tree, shrub, or herb layers, scored lower on positive and negative correlations. The quality assessment scores of TS-habitat and SG-habitat were 50 and 47, respectively, indicating excellent habitat quality. T-habitat (35) and TSG-habitat (37) also exhibited relatively good habitat quality.

4. Discussion

From July 2022 to June 2023, 12 surveys were conducted in seven urban mountain parks in Fuzhou City, yielding a collection of 2506 butterflies. These butterflies belonged to 7 families, 36 genera, and 46 species. Papilionidae and Nymphalidae families collectively represented about half of the recorded species, underscoring their significance within the butterfly community. Conversely, the Hesperiidae family was less represented, with only four species noted. Papilionidae and Nymphalidae butterflies are notable for their large size, diverse species, and generalist behavior, allowing them to adapt to various food sources and geographic regions [39,40]. This adaptability fostered their stability within the community structure, contributing to ecosystem balance and diversity. These findings aligned with previous research outcomes [41,42,43]. Urban mountain parks are an important part of the urban green space system, playing a vital role in maintaining the ecological balance of the city. The vegetation structure correlated with butterfly diversity [44,45], and differences between plant compositions within different habitat types also led to changes in butterfly community diversity [46]. The current landscape design in urban mountain parks predominantly prioritizes the public’s demands for leisure and recreation, with less emphasis on biodiversity studies within park green spaces. However, given the moderate visitor flow and the diverse habitats, urban mountain parks serve as valuable models for developing and enhancing butterfly habitats. Their significance as reference points for biodiversity conservation is particularly notable when compared to urban parks with higher visitor numbers.

4.1. Butterfly Communities in Relation to Different Vegetation Structures

Butterflies are vital pollinators with a crucial role in ecosystems. Throughout their life cycle, most of them form intricate symbiotic relationships with plants. Larvae rely on plants for shelter and food, while adults feed on nectar to sustain themselves and support their development and survival [47,48]. Consequently, alterations in vegetation structure directly impact the composition and distribution of butterfly communities. Changes in habitat vegetation structure lead to corresponding shifts in butterfly communities [49,50]. The study revealed that factors within the tree, shrub, and herb layers significantly influence butterfly communities. Tree layer richness positively correlated with butterfly diversity, richness, and abundance. Additionally, butterfly abundance demonstrated a significant positive correlation with tree abundance. Similarly, shrub layer richness and average shrub height exhibited significant positive correlations with butterfly diversity, richness, and evenness. Furthermore, shrub layer diversity and cover were significantly and positively correlated with butterfly diversity and evenness. These findings underscore the critical impact of species composition within the tree and shrub layers on butterfly diversity and richness. In contrast, herbaceous cover exhibited a significant negative correlation with butterfly diversity and abundance, consistent with previous research results [51]. The tree layer offers shade and shelter for butterflies, whereas the shrub and herb layers serve as rich food sources. Interestingly, herb layer richness did not correlate significantly with butterfly communities, unlike the shrub layer. This contrasts with the findings of Liu et al. [52]. Arboreal richness exhibited a significant positive correlation with butterfly diversity and abundance indices, while no significant relationship was observed with herbaceous layer richness. Herbaceous layers are host-plant specific, with butterflies often centered around their hosts. In our study, the prevalence of trees and shrubs as butterfly host plants outweighed that of herbs. Additionally, habitats dominated by herbaceous layers, such as G-habitat, TG-habitat, and TSG-habitat, showed relatively low butterfly diversity values. This suggested that herbaceous layer richness and cover offer less attraction and protection for butterflies.

4.2. Butterfly Communities in Relation to Different Vegetation Habitat Types

The planning and design of urban mountain parks should address residents’ recreational and relaxation needs while concurrently fostering suitable living conditions for wildlife and providing appropriate habitat environments [53,54]. Variations in vegetation structure give rise to distinctions among habitats, consequently influencing butterfly communities. Urban mountain parks in Fuzhou City prioritize the development of diverse vegetation landscapes with numerous vegetation habitats. Leveraging the 13 vegetation habitat variables obtained from botanical surveys conducted across these parks, clustering analysis was performed, resulting in the identification of seven distinct habitat types.

The analysis revealed significant variations in butterfly diversity and abundance indices across different habitat types, underscoring the impact of vegetation habitats on butterfly community distribution. T-habitat and SG-habitat exhibited notably high levels of butterfly diversity, richness, and abundance indices, suggesting favorable living conditions for butterflies in these habitats. Conversely, the N-habitat, characterized by limited vegetation growth and minimal presence of trees, shrubs, and grasses, displayed lower butterfly diversity and richness indices. This aligned with previous findings by Cheng Fan et al. [55,56,57]. The abundance of vegetation in T-habitat and SG-habitat types, coupled with moderate plant cover and habitat openness, offers ample light resources for butterflies, supporting findings from Yin Liwen’s study [58]. Moreover, Spearman correlation analysis revealed a positive correlation between butterfly community diversity and abundance with vegetation diversity and abundance indices, aligning with studies by Yan Weidong et al. [46,59,60]. Furthermore, no significant differences were observed in the butterfly evenness index across habitat types. This index signified the relative distribution of various species within the butterfly community. The lack of significant differences suggested a balanced distribution of butterfly species across vegetation habitats in urban mountain parks. This balance might stem from the stable structure of vegetation habitats, where diversity and structural complexity create suitable conditions for the survival of different butterfly species.

4.3. Spatial and Temporal Variation of Butterfly Communities in Different Vegetation Habitats

Seasonal variation is a fundamental feature of biodiversity change on ecological time scales [61]. This study analyzed spatial and temporal changes in butterfly communities across various vegetation habitat types within seven urban mountain parks in Fuzhou City. Our findings revealed that indices of butterfly diversity and abundance varied with seasonal changes, indicating the adaptability of butterfly communities to environmental conditions [46]. The pattern of butterfly abundance and peak occurrence differed among habitats and seasons. Overall, butterfly abundance, richness, diversity, and evenness index values were lowest in winter across all habitats. Summer marked the peak values of butterfly richness and diversity in all habitats. However, the number of individual butterflies peaked in spring for G-habitat and N-habitat, while for other habitats, it peaked in summer. Regarding butterfly evenness, all habitat types, except N-habitat, exhibited the highest values in autumn. These seasonal variations were attributed to diverse climatic conditions, plant phenology patterns, availability of host plants, and nectar plants, which collectively influence the habitat’s environmental factors throughout the year [62,63]. The summer and autumn temperatures in Fuzhou create favorable conditions for butterfly communities, promoting their survival and reproductive activities with abundant food resources. During winter, the lower temperatures and relatively scarce food resources lead to decreased butterfly diversity. Consistently, this pattern of low community indices in winter and peaks in summer and autumn aligns with previous findings by Liao Mingwei et al. [63,64].

4.4. Limitation and Development

The study acknowledges several limitations due to its one-year data collection period, which may not fully capture the long-term dynamics of butterfly communities. Climate change and inter-annual variability could influence butterfly composition and diversity over time, necessitating extended monitoring for a more robust understanding. Additionally, the focus on urban mountain parks excludes other park types in the city, which may have differing impacts on butterfly communities. Future research should encompass multiple park types to comprehensively assess these impacts and enhance the study’s practical applicability. While the study examines the influence of plant communities on butterfly diversity, other environmental factors like climate and altitude also play significant roles. Future studies should incorporate a broader range of potential influencing factors to provide a more accurate depiction of butterfly community ecology. Enhancing preliminary survey efforts by diversifying the sampled park types and increasing survey frequency will be crucial. Long-term, systematic monitoring should be prioritized for comprehensive and continuous data collection. Comparative studies across different time scales and park types will offer valuable insights into urban park butterfly diversity. Moreover, a comprehensive consideration of additional environmental factors influencing butterfly diversity is essential. Exploring the integrated mechanisms of these factors in butterfly communities will provide scientific guidance for conservation efforts and contribute to the theoretical foundation for urban park landscape construction and enhancement.

5. Conclusions

This study delved into the intricate relationship between urban butterfly communities and vegetation landscapes across seven urban mountain parks in Fuzhou City. Through comprehensive investigations, we gained insights into butterfly distribution patterns across different vegetation habitats, seasonal variations, and the impact of vegetation habitat factors on butterflies in mountain parks. The key findings are summarized as follows:

(1)

We documented 46 butterfly species spanning 7 families and 36 genera for a year-long monthly survey, totaling 2506 butterflies per survey session. Notably, Papilionidae and Nymphalidae collectively accounted for half of the recorded butterfly species, underscoring their significance in the local butterfly community.

(2)

Significant variations in butterfly diversity and abundance indices were observed across the seven urban mountain parks, highlighting the influence of different habitat types (p < 0.05). Seasonal trends indicated a consistent pattern of high diversity values in summer and autumn, contrasting with lower values in spring and winter across most habitat types, except N-habitat.

(3)

Correlation analyses revealed significant relationships between butterfly community characteristics and vegetation habitat factors within the tree, shrub, and herb layers (p < 0.05). Model selection based on the Akaike information criterion (AICc) highlighted the importance of arboreal layer vegetation in all best models. However, shrub layer vegetation was not included in the best model for butterfly abundance, and herb layer vegetation was absent in the best model for butterfly evenness.

(4)

Environmental quality assessments ranked the habitat types within urban mountain parks, with TS-habitat demonstrating the highest quality, followed by SG-habitat, TSG-habitat, T-habitat, TG-habitat, G-habitat, and N-habitat.

To deepen our understanding of long-term dynamics and diversity in urban park butterfly communities, future studies should intensify preliminary survey efforts by expanding sample park types and increasing survey frequency. Long-term, systematic monitoring is essential for comprehensive data collection. Additionally, a holistic consideration of various environmental factors influencing butterfly diversity is warranted, along with an in-depth exploration of their integrated mechanisms. Such endeavors will provide a robust theoretical foundation for constructing and enhancing urban park landscapes.