Using Tree Inventory to Assess Urban Treescape Diversity and Health in Popular Residential Typologies in the Poznań Metropolitan Area (Poland)

Abstract

Urban landscapes have become widespread as urban areas have grown. Studying the urban environment in terms of the ecosystem services provided is a key trend in contemporary science. This article aims to examine selected popular typologies of residential streetscapes in terms of their tree species diversity and the health of their greenery. The method combined an on-site tree inventory and selected indices relevant to the species richness, diversity, evenness, and nativity. Their correlation with the Vegetation Indices (VIs), expressing the health of the greenery and its density, was assessed. The main findings included the identification of positive correlations between the mean VI values and the diversity and evenness indices and a negative correlation with the tree nativity. The diversity and evenness indices could be used to inform landscape planning decisions and to evaluate both existing and projected treescapes. The nativity of trees should not be prioritized during planting selection; rather, the soil and climate conditions should be considered. As a result of this study, a comprehensive framework for assessing the greenness of streetscapes was developed. Its implementation will aid in directing greenery planning in residential areas towards sustainable development and regenerative urbanism.

1. Introduction

The United Nations’ forecast for 2050 indicates that 68% of the global population will reside in urban areas [1], emphasizing the need for interdisciplinary research to safeguard both urban ecosystem services (UESs) and urban biodiversity (UB) [2]. Researchers have investigated the relationship between ecosystem service delivery levels and biodiversity [3] using three main research approaches. Most studies have focused on (i) the spatiotemporal changes in UESs at different scales and (ii) how they affect regulating ecosystem services, such as carbon sequestration and air pollution removal [2]. The third research area (iii) involves the use of interrelated indicators to simplify decision-making and taking action [4].

Developing indicators for urban ecosystem services and biodiversity is crucial for decision-makers. They can effectively address biodiversity loss by linking these indicators to form comprehensive groups known as linked indicator sets [5]. The UN has set goals to prevent further biodiversity loss due to the global decline in natural ecosystems [6]. Species loss has been linked to climate change, the overconsumption of ecosystem services, and declining human well-being, making it necessary to integrate biodiversity conservation with other policies, including poverty alleviation agendas [7]. Consequently, developing sets of linked indicators is essential for tackling the complex issues facing urban ecosystems. Developing standard procedures for assessing biodiversity in urban areas is of paramount importance [8].

Urban landscapes are typically divided into distinct patches defined by their land use, leading to habitat fragmentation [9,10]. Vegetation-covered permeable surfaces, including small greenery patches, such as private gardens and streetscapes, are considered part of the diverse green infrastructure [11], contributing to the continuity and richness of the entire system. Therefore, residential districts and their greenery play a significant, yet still underestimated, role in the urban ecosystem service network [12]. This highlights the importance of performing diversity analyses at multiple levels, including the neighborhood level.

Assessing urban biodiversity is an enormous challenge for researchers and practitioners because of its complexity, multilayered nature, and scale dependence [9,11,13]. Made up of multiple groups of organisms, biodiversity is predominately measured based on the taxonomic diversity of species [3,14]. Compared with natural ecosystems, urban ecosystems have been observed to undergo biotic homogenization [15], increasing their taxonomic similarity despite diverse local conditions. Consequently, distinct urban communities, such as those housing satellites around different cities, have the potential to replace native species with non-native or synanthropic species [16,17].

Although urban biodiversity results from the diversity of all groups of organisms, its overall level can be linked to the taxonomic diversity of urban trees. Recent research has shown that the presence of fruiting trees contributes to the bird species diversity, increasing wildlife participation [18], in addition to resulting in a high carbon-binding capacity and economic benefits [19]. Similarly, flowering trees have been found to attract birds and pollinators [20], and scattered trees appear to be crucial for maintaining insect diversity [21,22]. The role of trees in conserving and promoting biodiversity in cities has been identified as having the potential to allow for more sustainable landscape management [23,24] and provides guidelines for planting selection [25].

Despite the existing knowledge on the role of street tree species in supporting biodiversity, most cities are dominated by tree monocultures. This accounts for the smaller number of animal species in urban ecosystems and makes tree communities more vulnerable to pests and diseases [26,27]. Moreover, densely populated cities have been found to have an increasing population of non-native trees within their treescapes [13,27]. Non-native tree species are often selected and planted for various purposes, frequently due to their capacity to effectively provide ecosystem services, such as esthetic enhancement or flood regulation. However, it is essential to consider the characteristics of these tree species, including their reproductive rates, the speed of their growth, and their adaptability to diverse conditions, to mitigate the risk of invasion [28]. While some studies have associated the predominance of native tree species in urban ecosystems with increased biodiversity [22,27,29,30], others have provided evidence that non-native species deliver ecosystem services at equal levels and contribute to biodiversity [31,32], enriching the basic set of native species [33]. Therefore, this study examined both native and non-native tree species while simultaneously grappling with the question of whether the presence of trees from one of these two categories yields more favorable outcomes concerning the health conditions of the greenery.

Like most empirical studies on urban biodiversity and ecosystem service provision [2], this study analyzed the taxonomic composition of the tree communities in selected residential districts. As shown in previous studies by various authors, urban biodiversity analysis based on street tree inventories provides valid results in diverse areas and at different scales [12,26,34,35,36,37,38,39]. Considering the lifespan of trees, their sedentary nature, and the fact that they create habitats for other species, scattered trees stabilize the biodiversity in urban ecosystems [21,40] and improve the landscape connectivity [41,42]. The species diversity and spatial distribution are other crucial aspects of urban forest (UF) planning and management [43]. The distribution of UFs within cities is predominantly shaped by social dynamics, including neighborhood features, local government regulations, and the decisions of individual community members [44]. Existing research has evidenced that UF variation can be explained by socioeconomic factors, the landscape cover, and landscape patterns, with socioeconomic factors also playing a vital role [45,46,47] in terms of driving diversity [48].

Urban trees, urban forests (UFs), and urban greenery in general must maintain a healthy condition to effectively provide ecosystem services and enhance the biodiversity within urban environments. The health status of such greenery can be evaluated through various means, including the assessment of its greenness or growth rate. The methodologies employed for these assessments can be categorized into direct methods, such as in situ measurements, and remote methods utilizing satellite sensors. The latter, known as Vegetation Indices (VIs), are increasingly attracting interest due to their provision of free data access and the capability to conduct analyses for any selected period, including retrospective analyses.

The most used remotely sensed indicator of the greenery condition is the Normalized Difference Vegetation Index (NDVI), first used by Rouse et al. in their Great Plains study [49]. The NDVI is calculated as a ratio between the red and near-infrared wavelength reflectance, and it serves to measure the greenness of vegetation, aiding in the evaluation of the plant density and monitoring shifts in plant health [50]. Diverse modifications were later introduced to the NDVI concept, aiming to reduce the interference from the canopy background, atmospheric contamination, and saturation problems [51,52]. One of these was the Soil-Adjusted Vegetation Index (SAVI), which was intended to remove canopy background noise in cases with incomplete vegetation canopies [53], hence making it useful in analyzing urban forests (UFs). The Enhanced Vegetation Index (EVI) was also developed to overcome the limitations inherent in the NDVI, particularly in contexts with dense vegetation and atmospheric effects. The EVI exhibits greater sensitivity to variations in the canopy structure, including in the Leaf Area Index and the canopy architecture, rendering it more suitable for the study of dense, tropical forests where the NDVI may experience saturation. On the other hand, it is more sensitive to the topographic conditions than is the NDVI [54]. Existing studies have found evidence that the EVI predicts the tree species diversity with significant predictive power [55], making it a reliable tool to analyze the tree species diversity using remotely sensed data. Nevertheless, the NDVI is still used as an indicator and has even been shown to be proportionally related to the cooling effect of urban trees [56]. Existing studies have also provided evidence that the NDVI could predict the urban forest LAI, similarly to how it does for natural forests [57].

The Leaf Area Index (LAI) serves as a fundamental measure of vegetation growth. The LAI is calculated as a function of the NDVI or EVI. The existing literature contains a discussion on which method has greater predictive power, and evidence has been found in favor of both indices [58,59,60]. The LAI has been widely used in forestry [60,61] and agriculture [62,63] as it illustrates the primary production levels and represents the dynamic flow of energy and nutrients in the ecosystems. Monitoring the LAI over extended periods is crucial for understanding how vegetation responds to climate change [64], which makes it an interesting metric with which to assess vegetation’s response to urban stress. Existing research has found evidence that rapid urbanization caused a drop in urban forests’ LAI levels [57].

The research presented in this article was informed by the evidenced relationship between urban trees and forests and general biodiversity levels. As highlighted in the existing literature, focusing research on tree species within urban environments offers a comprehensive overview of the general biodiversity status. Simultaneously, it allows for the potential expansion of the methodology to encompass analyses of the health condition of urban greenery with the use of remotely sensed indices. This study responded to the need to record and assess the taxonomic tree diversity in urban forests by focusing on residential districts. The usefulness of the results lies in their informative role in the decision-making process related to sustainable and resilient streetscape design. With such practice-oriented goals, this study began by evaluating and comparing the diversity levels of the existing greenery in the selected housing districts in the study area. Analyzing the existing tree communities in selected residential areas and considering their taxonomic diversity, distribution, and health conditions informed new landscaping projects. The results helped answer the question of how to design sustainable, resilient streetscapes. Providing guidelines regarding tree selection procedures could improve streetscape design standards in housing districts, shifting the current paradigm towards regenerative urbanism. Transferring our acquired knowledge into landscape planning guidelines was a crucial component of this research project, as the existing literature identifies and highlights the need for this [11,65].

Representative of the general levels of biodiversity in urban areas, as the state of knowledge indicates, tree communities are frequently specified by architects as part of streetscape designs. Therefore, the results of this study bridge the gap between ecological knowledge and planning practices, informing planting selection. The usefulness of this work results from the current suburbanization trends in the study area, which are causing an increase in urban greenery at the expense of rural land and, consequently, the popularization of unified landscape typologies with standardized treescapes. The goal of our study was to redirect this practice of planning greenery in residential areas towards sustainability.

The research questions were formulated as follows:

• Which of the compared residential districts is characterized by the most diverse tree community?
• Which district has the highest percentage of native trees?
• Which of the Vegetation Indices (VIs) is most reliable in recording the condition of the urban canopy in the compared districts?
• The canopy of which district scores the best in terms of the remotely sensed Vegetation Indices (VIs)?
• Does tree species diversity improve the health of urban greenery in the study area?
• Are there tree species in the study area that perform better in terms of VI values?
• Is the tree species diversity a valid indicator for predicting the urban ecosystem health?
• Does the nativeness of tree species translate into a good condition of the greenery?

2. Materials and Methods

2.1. Study Area

The research covered three multi-family housing estates in the metropolitan area of Poznań, Poland. Of the six case study districts analyzed in previous research [66], this study considered three of the larger examples. The reason for limiting the selection was the previously obtained conclusion that scale divergence may adversely affect research reliability. The three districts are situated on different sides of the Poznań metropolitan area (Figure 1). They range from approximately 4 to 7.5 hectares in terms of their surface area and have counts of approximately 200 to 350 trees. In terms of age, the three examples represent typical suburban planning solutions popular in their spacetimes: the 1960s, 1990s, and 2020s.

The first case study was a suburban multi-family district, Lubonianka, located in Luboń, south of the Poznań city center. It was planned in the 1960s and was gradually developed over the next two decades. It consists of 12 evenly distributed rectangular blocks, 1 of which has service and administrative functions, whereas the remaining 11 blocks are residential. The blocks are arranged in an orthogonal layout, with longer facades facing east and west. Vehicle access to the blocks is provided by two streets that border the complex to the north and south. The space between the two rows of blocks forms a pedestrian zone with greenery (lawns and trees) and playgrounds. This space has a public character and is generally accessible to residents. The aforementioned spatial features of the Lubonianka estate are typical of settlements from that period.

The estate comprises 440 flats, as reported by the property administrators. This gives us an approximate number of inhabitants of about 1144 persons. The total developed area covered by the study area is 38,026 m², of which 18,778 m² is unsealed land, resulting in 49% of the area being biologically active [66]. The tree community in the estate consists of 206 specimens belonging to 20 different species (Figure 2).

The second example district was the Cegielskiego estate, situated in the commune of Swarzędz, on the western fringe of Poznań. The estate was built during Poland’s political and economic transformation in the 1990s. From the point of view of architectural design, the estate combines the most popular solutions of the previous decades (i.e., big slab technology) with postmodern elements (e.g., symmetric entrance porches and sloped roofs). Regarding the urban layout, the blocks form enclosures with semi-public patios. Vehicle traffic is limited to the outer perimeters of the blocks, whereas the inner patios are pedestrianized. They include greenery, outdoor rest areas, and playgrounds.

According to information provided by the property administration, the district consists of six large blocks containing 654 flats. This gives us an approximate number of 1707 residents. The total development area covers 49,891 m², with 19,827 m² of green area, giving us a percentage of 40% of the area that is biologically active [66]. The tree community in the district consists of 215 specimens belonging to 19 different species (Figure 3).

The last case study, the Grafitowe estate in the commune of Dopiewo, exemplifies one of the contemporarily popular residential typologies in suburban areas. The district consists of 33 low-rise apartment houses characterized by contemporary and economical architectural solutions. The shapes of the individual buildings are compact, and common areas (e.g., communication areas) are optimized. The estate represents the most popular contemporary trend in planning and landscape design in Poland. Green areas at ground level are divided into private gardens and public communication areas. The gardens are private parts of ground-floor apartments and are fenced. Streets and pedestrian paths line a minimalistic streetscape consisting of lawns and grafted trees. A common playground is situated at the western end of the entire complex.

According to the in situ analysis conducted by the authors, the estate comprises 544 homes with approximately 1534 residents [66]. A total of 347 trees representing eight different species were identified in the estate (Figure 4).

2.2. Materials

The materials used in this research were partially developed in relation to previous research devoted to assessing the greenery content in housing districts in the study area [66]. They included imagery obtained from the spatial information system of the Poznań metropolitan area [67], which allowed for the generation of updated and scaled maps, which could then be imported into CAD or GIS software (ArchiCAD 24 EDU and QGIS 3.40). Subsequently, the greenery (lawns and trees) was inventoried and marked on the maps according to its existing state. This step used imagery provided by Google Maps and Google Street View, Poland’s online tree information service [68], site visits, and on-site inventories.

For this study, spectrometric data were obtained from the Copernicus Sentinel–2B satellites [69] using the online application Copernicus Browser [70]. Specifically, bands 2 (blue), 4 (red) and 8 (near-infrared) at a resolution of 10 m were the main focus points, as they were the basis for the VIs’ calculation. We sought imagery from the months with the peak growth of vegetation in our region, which occurs between mid-May and mid-August, for 2021, 2022, 2023, and 2024, that is, after the implementation of the latest case studies in these districts. The maximum cloud coverage was set at 30%, with a preference of 20%. This resulted in one dataset per year, dated 13 August 2021, 3 August 2022, 9 July 2023, and 14 May 2024.

After downloading the data, bands 2 (blue), 4 (red), and 8 (near-infrared) were imported into QGIS 3.40 software, where selected VIs were calculated using the formulas described below. In addition to the graphical presentation of the NDVI, SAVI, and EVI values, the minimum, median, mean, maximum, and range values were calculated for the three case study districts, for their administrative units (communes), and for their tree canopies, as well as for separate tree clusters.

2.3. Methodology

The three selected districts were subjected to and on-site tree inventory, including the location of the trees, species recognition, and trunk circumference measurements taken at a height of approximately 1.20 m. The inventory took place in summer (June–August) 2023, and the data collected were used to calculate selected indicators reflecting the diversity of the tree species and the general condition of the ecosystem. The set of interlinked indices used in this study included a selection of classic formulas used to assess diversity (e.g., the species richness, Simon’s Diversity Index, Shannon’s Diversity Index, and Pielou’s Evenness) [71,72,73,74,75,76,77,78,79]. An attempt was also made to analyze the nativeness of the species based on the literature [80,81]. The selection was completed by remotely sensed Vegetation Indices (VIs), metrics for quantifying the health and density of vegetation using sensor data: the Normalized Difference Vegetation Index (NDVI), the Soil-Adjusted Vegetation Index (SAVI), the Enhanced Vegetation Index (EVI), and the Leaf Area Index (LAI) [82]. Subsequently, the correlation between the two groups of indicators was calculated, determining which factors reflecting the species richness, diversity, evenness, and nativity could exert a positive influence on the health conditions of vegetation. The research results and observations constituted the basis for our conclusions, which comprised guidelines on further using the selected indices to inform landscape design projects (Figure 5).

The novelty of this approach lay in the combination of tools used from various disciplines. First, the case study selection procedure was grounded in the analysis of the architectural and spatial typologies. The on-site inventory was conducted as per the generally accepted principles of inventories in landscape architecture. The set of indices used to assess the richness, diversity, and evenness of tree species stemmed from ecology. The tree nativity analysis was informed by the existing catalogs of the vascular plant species of the region. The NDVI, SAVI, EVI, and LAI are widely used in diverse disciplines because of their accuracy in reflecting the general state of vegetation. Usually used on larger scales, VIs were applied in this study at the local scale, allowing for a comparison of housing typologies in terms of the streetscape resilience. Finally, the Pearson correlation was measured between the indicators of the species richness, diversity, evenness, and nativity on one side and the VIs on the other side to analyze their interdependence.

2.3.1. Indicators of Species Richness, Abundance, Diversity, and Evenness

The species richness (S) is the simplest measure of species diversity and can be either a count of the number of species or a list of species inhabiting a particular area or habitat [72]. The abundance (N) corresponds to the number of observations or individuals. Their relationships form the basis of defining diversity. In ecology, it is customary to distinguish between taxonomic and functional diversities. The first is defined as “the number and the relative abundance of species in a community”, and the latter refers to “different types of processes in a community that are important to its structure and dynamic stability” [83] (p. 205). These two seemingly competing approaches to measuring diversity are strongly interrelated, translating into a full picture of diversity and informing relevant governance policies [83]. The following study was conducted from the perspective of taxonomic diversity because it concerned a group of organisms with similar functionality in the ecosystem (i.e., trees). To obtain comparable results, the diversity of the tree species was related to the total area of the unsealed surface within the case study boundary, yielding a Species Richness Index. In this study, the Species Richness Index (R_S) was calculated for the tree communities in the three selected case studies using the following equation:

$${\mathrm{R}}_{\mathrm{S}}={\displaystyle \frac{\mathrm{S}}{\mathrm{A}}}$$

(1)

where R_S is the Species Richness Index, S is the number of tree species identified during the on-site inventory for each case study, and A is the total area of the unsealed surface in the district. The area unit used was 1000 m².

The Biodiversity Index (R_BIO), the most commonly used index for characterizing forest ecosystems, is defined as the ratio of the number of species (S) to the number of individuals (N) in a sample [73,84]. This study used the following formula:

$${\mathrm{R}}_{\mathrm{B}\mathrm{I}\mathrm{O}}={\displaystyle \frac{\mathrm{S}}{\mathrm{N}}}$$

(2)

where R_BIO is the Biodiversity Index, S describes the total number of species identified in each case study area during the on-site inventory, and N denotes the total number of trees counted.

Owing to its simplicity, the Biodiversity Index formula does not consider the sample size or total surface area. To fill this gap, this study used additional diversity indices. Akin to the Biodiversity Index, Menhinick’s Index assesses the relationship between the species richness and the sample size. It normalizes the species richness by the community size, thereby reducing the potential impact of variations in the sample size across communities [71]. It is defined as the ratio of the number of species to the square root of the number of individuals in a sample [73]. In this study, it was calculated using the following formula:

$${\mathrm{R}}_{\mathrm{M}\mathrm{E}\mathrm{N}}={\displaystyle \frac{\mathrm{S}}{\sqrt{\mathrm{N}}}}$$

(3)

where R_MEN is Menhinick’s Index, S is the total number of species, and N is the total number of individuals in the sample.

Simpson’s Diversity Index (D) is one of the most widely used and reliable methods for evaluating the community biodiversity. It is a measure of species diversity that considers both the number of species present and the relative abundance of each species. It is determined by the relationship between the number of different species in a habitat (species richness) and the number of organisms (individuals) representing each species. Its original form is calculated using the following formula [74]:

$$\mathrm{D}={\displaystyle \frac{\sum {\mathrm{n}}_{\mathrm{i}}({\mathrm{n}}_{\mathrm{i}}-1)}{\mathrm{N}(\mathrm{N}-1)}}$$

(4)

where D is Simpson’s Diversity Index, n_i is the number of individuals belonging to the ith species, and N is the total number of individuals in the community. As the value of D increases, the diversity decreases. For a more intuitive understanding of the results, this study used Simpson’s Reciprocal Index, calculated as follows:

$${\mathrm{D}}_{\mathrm{I}}={\displaystyle \frac{1}{\mathrm{D}}}={\displaystyle \frac{\mathrm{N}(\mathrm{N}-1)}{\sum {\mathrm{n}}_{\mathrm{i}}({\mathrm{n}}_{\mathrm{i}}-1)}}$$

(5)

where D_I is Simpson’s Reciprocal Index, n_i is the number of individuals belonging to the ith species, and N is the total number of individuals in the community.

The Shannon Diversity Index, often referred to as the Shannon–Wiener Index or Shannon entropy, is a measure of the variety of species found within a community. It is widely used in ecology and has proven to be a valuable tool for understanding and quantifying biodiversity [71]. This index was developed by Claude Shannon in 1948 and provides a numerical value that represents the diversity of a community by considering both the number of species present and their relative abundance [75]. The Shannon Diversity Index was calculated by summing the proportion of the number of individuals of each species multiplied by the logarithm of that number.

$$\mathrm{H}=-\sum _{\mathrm{i}=1}^{\mathrm{S}}{\mathrm{p}}_{\mathrm{i}}\times \mathrm{l}\mathrm{n}({\mathrm{p}}_{\mathrm{i}})$$

(6)

where H is the Shannon Diversity Index, S is the total number of individual species, p_i is the proportional abundance of the ith species, and ln(p_i) is the natural logarithm of this number.

The Shannon Equitability Index (E_H) [75], akin to Pielou’s Evenness (J) [77], is a useful tool for determining the uniformity of the species abundance in a particular community [71]. This index is based on the concept of evenness, which is characterized by similarity in the abundance levels of different species in the community. This was calculated using the following formula:

$${\mathrm{E}}_{\mathrm{H}}={\displaystyle \frac{\mathrm{H}}{\mathrm{l}\mathrm{n}\mathrm{S}}}$$

(7)

where E_H is the Shannon Equitability Index, H is the Shannon Diversity Index (calculated as described above), and S is the total number of individual species in the community.

2.3.2. Tree Nativity Analysis

In this study, the nativity of the tree species in the region was measured using the simple ratio of the number of native tree species to the overall number of species in the community (S), as well as the ratio of the number of native trees to the total number of individuals (observations). The native or non-native character of a species was defined according to the information provided in the relevant literature: a key to the identification of vascular plants in lowland Poland [81], a guide to recognizing tree and shrub species in the region of Greater Poland [80], and a public encyclopedia. Grafted trees of native species were considered native, prioritizing the consideration of these taxa. Two formulas were used to assess these proportions:

$${\mathrm{R}}_{\mathrm{S}\mathrm{N}}={\displaystyle \frac{{\mathrm{S}}_{\mathrm{N}}}{\mathrm{S}}}$$

(8)

where R_SN is the proportion of native tree species, S_N is the number of species defined as native to the region, and S is the total number of species in the community, and

$${\mathrm{R}}_{\mathrm{N}\mathrm{T}}={\displaystyle \frac{{\mathrm{T}}_{\mathrm{N}}}{\mathrm{T}}}$$

(9)

where R_NT is the proportion of native tree specimens, T_N is the number of native tree observations (individuals), and T is the total number of trees in the community.

2.3.3. Vegetation Indices (VIs)

The Normalized Difference Vegetation Index (NDVI) is a popular metric used to quantify vegetation greenness. It is useful for understanding the vegetation density and assessing changes in plant health [85]. Its concept was based on the rationale that pigments in plant leaves strongly absorb visible light for use in photosynthesis, while the structure of the leaves, on the other hand, strongly reflects near-infrared light [82]. The NDVI is calculated using spectrometric data sourced from remote sensors, such as satellites, as a ratio between the red (R) and near-infrared (NIR) values:

$$\mathrm{N}\mathrm{D}\mathrm{V}\mathrm{I}=(\mathrm{N}\mathrm{I}\mathrm{R}-\mathrm{R}\mathrm{e}\mathrm{d})/(\mathrm{N}\mathrm{I}\mathrm{R}+\mathrm{R}\mathrm{e}\mathrm{d})$$

(10)

where NDVI is the Normalized Difference Vegetation Index and NIR and Red stand for the spectral reflectance measurements acquired in the near-infrared and red (visible) regions, respectively [86]. The NDVI value ranges from −1 to 1. The closer it is to 1, the healthier the greenery. Areas with no vegetation typically yield an NDVI value of zero. Surface water typically yields an NDVI of −1.

The Soil-Adjusted Vegetation Index (SAVI) was developed from the Normalized Difference Vegetation Index (NDVI) to reduce the influence of the soil brightness on spectral Vegetation Indices using red and near-infrared (NIR) wavelengths. Elaborated by Huete, this transformation shifts the origin of the reflectance spectra in the NIR − red wavelength space to account for soil–vegetation interactions and the extinction of red and NIR flux when passing through canopies [53]. The SAVI is defined as follows [87]:

$$\mathrm{S}\mathrm{A}\mathrm{V}\mathrm{I}=\left(\right(\mathrm{N}\mathrm{I}\mathrm{R}-\mathrm{R}\mathrm{e}\mathrm{d})/(\mathrm{N}\mathrm{I}\mathrm{R}+\mathrm{R}\mathrm{e}\mathrm{d}+\mathrm{L}\left)\right)\times (1+\mathrm{L})$$

(11)

where SAVI is the Soil-Adjusted Vegetation Index, NIR and Red are the near-infrared and red spectral reflectance measurements, and L is the soil brightness correction factor (and could range from 0 −1). Following the instructions from the Sentinel, the value of L was set at L = 0.428. The values of the SAVI range from −1 to +1, in a way analogous to the NDVI.

The Enhanced Vegetation Index (EVI) improves the accuracy of the NDVI by using blue wavelengths to correct for soil background signals and atmospheric influences. It is defined as follows [52,88]:

$$\mathrm{E}\mathrm{V}\mathrm{I}=\mathrm{G}\times (\mathrm{N}\mathrm{I}\mathrm{R}-\mathrm{R}\mathrm{E}\mathrm{D})/(\left(\mathrm{N}\mathrm{I}\mathrm{R}+\mathrm{C}1\times \mathrm{R}\mathrm{E}\mathrm{D}-\mathrm{C}2\times \mathrm{B}\mathrm{L}\mathrm{U}\mathrm{E}\right)+\mathrm{L})$$

(12)

where EVI is the Enhanced Vegetation Index, G is the gain factor, L is a soil adjustment factor, and C1 and C2 are coefficients used to correct aerosol scattering in the red band through the use of the blue band. NIR, Red, and Blue represent the reflectance at the near-infrared (NIR), red, and blue wavelengths. G = 2.5, C1 = 6.0, C2 = 7.5, and L = 1 [51,54]. The range of values for the EVI is −1 to 1, with healthy vegetation generally scoring around 0.20 to 0.80 [88].

The indices of the NDVI and EVI were also used to calculate the Leaf Area Index (LAI) according to the following equations [59,60]:

$$\mathrm{L}\mathrm{A}\mathrm{I}=0.57\times \mathrm{e}\mathrm{x}\mathrm{p}(2.33\times \mathrm{N}\mathrm{D}\mathrm{V}\mathrm{I})$$

(13)

$$\mathrm{L}\mathrm{A}\mathrm{I}=3.618\times \mathrm{E}\mathrm{V}\mathrm{I}-0.118$$

(14)

In this study, the Vegetation Indices (VIs) were used as a reflection of the streetscape conditions in the three case study areas. We investigated whether different urban typologies combined with greenery, designed in a specific way, including the selection of tree species, translated into the health status of the greenery.

3. Results

The research results can be presented in two categories: (i) the results from the assessment of the existing typical suburban residential treescapes in the study area in terms of the tree species diversity, nativity, and general health condition, and (ii) the correlation between the tree species richness, diversity, evenness, and nativity and the remotely sensed general condition of the greenery.

3.1. The Diversity, Nativity, and Condition of Popular Treescapes in Multi-Family Residential Districts

The first set of indices calculated in this study (

Table 1. Calculated values of indices relevant to tree species composition in three case study districts.

Group of Indices	Indices	Lubonianka	Cegielskiego	Grafitowe
Richness	Species richness (S)	20	19	8
	Abundance (N)	206	215	347
	Species Richness Index (RS)	1.07	0.96	0.27
Diversity	Biodiversity Index (RBIO)	0.1	0.09	0.02
	Menhinick’s Index (RMEN)	3.76	3.54	1.37
	Simpson’s Diversity Index (D)	0.12	0.14	0.34
	Simpson’s Reciprocal Index (DI)	8.53	7.13	2.97
	Shannon Diversity Index (H)	2.44	2.33	1.42
Evenness	Shannon Equitability Index (EH)	0.82	0.79	0.68
Nativity	Proportion of native tree species (RSN)	0.65	0.63	1
	Proportion of native tree specimens (RNT)	0.84	0.9	1

) reflected the tree species composition in three exemplary residential districts concerning the species richness, diversity, and evenness. An analysis of the tree species nativity was also performed and quantified using two indices. Additionally, the health condition of the greenery was reflected by the mean values of the selected remotely sensed Vegetation Indices (VIs) (see Section 3.1.3).

3.1.1. The Greater Diversity of the Tree Community in the Oldest of Three Suburban Housing Estates in Terms of the Species Richness, Diversity, and Evenness

The species richness reflected the number of tree species observed in each case study area. Its value was the highest for the two older neighborhoods and the lowest for the newest district covered by the study. This situation can be partially explained by the passing of time. Tree communities in older districts no longer reflect the original planning and landscaping solutions. Some trees have been planted by residents over the years, in addition to self-seeders. In contrast, the newest of the three districts contained almost exclusively trees planted as part of the designed landscape. For this same reason, however, the state of the tree community provided the best basis for assessing the greenery in terms of the current landscape planning practices.

The species richness in the newest of the three districts (Grafitowe) was 60% lower than in the oldest (Lubonianka), whereas the abundance (N) was 68% greater in the newest district (Figure 6a,b). This indicates that the tree community in this district has been monocultured, which remains a common practice in Poland. Moreover, the ecosystem was dominated by the seedlings of two species of ornamental grafted trees: a small variety of cherry (Prunus eminens Umbraculifera) and Norway maple (Acer Globosum).

The Species Richness Index (R_S), which reflected the number of species in relation to the green area in each case study district, was four times greater in the oldest district (Lubonianka) than in the newest district (Grafitowe). The second case study (Cegielskiego) achieved a result close to that of the leader, three and a half times higher than that of the newest estate (Figure 6c).

The diversity indices confirmed the disproportion between the two old housing estates and one new housing estate in terms of the tree species variety. The popularly used Biodiversity Index (R_BIO), defined as the ratio of the number of species to the number of individuals in the sample, showed a five-fold advantage for old housing estates over new ones (Figure 7a). Menhinick’s Index (R_MEN), which reduces the potential impact of variations in the sample size across communities, was 64% lower for the new estate (Grafitowe) compared with the oldest estate (Lubonianka) and 61% lower than that of the middle estate (Cegielskiego) (Figure 7b). Similarly, Simpson’s Diversity Index (D) and Simpson’s Reciprocal Index (D_I) indicated that the result for the new housing estate was 65% lower than that for the oldest estate and 58% lower than that for the estate that was chronologically in the middle (Figure 7c). The Shannon Diversity Index (H) considers both the number of species present and their relative abundance. The value of this indicator was again the highest for the oldest of the three districts (Table 1). The newest of the three districts had a result that was 42% lower than that of the leader and 39% lower than that of the middle district (Figure 7d).

Answer to Research Question 1: Lubonianka, the oldest of the three districts, had the most diverse community of trees.

The evenness refers to the comparable abundance of various species within a community ecosystem. A greater value in terms of evenness indicators indicates that the individuals in a community are distributed more equitably among the species. The Shannon Equitability Index (E_H), which specifies the uniformity of the species abundance in a particular community, yielded less differentiated results. The value of the oldest estate compared to the newest estate in this respect was 17% higher, while the result for the middle example was 14% higher than that of the newest estate (Figure 8).

3.1.2. The Newest of the Three Compared Districts Had the Highest Percentage of Native Trees

In contrast to all the previous indicators, the analysis of the tree species’ nativity rates showed an advantage of the new housing estate over the two older ones. The proportion of native tree species (R_SN) was the highest for the newest estate (Grafitowe). The result for the oldest estate was 35% lower, and the result for the middle example was 37% lower (Figure 9a). The proportion of native tree specimens (R_NT) was also the highest for the newest estate. The result for the oldest one was 16% lower, and that for the middle one was 10% lower (Figure 9b). As the newest of the surveyed estates, Grafitowe had not yet had any self-seeded plants or seedlings added by the residents. Therefore, the vision for its streetscapes created by its designers remained visible. Interestingly, only native tree species had been planted in the newest estate (Grafitowe). However, most of them had features that differed from those that occur naturally. For example, hornbeam trees (Carpinus betulus Fastigiata) were shaped differently to usual to serve as hedges. The maple trees planted in the district were mainly of the Globosum variety (Acer platanoides Globosum), with limited growth. Finally, the dominant species, whose representatives comprised 53% of all individuals, were small, grafted cherry trees (Prunus eminens Umbraculifera), a naturally occurring hybrid of two other cherry species native to central Europe.

Answer to Research Question 2: Grafitowe, the newest district, had the highest proportion of tree nativity.

3.1.3. The Most Diverse and Mature Tree Canopy Scored the Best in Terms of the VI Values

The values of the selected Vegetation Indices (VIs)—the NDVI, SAVI, EVI, and LAI—were calculated for predefined dates, determined by criteria pertaining to the growing season and obtaining a view of the Earth with minimal cloud obstruction (see Section 2.3.3). The results of the calculations (

Table 2. VI values obtained for preselected dates over a four-year period for the three case study districts and their canopies, rounded to three decimal places.

Date	VI Values	Lubonianka Canopy	Lubonianka District	Cegielskiego Canopy	Cegielskiego District	Grafitowe Canopy	Grafitowe District
13 August 2021	Mean NDVI	0.449	0.351	0.429	0.311	0.365	0.256
	Min NDVI	0.223	0.011	0.125	0.039	0.090	−0.002
	Max NDVI	0.760	0.774	0.649	0.652	0.700	0.723
	Mean SAVI	0.641	0.515	0.612	0.444	0.514	0.293
	Min SAVI	0.318	0.015	0.179	0.056	0.129	−0.003
	Max SAVI	1	1	0.927	0.932	0.999	1
	Mean EVI	0.322	0.263	0.240	0.178	0.220	0.164
	Min EVI	0.081	0.005	0.043	0.017	0.072	−0.001
	Max EVI	1	1	0.369	0.452	0.526	0.570
	LAI (NDVI-based)	1.622	-	1.548	-	1.334	-
	LAI (EVI-based)	1047	-	0.751	-	0.676	-
3 August 2022	Mean NDVI	0.400	0.264	0.247	0.167	0.229	0.166
	Min NDVI	0.182	0.012	0.055	0.008	0.073	0.001
	Max NDVI	0.557	0.557	0.415	0.415	0.436	0.544
	Mean SAVI	0.572	0.397	0.353	0.238	0.326	0.237
	Min SAVI	0.260	0.018	0.079	0.012	0.104	0.001
	Max SAVI	0.795	0.795	0.593	0.593	0.623	0.777
	Mean EVI	0.419	0.291	0.256	0.175	0.246	0.182
	Min EVI	0.152	0.013	0.049	0.007	0.090	0.001
	Max EVI	0.669	0.669	0.399	0.412	0.495	0.632
	LAI (NDVI-based)	1.449	-	1.014	-	0.970	-
	LAI (EVI-based)	1.397	-	0.807	-	0.771	-
9 July 2023	Mean NDVI	0.407	0.282	0.289	0.200	0.233	0.178
	Min NDVI	0.199	0.009	0.105	0.031	0.061	0.005
	Max NDVI	0.572	0.608	0.434	0.434	0.430	0.442
	Mean SAVI	0.597	0.421	0.396	0.286	0.328	0.250
	Min SAVI	0.285	0.012	0.170	0.022	0.090	0.002
	Max SAVI	0.868	0.868	0.620	0.620	0.614	0.632
	Mean EVI	0.430	0.304	0.286	0.210	0.237	0.186
	Min EVI	0.207	0.011	0.109	0.020	0.071	0.002
	Max EVI	0.714	0.714	0.439	0.464	0.462	0.499
	LAI (NDVI-based)	1.470	-	1.119	-	0.981	-
	LAI (EVI-based)	1.439	-	0.918	-	0.739	-
14 May 2024	Mean NDVI	0.425	0.303	0.302	0.215	0.252	0.209
	Min NDVI	0.202	0.039	0.112	0.043	0.080	0.036
	Max NDVI	0. 586	0.587	0.478	0.478	0.469	0.512
	Mean SAVI	0.607	0.451	0.431	0.307	0.360	0.299
	Min SAVI	0.288	0.066	0.159	0.062	0.115	0.051
	Max SAVI	0.837	0.838	0.682	0.682	0.669	0.731
	Mean EVI	0.453	0.339	0.322	0.228	0.268	0.229
	Min EVI	0.185	0.054	0.104	0.044	0.091	0.036
	Max EVI	0.714	0.714	0.504	0.510	0.534	0.592
	LAI (NDVI-based)	1.534	-	1.153	-	1.026	-
	LAI (EVI-based)	1.519	-	1.048	-	0.850	-
4-Year Average	Mean NDVI	0.420	0.300	0.317	0.223	0.270	0.202
	Min NDVI	0.201	0.018	0.099	0.031	0.076	0.010
	Max NDVI	0.619	0.632	0.494	0.495	0.509	0.555
	Mean SAVI	0.604	0.446	0.448	0.319	0.384	0.288
	Min SAVI	0.288	0.028	0.147	0.038	0.109	0.013
	Max SAVI	0.896	0.875	0.705	0.706	0.726	0.793
	Mean EVI	0.406	0.300	0.276	0.198	0.242	0.190
	Min EVI	0.156	0.021	0.076	0.022	0.081	0.009
	Max EVI	0.774	0.774	0.428	0.459	0.504	0.573
	LAI (NDVI-based)	1.519	-	1.208	-	1.078	-
	LAI (EVI-based)	1.351	-	0.881	-	0.759	-

) reflected the annual changes in the state of the greenery, which were linked to the changing weather conditions. The average temperature in Poznań is 9.7 °C, with an annual precipitation of about 650 mm [89]. July is the rainiest and warmest month in Poznań and its surroundings, with an average rainfall of 92 mm and a medium temperature of 19.8 °C, based on data from 1991 to 2021 [89]. The years of 2021 and 2022 were dry, with annual precipitations of approximately 500–550 mm and 400–450 mm, respectively [90]. July, statistically the rainiest month, brought less than 50 mm of rain in 2021 and less than 30 mm in 2022 [90]. Drought contributed to the decrease in the mean NDVI and SAVI values (Table 2, Figure 10). After 2022, the values of those two indices started to increase steadily again, which reflected the recovery of the greenery after the drought. Interestingly, the EVI values showed a modest but stable upward trend despite the drought in 2021 and 2022. This could be explained by the growth of the tree canopies and the fact that an increase in this is better reflected by the EVI values, as the EVI formula was developed to overcome the vulnerability of the NDVI to interference from the soil and atmospheric influences. While dry grasses and herbaceous vegetation contributed to lowered NDVI values during the drought, the EVI results were not affected and reflected the stable growth of the tree canopy.

Answer to Research Question 3: the EVI was more precise in detecting the condition of the tree canopies in residential districts, while the NDVI and SAVI were affected by the seasonal droughts and yielded a more general image of the state of the total greenery, including herbs, grasses, and shrubs.

While comparing the average NDVI, SAVI, EVI, and LAI values for the three housing estates that were the subject of this analysis, it should be noted that Lubonianka, the oldest of the three districts, scored the best. Its results were higher, in some cases by more than 50%, than those of the two newer estates, Cegielskiego and Grafitowe (Table 2, Figure 10). Another observation is that although the values obtained for the tree canopies were higher than those for the districts in their entirety, the parallelism of the results is clearly visible. Lubonianka’s canopy was formed of the most diverse combination of tree species, including both native and non-native species. Due to its age, it also had the largest percentage of mature trees. The fact that the highest VI values were obtained for this district provides evidence of the impact that trees have on the overall condition of urban greenery, which tends to be healthier and more stable when mature trees of diverse species are present.

Answer to Research Question 4: The tree canopy in Lubonianka, the oldest district, had the highest mean values for all the VIs considered.

Answer to Research Question 5: A greater diversity of tree species contributed to better VI values both for the canopy and for the residential district in its entirety.

In this study, the Leaf Area Index (LAI) was calculated using two formulas, which were NDVI-derived and EVI-derived (Equations (13) and (14)). The results showed a difference in the LAI values obtained using the two calculation methods (Figure 11). The NDVI-derived LAI showed a drop in its values during the dry period of 2021 and 2022 and a stable upward tendency after. The EVI-derived LAI values showed proportional growth across the entire period of comparison. Looking for an answer to the question regarding which LAI formula yields more stable results for urban greenery, we could refer to the previous observation that NDVI values tend to be more affected by the background [52], e.g., the dried-out state of herbaceous vegetation, while the EVI is more precise in reflecting the condition of the tree canopy. Therefore, the EVI-based LAI seems to be more appropriate to assess the state of urban trees and forests.

The NDVI, SAVI, and EVI values, calculated based on the imagery obtained from the Copernicus Sentinel–2B satellites, could also be expressed graphically in pixel form at 10 m resolutions (Figure 12, Figure 13 and Figure 14). The visual analysis of the graphic material confirmed that the Lubonianka estate was the most stable of the three case study areas in terms of maintaining high VI values. Based on the images, it can also be concluded that crops had higher VI values than residential streetscapes; this is linked to crop irrigation.

3.1.4. Mature Trees, Access to Water, and Larger Permeable Surface Contributed to Higher VI Values

The 10 m resolution of the satellite imagery allowed for the isolation of the VI values for separate tree clusters. Eight tree clusters with the highest VI values, mostly situated in the Lubonianka district, were selected for the further analysis of their species compositions (Figure 15,

Table 3. Composition of eight tree clusters with high VI results.

Tree Cluster	Species	Trunk Circumference Class [cm]	Number of Trees
1	Pedunculate Oak	>200	1
	White Poplar	100–200	7
	White Poplar	>200	2
	White Willow	100–200	1
	White Willow	>200	1
2	Common Linden	100–200	8
3	Birch	<100	1
	Common Linden	100–200	2
	Maple	100–200	2
	Robinia Pseudoacacia	100–200	4
	Swedish Whitebeam	100–200	1
	White Poplar	100–200	1
4	Horse Chestnut	100–200	2
	Maple	<100	2
	Maple	100–200	2
	Plum	<100	1
	Robinia Pseudoacacia	100–200	2
	White Poplar	100–200	2
5	Common Linden	<100	3
	Common Linden	100–200	1
	Pine	100–200	1
6	Common Beech	<60	1
	Common Linden	<100	5
	Common Linden	100–200	1
	Hawthorn	<60	1
7	Common Beech	<100	3
	Common Linden	<60	3
	Hawthorn	<60	8
8	Common Linden	100–200	12
	Hornbeam	<20	6
	Pine	<40	3
	Pine	<100	8
	White Willow	<40	9
Sum			103

).

Among the tree clusters exhibiting high VI values, all but one comprised a mixture of different species. Notably, the most prevalent species across the eight compared clusters was the common linden (36%), which is a native species. But there also were some non-native trees, like horse chestnut, and a few specimens of invasive Robinia Pseudoacacia. An important observation is that all eight clusters comprised a considerable proportion of mature trees with a trunk circumference of over 100 cm. Moreover, the habitat quality was of key importance: Tree Clusters 1 and 8 had privileged access to water, being situated next to a ditch. Tree Clusters 1, 2, 5, and 8 had a large permeable surface area around them.

Answer to Research Question 6: The most common tree species in clusters with high VI values was common linden. It was also ascertained that better access to water, a larger green area (permeable surface) around the tree, and a greater distance from buildings contributed to the better condition of the canopy.

3.2. The Correlation of the VI Values Was Positive with the Diversity and Evenness and Negative with the Nativity

Pearson’s correlation formula was applied to assess the relationship between indices relevant to the species richness, diversity, evenness, and nativity and the health condition of the greenery reflected by the NDVI, SAVI, EVI, and LAI (

Table 4. Pearson’s correlation coefficients of selected indicators of species richness, diversity, evenness, and nativity and VI values over four-year period, rounded to three decimal places.

Indices	NDVI Mean for Canopy	SAVI Mean for Canopy	EVI Mean for Canopy	LAI (NDVI-Based)	LAI (EVI-Based)
Species Richness Index (RS)	0.633	0.627	0.670	0.538	0.627
	p = 0.027	p = 0.029	p = 0.017	p = 0.071	p = 0.029
Abundance (N)	−0.601	−0.593	−0.626	−0.504	−0.593
	p = 0.039	p = 0.042	p = 0.029	p = 0.094	p = 0.042
Biodiversity Index (RBIO)	0.628	0.621	0.662	0.532	0.621
	p = 0.029	p = 0.031	p = 0.024	p = 0.075	p = 0.031
Menhinick’s Index (RMEN)	0. 614	0.606	0.643	0.517	0.606
	p = 0.034	p = 0.037	p = 0.024	p = 0.085	p = 0.037
Simpson’s Reciprocal Index (DI)	0.681	0.677	0.734	0.588	0.677
	p = 0.015	p = 0.016	p = 0.007	p = 0.044	p = 0.016
Shannon Diversity Index (H)	0.621	0.614	0.652	0.524	0.614
	p = 0.031	p = 0.034	p = 0.022	p = 0.080	p = 0.034
Shannon Equitability Index (EH)	0.666	0.661	0.713	0.572	0.661
	p = 0.018	p = 0.019	p = 0.009	p = 0.052	p = 0.019
Proportion of native tree species (RSN)	−0.547	−0.537	−0.556	−0.448	−0.537
	p = 0.066	p = 0.072	p = 0.061	p = 0.144	p = 0.072
Proportion of native tree specimens (RNT)	−0.724	−0.722	−0.796	−0.637	−0.722
	p = 0.008	p = 0.008	p = 0.002	p = 0.026	p = 0.008

). The calculation of Simpson’s Diversity Index (D) was omitted in favor of Simpson’s Reciprocal Index (D_I) to maintain the logic that higher indices are advantageous. The correlation analysis considered mean values from different years, not their averages, to better reflect their real variation.

The correlation coefficients showed a strong positive relationship between the richness, diversity, and evenness indices and the mean NDVI, SAVI, and EVI values. The correlation was significant with a p value < 0.05. The two most strongly correlated indices were Simpson’s Reciprocal Index (D_I) and Shannon Equitability Index (E_H), and their relationship was strongest with the EVI mean values, with p value < 0.01. The LAI values derived from the NDVI had a moderate to strong correlation with the indices of the species richness, diversity, and evenness, but the p value did not allow us to reject the null hypothesis. The EVI-derived LAI values, on the other hand, exhibited a strong correlation with these indices, and it was significant, with a p value < 0.05.

Answer to Research Question 5: The tree species diversity had a strong positive correlation with the health of urban greenery as expressed by VI values.

Answer to Research Question 7: The diversity of the tree species and the evenness, expressed using Simpson’s Reciprocal Index (D_I) and the Shannon Equitability Index (E_H), were valid indicators for assessing the health of urban greenery. They exhibited the strongest relationship with the EVI.

Interestingly, the abundance and both nativity proportions showed a negative correlation with the mean VI values. The proportion of native tree species (R_SN) showed a moderate negative correlation, but the p value > 0.05 would not allow us to reject the null hypothesis. On average, the proportion of native tree specimens (RNT) showed a strong negative correlation with the mean NDVI, SAVI, EVI, and EVI-derived LAI values, statistically significant a with p value < 0.01. The correlations of all the indices with the minimum, maximum, and range of the NDVI values were weak or very weak.

Answer to Research Question 8: The nativeness of the tree species did not translate directly into good greenery health conditions (in the study areas).

4. Discussion

The results obtained can be discussed in terms of three key aspects relevant to the research objectives: (i) ascertaining the existing tree diversity and greenery health conditions in the areas covered by the study, (ii) the reasons that led to the existing situation, and (iii) the applicability of the indices related to the richness, diversity, evenness, and nativity of trees in streetscape design practices aimed at increasing the health of greenery as expressed by VI values.

4.1. Diversity and Health Conditions of Greenery in Residential Districts of Study Area

Three case studies illustrated the urban standards at different times: the 1960s, 1990s, and the 2020s. The results complement previous observations of the greenery content and ecological value of the tree communities in residential districts in the study area, both of which showed a downward trend [66]. In addition to the previously ascertained factors contributing to such a situation, including insufficient space reserved for the growth of trees or the selection of tree species with limited lifespans and canopy development capacities, new observations have now been made, indicating the low diversity of the currently created greenery complexes and their poor health status compared to older examples. All indices reflecting the species richness and diversity showed pronounced declines of between 40 and 80%. The indicator of evenness, which reflects the relative frequency of different species within a community, also showed a decrease in the range of 14–17%. An upward trend was observed only in the indicators associated with species nativeness due to the recent implementation of the greenery project in the newest case study district.

The decline in urban standards in terms of creating resilient streetscapes in residential districts, observed in this study, is consistent with research results concerning the potential for using nature-based solutions in the study area. Zwierzchowska, Haase, and Dushkova analyzed residential districts in Poznań and Berlin and ascertained that socmodernist multi-story estates had great potential for implementing nature-based solutions [91]. This is important in the context of implementing sustainable and regenerative urban planning, as nature-based solutions within green infrastructure play a significant role in urban resilience and human well-being [92].

Another study conducted in the research area measured the tree diversity in urban parks, which were found to be green spaces with high native plant richness and diversity [93]. However, it has also been noted that the condition and diversity of Polish urban forests are predominantly assessed using conventional land inventory techniques [94]. Since this approach is labor-intensive, Długoński, Wellmann, and Haase concluded that remote sensing technologies such as LiDAR should be used to assess urban forests’ status, particularly during and after disturbances or following intense droughts [94]. Our study employed both methodologies and demonstrated that older treescapes within residential districts in the study area, characterized by a higher diversity of species, attained superior VI values.

Regardless of the abundance of trees and their nativity ratio, the decrease in richness and diversity in the newer districts was correlated with a decline in the VI values. Lubonianka was the only district to maintain maximum NDVI values above or close to the 0.6 threshold corresponding to dense greenery [95,96]. The maximum SAVI values, averaged over a four-year period, were close to 0.9 for Lubonianka, corresponding to the high density of the greenery. The maximum EVI values were averaged to 0.774, signifying healthy vegetation. The LAI values for Lubonianka were not only the highest (1.5) but were also nearly equal for the two calculation methods (i.e., the NDVI-based and EVI-based methods). These results are consistent with the current state of knowledge. Many studies have indicated that plant diversity increases the productivity and stability of ecosystems [97] and that tree species diversity promotes a wide range of ecosystem functions and services [98]. Moreover, tree diversity in forestation projects protects trees and organisms, reduces the production risks, and enhances natural regeneration [99]. Although researchers in the forestry discipline have discussed the benefits and disadvantages of tree diversity [99], they mostly agree that the mechanisms of tree species selection can guarantee a favorable mix of species if they consider the patterns of coexistence of species [97,98].

It is necessary to transfer knowledge of the coexistence of tree species from forestry to the design of urban greenery, as increasing biodiversity is a key strategy for building resilience in urban landscapes [100]. The positive impact of tree species diversity on the resilience of urban greenery [101] is sometimes undermined by the introduction of alien invasive species, as urban ecosystems are hotspots for biological invasions [102]. Although researchers have found evidence of both disadvantages [103,104] and benefits [105] of introducing non-native plants, the idea of the ecological advantage of native species has been consolidated in both science and public awareness [106], even though not all non-native species are invasive. Native trees have proven beneficial in providing vital urban ecosystem services and supporting biodiversity [66,106,107,108]. However, their capacity to withstand urban stress is still being researched and discussed [109], and their adaptation to climate change has been identified as one of the greatest limitations in the application of nature-based solutions in cities [110]. On the other hand, there is a risk that a species planted for its high resistance to urban stress may become an invasive species in the future [111] and will spread from urban areas into the natural environment [112] at an increasing rate after reaching a critical point [113].

In this study, a negative correlation was observed between the nativity ratio of trees and the VI values. This situation is a result of complex factors, including the young age of the example district with 100% native trees (Grafitowe) and the dominant selection of grafted and trimmed trees within this district. This phenomenon can also be partially attributed to design factors, specifically the presence of impervious surfaces in proximity to the trees, which are known to inhibit tree growth [47]. An important aspect that was not addressed in this study and should be incorporated in future research is the consideration of the soil type and water availability. Nonetheless, the observed situation indicates that nativity does not guarantee healthy urban greenery. This is an important finding in the context of the aim of this work, which was to identify which tree selections performed better in streetscapes in the study area. Further analysis is necessary to determine the causes of the current situation and identify directions for improvement.

4.2. Reasons That Led to the Decline in the Diversity and Health of Greenery in the Study Area

The decline in the diversity and health of greenery in residential districts is a direct result of poor spatial planning, as is the worldwide decline in biodiversity [114]. Poland’s planning framework sets the minimum proportion of the biologically active area to ensure natural vegetation and rainwater retention. Of the thirteen types of development zones, ten (including residential districts) require a minimum of 30% biologically active area [115]. However, this law does not define the proportion of specific types of greenery, such as parks and urban forests, in residential districts. Previous studies have defined 40% forest cover as the optimal level for maintaining biodiversity in human-modified landscapes [116]. Similar standards have been implemented in Polish cities since the 1960s, as exemplified in this study by the Lubonianka estate; however, after the political transformation, they were abandoned in favor of free competition in the real estate market. The Cegielskiego estate is an example from this difficult transformation period. Its planning was founded on previous socialist standards, with more space allocated to tall greenery, but the treescape was not solicitously developed and its creation eventually consisted of spontaneous plantings by residents, with a significant proportion of self-seeders.

Currently, standards for the availability of public and recreational greenery are established for urbanized areas along policies on locating basic services in residential areas [117]. Polish cities, e.g., Warsaw or Poznań [118,119], define their strategies for residential development focusing most closely on social issues such as the availability of flats, ensuring the satisfaction of the needs of the aging society, safety, and the availability of services and recreational greenery [120,121]. However, such policies are inefficient in solving streetscape planning problems, particularly in suburban fringes. The dominant factor in the real estate sector is the return on investment, and greenery is treated as a factor that increases the market value of apartments and plots [122,123]. It has become common practice to locate densely built-up estates, without their own landscaping, around existing greenery resources (e.g., state forests) to meet the greenery availability criteria. Some new housing districts, as exemplified by the Grafitowe estate, adhere to planning requirements regarding the proportion of green areas, but the design of their streetscapes is driven by economic factors. Tree species selection is determined based on esthetic, security, and insurance issues (e.g., avoiding the growth of tree branches and roots that could endanger buildings) and minimizing maintenance costs (e.g., pruning and leaf removal). It has also become standard to limit the diversity of species selected. The repetitive catalog architecture is accompanied by minimal greenery, which is neat and economical. Despite the growing share of green areas on urbanized land (e.g., lawns, greenery in housing estates, and street greenery) [124], their ecological value is low because of the above-mentioned streetscape design decision criteria. As an effect of such choices, monocultural landscapes have been created, scoring low in terms of ecosystem resilience. Another reason for the decrease in the ecological value of urban greenery is excessive maintenance. As Borysiak, Breuste, and Mizgajski proved, excessively manicured greenery has a negative effect on the plant species richness and diversity [93].

An effective solution to this problem requires a comprehensive approach to planning streetscapes and common green areas in residential districts. A strategy for introducing biodiversity to cities has been promoted in Poland by the Ministry of Culture and National Heritage [125], but it applies to public investments. Private-sector projects are not obligated to implement these guidelines. This can be improved by including them in local plans. Along with the implementation of a participatory planning model, local plans can be complemented by an alternative-based approach to estimate the potential of residential landscapes to provide urban ecosystem services of high priority to diverse stakeholders [126,127].

4.3. Applicability of Tested Indices for Assessing State of Urban Greenery and Informing Design

The usability of richness and diversity indicators to quantify and analyze ecosystem compositions has been widely debated by scientists, especially in forest structure research. Some studies have shown the inaccuracy of using the concept of diversity when analyzing entropy among tree size classes; however, they have also shown that measuring the equitability based on individual trees yields more satisfactory results [128]. It has also been stated that the use of the Shannon Diversity Index (H) should be discontinued when assessing biodiversity due to calculation bias [129]. On the other hand, evenness measures (the Shannon Equitability or Pielou’s Evenness) have been proven to be reliable instruments for quantifying the relative evenness, which can be possible even with an incomplete sample [79]. A recent review that summarized seventeen diversity indices and discussed their results mentioned the Shannon Equitability as one of eight robust indices for comparing biodiversity in ecosystems [71].

Diversity indices have shown a positive relationship with VI values, reflecting the health of greenery. Studies conducted in diverse regions worldwide have validated this correlation [130,131]. In a manner consistent with the findings of this study, it has been observed that high tree diversity is correlated with elevated values of VIs [132,133]. Furthermore, a positive correlation between NDVI, SAVI, and EVI values and diversity has been established [134]. Among these indices, the EVI has demonstrated the strongest correlation with the floristic diversity [135]. The correlation between the plant species diversity indices and mean VI values ascertained in this study was the reverse of that found for the direct application of remote sensing for assessing the plant species diversity [130]. As demonstrated by the current state of knowledge, the mean values of the NDVI and EVI, along with their standard deviations, can serve as predictors of the plant species diversity [55,131]. Several existing studies have substantiated the scientific validity of utilizing remotely sensed data for predicting and modeling the plant diversity across various environments [45,136,137]. This study employed this correlation in a reversed way to underscore the significance of the tree species diversity for the vitality of greenery in residential areas. The two indices most strongly correlated with the VI values, namely Simpson’s Reciprocal Index (D_I) and the Shannon Equitability Index (E_H), could be integral components of urban forest analysis. These indices are valuable both in evaluating the current state of greenery and in assessing project variants, thereby informing and supporting the decision-making process. Future research could determine the optimal levels of these indices for greenery projects in urbanized areas.

The use of diversity indices to inform landscape design decisions presented in this study is a novel approach. The potential applications of this method are categorized into two primary areas: the ecological evaluation of urban greenery and preparatory research for design purposes. In the first application, the method facilitates a comprehensive, multi-criteria assessment of existing urban forests (UFs). In the second application, it enables local pre-design analyses to be conducted, which gather information and identify the most effective tall greenery complexes in a specific region, thereby informing design practices. The decisive factors contributing to the application value of the proposed method are its simplicity and the availability of data (incl. satellite imagery). Tree inventories in situ are routinely performed as part of pre-design preparatory studies, but the acquired data are not used efficiently in current practice. This study showed that performing additional calculations relating to the diversity and evenness of tree species can inform landscape design and decision-making processes to achieve more resilient streetscapes. It has been proven that the diversity of tree species, the largest plants in urban spaces, has an impact on the health of greenery and, thus, on the resilience of the urban environment. Further research should answer the question regarding the threshold at which we can observe a positive effect of diversity on VI values.

The research conducted yielded specific practical recommendations to support sustainable urban development. These recommendations pertain to the formulation of a tiered approach for the assessment and design of urban treescapes. Given the increased availability of remotely sensed data and the potential for automating the calculation of Vegetation Indices (VIs), while acknowledging the labor-intensive nature of on-site inventories, it is advisable to reverse the sequence of stages employed in this study, as follows:

(1)

Calculate the Vegetation Indices (VIs) for the study area utilizing remotely sensed spectral data;

(2)

Identify patches of urban forestry with diverse VI values for further analysis, giving special attention to the tree canopies with the highest mean VI values;

(3)

Perform on-site inventories of the identified exemplary greenery patches;

(4)

Calculate indices relevant to the species richness, diversity, evenness, and nativity;

(5)

Create recommendations for local greenery planning.

The limitations of this study are primarily attributable to its geographical constraints and scale. Due to the inherent variability in the environmental, soil, and climatic conditions, the results obtained may be specific to a particular region and not universally applicable. It is essential to apply the method locally to minimize the potential errors. The regional analysis of tree health in urban environments is necessary to inform planting selection. This study faced challenges in generalizing the findings from empirical research conducted in a limited area. The further analysis of additional examples is needed to confirm the observed correlations. Additionally, this study did not address the issues of urbanization on a regional scale and the socioeconomic impacts that cause stress to urban greenery. Incorporating these aspects would help differentiate the impact of the tree diversity from that of environmental and social factors. Finally, data availability can be a limiting factor. While satellite imagery is widely accessible, conducting a tree inventory is a time-consuming and labor-intensive process. A publicly accessible system could be developed to collect data from inventories conducted by various stakeholders during preparatory pre-design site analyses.

5. Conclusions

The novelty of this study lay in the development of a comprehensive scheme for evaluating streetscapes in terms of their tree species richness, diversity, evenness, and nativity and the extent to which these affected the health status of the greenery. A tree inventory was the basis for calculating indicators relevant to the species richness, diversity, evenness, and nativity rates, which we followed by demonstrating their correlation with the health status of greenery expressed by VI values. The goal was to inform streetscape design decisions in the study area.

The results showed the following:

(1)

The oldest residential district in the comparison had the most diverse community of trees, while the newest one had the highest degree of tree nativity.

(2)

The tree species diversity and evenness were positively correlated with the mean Vegetation Index (VI) values, reflecting the good health of the greenery, and the oldest of the three compared districts ranked first in terms of the VI values. The current model of suburban planning fails to deliver a healthy treescape.

(3)

The two indices most strongly correlated with the VI values, i.e., Simpson’s Reciprocal Index (D_I) and the Shannon Equitability Index (E_H), can be used to assess the health potential of urban greenery, both existing and projected.

(4)

The Enhanced Vegetation Index (EVI) was the most precise in detecting the condition of the tree canopies in the study area, while the Normalized Difference Vegetation Index (NDVI) and the Soil-Adjusted Vegetation Index (SAVI) were more affected by seasonal droughts and yielded a more general image of the state of the greenery in the residential districts compared.

(5)

The taxonomic nativeness of the tree species had a negative correlation with the mean NDVI, SAVI, EVI, and LAI values and did not guarantee good greenery health.

(6)

The most common tree species in the clusters with high VI values was common linden. However, the habitat quality was of key importance, including the access to water and amount of free space around the trees.

In relation to the objective of this study, which was to establish guidelines for redirecting suburban planning towards sustainable urbanism, it is essential to differentiate between general and local recommendations. The general conclusions facilitated the development of a tiered framework proposal for assessing the status of urban greenery, as presented in the Discussion (see Section 4.3). The recommendation for the local streetscape designers that can be formulated at this stage of the research is to increase the diversity of the selected trees and to focus on adapting the selected tree species to the environmental conditions. Determining the target minimum diversity level requires further research; however, it can be noted that the best result in the comparison made in this study was a Simpson’s Reciprocal Index (D_I) equal to 8.5 (for the example of Lubonianka).

This study has also shown that high VI results were locally achieved for the common linden species. Nonetheless, its popularization would be inappropriate without considering other factors, such as the allergenicity of the species to humans. Another recommendation suggests that when selecting tree species, primary consideration should be given to their climatic and soil requirements, considering the conditions of the urban environment. It is advisable to select species that exhibit the greatest capacity for survival and produce substantial canopies. The criterion of nativeness should not be considered a limiting factor in this selection process; however, the implementation of anti-invasion selection rules must be prioritized.

Further research should strive to determine the optimal level of the diversity and evenness indicators to ensure good health conditions for greenery, as well as to propose exemplary plant selections based on studies of the mutual relationships between species and their influence on human living conditions.