Characteristics of Planting Structures in Public-Type Private Gardens in Urban Areas of South Korea

Abstract

This study analyzed the planting characteristics and spatial patterns of public-type private gardens in urban areas. Five gardens in Daejeon and Ulsan were surveyed using quadrats to record tree locations and sizes and were digitized for layout mapping. Planting and analysis units were defined, and spatial patterns were examined using degree centrality. The gardens were classified into one site under mixed artificial–natural management and four sites under artificial management with commercial linkage. The mixed site featured both canopy and shrub layers, with spontaneous vegetation surrounding Pinus thunbergii, Pinus densiflora, and Prunus yedoensis. The commercial sites included either canopy-only or canopy-shrub structures. Lagerstroemia indica, P. densiflora, and Euonymus japonicus. were predominant in the temperate central region, while P. densiflora and Diospyros kaki. dominated in the southern region. This study identified the potential of public-type private gardens as planting models and their capacity to contribute to urban environmental improvement.

1. Introduction

Gardens are increasingly recognized not merely as places where plants are cultivated and maintained, but as key socio-ecological spaces in which people and nature interact directly. Santos et al. [1] define garden as “a space designed for the cultivation, display, and appreciation of plants,” highlighting the garden as a cultural apparatus through which humans organize and experience nature. Similarly, Cameron et al. [2] conceptualize residential gardens as green spaces adjacent to dwellings where residents exercise management autonomy, distinguishing them from parks and shared green spaces, and positioning them as core components of urban green infrastructure.

Aligned with these conceptual attributes, gardens perform multiple functions. Most notably, they play a central role in conserving urban biodiversity. In several countries including the United Kingdom, Sweden, and New Zealand, gardens account for approximately 16–47% of urban green space and provide habitat for a wide range of taxa, including insects, birds, and amphibians [3]. Beyond biodiversity conservation, gardens deliver diverse ecosystem services. Cameron et al. [2] have noted that gardens contribute to climate regulation, hydrological regulation, carbon sequestration, and soil conservation, while Cabral et al. [4] further identify the urban heat-island mitigation, water-quality improvement, and even food production as relevant provisioning and regulating services. Gardens also carry significant cultural value.

Moreover, gardens contribute to human health. A meta-analysis by Soga et al. [5] has shown that gardening reduces depressive symptoms and anxiety while improving life satisfaction, and experimental evidence reported by Joubert et al. [6] has demonstrated that horticultural therapy yields measurable improvements in mental health. In sum, within urban environments, gardens function as complex spaces that extend well beyond “greenness,” supporting ecological processes, cultural meanings, and human well-being.

These multilayered functions have positioned the gardens not merely as a physical site but as a cultural phenomenon, a status borne out by diverse cases in contemporary society. A representative example is the United Kingdom’s RHS Chelsea Flower Show. This international exhibition showcases diverse garden typologies including show gardens, sanctuary gardens, balcony gardens, and container gardens and, by synthesizing aesthetic values with the latest design trends, has helped to disseminate and elevate the sociocultural status of gardens worldwide.

Similarly, to developments in the United Kingdom, interest in the public value and cultural role of gardens has been steadily growing in Korea. Notably, the 2010 Gyeonggi Garden Expo, the first garden exposition organized at the municipal level, marked an important starting point for public engagement with garden culture and its diffusion. Building on this, the 2013 Suncheon Bay International Garden Expo, the country’s first international garden exposition, broadened public awareness of garden culture and reframed the public functions and cultural meanings that gardens may assume within the urban landscape.

Institutional efforts to strengthen the public character of gardens continue to emerge internationally. The United Kingdom’s National Garden Scheme (NGS) temporarily opens private domestic gardens to the public and donates the proceeds from admission fees to social welfare, thereby exemplifying how private property can assume a public function. Comparable initiatives include France’s Jardins familiaux and the United States’ Community Gardens Act.

In particular, in Korea, the 2015 amendment to the Act on the Creation and Promotion of Arboreta and Gardens established a statutory framework classifying gardens as national, local, and private, and provided the institutional basis for registering qualifying private gardens as “public-type private gardens.” Registration criteria include, inter alia, a green area ratio of at least 40% of the total site area and, for large-scale gardens, the provision of professional staff and public convenience facilities [7].

In tandem with these developments, the Seoul Metropolitan Government enacted the Ordinance on the Creation and Promotion of Garden Culture in 2019. Subsequent revisions to Article 10 have expanded and refined support mechanisms: In 2024, a new clause authorized partial subsidies to defray expenses for plant conservation and propagation as well as facility management in private gardens recognized as having conservation value; in July 2025, the ordinance was amended to provide that the Mayor shall endeavor to revitalize private gardens and may, pursuant to Article 18-6(3) of the Act on the Creation and Promotion of Arboreta and Gardens, encourage their general opening to the public [8]. Public-type private gardens thus differ from conventional urban green space categories in that they retain the attributes of private property while guaranteeing public accessibility, endowing them with dual character.

Ultimately, the institutional adoption of public-type private gardens accords with a national vision that understands gardens as everyday nature and as spaces of care and healing that enhance quality of life. The government of the Republic of Korea has articulated the vision of “gardens accessible anytime, anywhere” and, through the Second Basic Plan for Garden Promotion (2021–2025), has pursued the promotion of garden culture and improved public access [9]. Within this policy trajectory, the recent registrations of public-type private gardens provide concrete cases of gardens functioning as spaces of public value.

While previous research has explored garden-related themes such as garden culture [10,11,12,13,14], garden design and planting analysis [15,16,17,18,19,20], garden creation and management [21,22,23,24,25], the concept of gardens [26,27], and ecosystem services provided by gardens [28], few studies have directly addressed the role of gardens as public-value spaces under legal frameworks. As exceptional private gardens are now being registered as public-type private gardens, research on their planting structure characteristics is increasingly important. These exemplary gardens can serve as reference models for planting in typical gardens and provide the foundational data for the establishment of a management plan for public-oriented gardens. Therefore, this study aimed to analyze the planting structure characteristics of public-type private gardens established in urban areas.

2. Methods

2.1. Research Performance System

Figure 1 presents the research framework. The objective of this study was to analyze the planting characteristics and spatial patterns of private gardens located in urban areas and were registered with the Korea Forest Service. The research process consisted of site selection, the establishment of planting and analysis units, and the derivation of patterns through network analysis. As of January 2024, 131 registered private gardens were examined using QGIS to identify those located in urban zones, resulting in 34 candidates. A land cover analysis with a 250 m buffer was then conducted, and five gardens with an urbanization ratio of over 50% were selected as the final study sites. At each site, a 20 m × 20 m quadrat was established to survey the location and characteristics of trees, including height, diameter at breast height (DBH), height to the first branch, and crown width. These data served as the foundation for the analysis of planting structure. A planting unit was defined as a planting section based on movement pathways. To identify the planting structure of each section, dominant composition was determined using a threshold of 30% for trees and herbaceous plants, and 10% for shrubs. Analysis units were determined based on urban park planting density standards and weighting criteria from the Ministry of Land, Infrastructure and Transport [29]. Type 1 (2.5 m radius) was applied to trees with a height of 3–4 m, and Type 2 (3.6 m radius) was applied to trees 5 m or taller. Network analysis using Gephi 0.10 was conducted to identify the top three species in terms of connectivity centrality, thereby revealing the structural characteristics of the planting patterns.

2.2. Site Selection

The spatial scope of this study included 131 private gardens registered with the Korea Forest Service as of January 2024. In this study, we sought to select public-type private gardens located in urban areas, where public accessibility is higher than in suburban gardens. Using QGIS and the 2024 V World land-use data [30], 34 gardens located within urban zones were identified. After excluding Jeju Island, land cover [31] ratios were calculated for the remaining 33 gardens using a 250 m buffer. Surrounding land uses were classified into two categories: urbanized areas (residential, commercial, industrial, cultural, recreational, transportation, public facilities, and artificial bare land) and green areas (artificial grasslands, orchards, rice paddies, fields, coniferous forests, broad leaved forests, mixed forests, and other cultivated land). Among gardens with surrounding urbanized land exceeding 50%, five sites were selected through field verification. The final study sites included three gardens in Daejeon Metropolitan City, Pine Landscape (No. 2), Walden (No. 3), and There (No. 5), and two in Ulsan Metropolitan City, Guam Garden (No. 2) and Nowije (No. 6), as shown in Figure 2.

Through the Garden Nuri homepage [32], basic information regarding the location, area, planting composition, and management characteristics of the selected gardens was reviewed. In addition, preliminary interviews with the owners were conducted to secure approval for on-site investigations and to supplement information on unique management practices. As a result, among the five gardens surveyed, Guam Garden exhibited a distinct difference from the others. The garden was equipped with a greywater recycling system and solar power facilities, enabling self-sufficiency in energy, water, and soil nutrients required for its maintenance. Furthermore, in terms of vegetation, seeds naturally dispersed by birds and wind were found to germinate and coexist in competition with pre-planted species, thereby maintaining a spontaneous and naturalistic landscape. Based on these characteristics, Guam Garden was classified as an Artificial and Natural Mixed Management Type, whereas the other four gardens, which were subject to continuous and intensive management linked to commercial facilities, were classified as a Commercial-Linked Artificial Management Type. However, it should be noted that the management approach of Guam Garden represented the only such case among the 131 officially registered private gardens, and the acquisition of comparable additional samples was limited.

2.3. Survey and Analysis Methods

(1)

Field Survey

The five study sites are located in the metropolitan cities of Daejeon and Ulsan and, based on the warmth index (WI), are classified into the central temperate zone (WI 85–100; distribution of deciduous broad-leaved forests) and the southern temperate zone (WI > 100; distribution of warm-temperate evergreen broad-leaved forests). Given the close association between this bioclimatic zonation and vegetation characteristics, it was used to contextualize the environmental setting of the sites. The field survey was conducted from May to July 2024. At each site, a quadrat measuring 20 m × 20 m was established. The location of planted trees and their figures were recorded. The recorded data included tree height, DBH, height to the first branch, and crown width. Based on the surveyed planting data, the data were digitized using AutoCAD 2025.

(2)

Analysis Methods

This study employed a network-based analytical approach to examine the planting structure of five registered private gardens. A planting unit was defined as a spatial segment delineated by movement pathways in the site plan. To evaluate the structural composition, dominance thresholds were applied: 30% for trees and herbaceous plants and 10% for shrubs. For shrubs, the 10% threshold corresponded to the garden shrub density standard of 0.099 plants/m², as defined by the Urban Park Planting Density Standard [29]. To analyze plant relationships, analysis units were defined based on weighted tree spacing: a 3.6 m radius was applied to trees 5 m or taller, and a 2.5 m radius was applied to trees ranging from 3 to 4 m in height.

Plant network structures were visualized and analyzed using Gephi 0.10, an open-source graph analysis tool. This method followed the framework established by Lee et al. [33] for urban vegetation networks and adopted the degree centrality approach proposed by Park [15] in her analysis of Gertrude Jekyll’s Wild Garden. Centrality analysis is a fundamental concept in network analysis that quantifies the importance of individual nodes within an overall network. Freeman and Scott [34] proposed three key centrality measures: degree centrality, closeness centrality, and betweenness centrality, each defined by the formulas shown in

Table 1. Network centrality analysis.

Degree Centrality	Closeness Centrality	Betweenness Centrality
${\displaystyle \frac{d_i}{n-1}}$	$C^{\prime }c\left(p_k\right)={\displaystyle \frac{n-1}{\sum _{i=1}^{n}d\left(p_i,p_k\right)}}$	$C_B^{\prime }={\displaystyle \sum _{j,k}}{\displaystyle \frac{g_{j,k}^{\left(i\right)}}{g_{j,k}}}$ $\left(j\ne k\ne i,\hspace{0.25em}j,k,i\in N\right)$
$d_i$ = Absolute degree of connection centrality $n$ = Total number of nodes in the network	$d\left(p_i,p_k\right)$ = represents the number of nodes that are connected to both node $p_i$ and node $p_k$.	$g_{j,k}$: Number of shortest paths between node j and node k $g_{j,k}^{\left(i\right)}$: Number of shortest paths between two nodes j and k that pass-through node I

. Degree centrality represents the number of direct connections a node has, indicating the strength of local interactions. Closeness centrality measures the average shortest path length from a node to all others, reflecting the efficiency of information dissemination. Betweenness centrality quantifies how often a node appears on the shortest paths between other nodes, representing its role as an intermediary [34]. Using these centrality measures, this study analyzed plant patterns and relationships in public-type private gardens within urban areas, identifying the top three species by degree centrality.

3. Results

3.1. Artificial and Natural Mixed Management Type

(1)

Planting Status

A survey of the planting status in Guam Garden revealed the presence of 28 tree species in the canopy layer, 24 species in the subcanopy layer, and 26 species in the shrub layer, as shown in

Table 2. Planting structure of Guam Garden.

Planting Level	Observed Species	Number of Species Appearing	Planting Density
Canopy layer	P. thunbergii, P. densiflora, Trachycarpus fortune, etc.	28	0.224 tree/m2
Subcanopy layer	Camellia japonica, Prunus serrulata, Fatsia japonica, Photinia glabra, etc.	24
Shrub layer	Ilex crenata, Rhododendron(spp.), Nandina domestica, etc.	26	0.179 m2/m2
Herb layer	Polygonatum odoratum (Mill.) Druce var. pluriflorum, Hosta(spp.), etc.	10	-

. In particular, across the combined canopy and subcanopy layers, a total of 41 species were recorded; of these, 24 were spontaneously germinated and 17 were planted. This high proportion of spontaneously germinated species, a pattern not observed in the other gardens, indicates that the site differs from the others in terms of species composition. The average height of the canopy layer was 4 ± 1 m, with an average diameter at DBH of 18 ± 11 cm. For the subcanopy layer, the average height was 3 ± 2 m, and the DBH was 7 ± 5 cm.

Analysis of the planting density for woody plants showed that the canopy and subcanopy layers had a density of 0.224 tree/m², which, compared with the other four gardens, was the highest (Pine landscape Garden: 0.109 tree/m²; Walden Garden: 0.222 tree/m²; There Garden: 0.123 tree/m²; Nowije Garden: 0.202 tree/m²). Meanwhile, the shrub layer had a density of 0.179 shrubs/m². In addition, the total area of planting units was 558 m², and the combined planted area across tree, shrub, and herbaceous layers was 630 m², corresponding to 113% of the total site area; this reflects the inclusion of the projected planting area resulting from tree canopy spread. Spatial analysis based on Guam Garden’s planting units divided the area into six distinct spaces, and the planting structure was identified as a combination of trees and shrubs, as illustrated in Figure 3.

(2)

Network Centrality Analysis

In Guam Garden, a total of 19 analysis units were used to generate data on plant relationships, consisting of 15 units of Type 1 and 4 units of Type 2. The results of centrality calculations, including degree centrality, closeness centrality, and betweenness centrality, are presented in

Table 3. Network centrality analysis of Guam Garden.

Id	Degree	Degree Centrality	Closeness Centrality	Betweenness Centrality
P. thunbergii	31	0.62	0.73	209
P. densiflora	29	0.58	0.61	160
P. yedoensis	7	0.14	0.45	39
O. fragrans	7	0.14	0.41	6
Metasequoia glyptostroboides	6	0.12	0.34	0
Ternstroemia gymnanthera	6	0.12	0.45	38

. The spatial distribution of planting patterns was visualized in Figure 4. In the visualization, the size of each node indicates its degree of connectivity, and the groups of trees, shrubs, and vines were distinguished by color, following the method presented by Park in 2019. The centrality analysis revealed that the plant network of Guam Garden consisted of 51 nodes and 88 edges. Among the species with identical degree centrality values, P. thunbergii, P. densiflora, P. yedoensis, and Osmanthus fragrans were ranked in order of descending degree centrality. P. thunbergii had the highest level of connectivity, with 31 connections and a degree centrality value of 0.62. This was followed by P. densiflora, which had 29 connections and a degree centrality value of 0.58. Both P. yedoensis and O. fragrans had 7 connections, corresponding to a degree centrality value of 0.14. Unlike the other private gardens, Guam Garden exhibited a structural pattern connecting planted species with naturally germinated species. Around the primary species P. thunbergii and P. densiflora, connections with surrounding species such as Albizia julibrissin, Viburnum carlesii, and Eriobotrya japonica were identified. P. thunbergii showed high values not only in degree centrality but also in closeness centrality and betweenness centrality. This indicates that the species is not only well connected within the network but also plays a critical role in facilitating the overall flow of interactions. These findings suggest that P. thunbergii served as a key mediator in the network, enhancing the potential for connectivity among other species.

3.2. Commercial-Linked Artificial Management Type

(1)

Planting status

According to the analysis of planting at the four commercially integrated and artificially managed sites, as shown in

Table 4. Planting structure of gardens under commercial-linked artificial management.

Site	Planting Level	Observed Species	Number of Species Appearing	Planting Density
Pine landscape Garden	Canopy layer	Acer palmatum, P. densiflora, Juniperus chinensis, Malus pumila, etc.	11	0.109 tree/m2
	Subcanopy layer	P. densiflora, P. densiflora for. Multicaulis	2
	Shrub layer	Taxus cuspidata, Rhododendron (spp.), Buxus sinica, etc.	11	0.211 m2/m2
	Herb layer	Hosta longipes, Ajuga reptans, etc.	6	-
Walden Garden	Canopy layer	Aesculus turbinata	1	0.222 tree/m2
	Subcanopy layer	Euonymus (spp.), Platycladus orientalis, Photinia glabra	12
	Shrub layer	Hydrangea (spp.), Phyllostachys reticulata, Rosa (spp.)	3	0.141 m2/m2
	Herb layer	Liriope muscari, Physostegia virginiana, etc.	6	-
There Garden	Canopy layer	A. palmatum, P. serrulata, P. densiflora, etc.	15	0.123 tree/m2
	Subcanopy layer	L. indica, Hydrangea (spp.), etc.	11
	Shrub layer	E. japonicus, Rhododendron (spp.), Rosa (spp.), etc.	11	0.153 m2/m2
	Herb layer	Liriope muscari, Chrysanthemum (spp.), Hosta longipes, etc.	25	-
Nowije Garden	Canopy layer	P. thunbergii, P. densiflora, M. glyptostroboides, etc.	48	0.202 tree/m2
	Subcanopy layer	P. densiflora, T. cuspidata, P. orientalis, etc.	31
	Shrub layer	Rhododendron (spp.), B. sinica, N. domestica, Hydrangea (spp.), etc.	52	0.334 m2/m2
	Herb layer	Liriope muscari, Hosta longipes, etc.	22	-

, the site named “Pine Landscape” included 11 species in the canopy layer, 2 species in the subcanopy layer, and 11 species in the shrub layer. The average height of the canopy layer was 3 ± 1 m, with an average diameter at breast height (DBH) of 39 ± 21 cm. The planting density of woody species was 0.109 tree/m² for the canopy and subcanopy layers and 0.211 shrubs/m² for the shrub layer. The total area of planting units was 303 m², and the actual planted area was 153 m², corresponding to 50% of the total site area. Based on the planting units, the site was divided into seven spatial zones, and the planting structure was classified as the “Other type”, characterized by paving and lawn, as illustrated in Figure 5.

In the site called “Walden,” the vegetation consisted of 1 species in the canopy layer, 12 species in the subcanopy layer, and 3 species in the shrub layer. The canopy layer had a height of 8 m and a DBH of 62 cm. For the subcanopy layer, the average height was 2 ± 1 m, and the average DBH was 24 ± 15 cm. The planting density was 0.222 tree/m² for the canopy and subcanopy layers, and 0.141 shrubs/m² for the shrub layer. At Walden Garden, the total area of planting units was 180 m², and the actual planted area was 81 m², representing 45% of the total site area. The site was divided into two spatial zones, and the planting structure was classified as the “Tree type.”

At the site named “There,” the planting composition included 15 species in the canopy layer, 11 species in the subcanopy layer, and 11 species in the shrub layer. The average height of the canopy layer was 6 ± 4 m, and the average DBH was 61 ± 35 cm. The subcanopy layer had an average height of 2 m and a DBH of 17 cm. The planting density was 0.147 tree/m² for the canopy and subcanopy layers and 0.183 shrubs/m² for the shrub layer. The total area of planting units was 665 m², and the planted area was 1072 m², corresponding to 161% of the total site area. Six planting units were identified through spatial analysis, and the structure was categorized as the “Tree and Herb” type.

In the site called “Nowije,” 48 species were identified in the canopy layer, 31 species in the subcanopy layer, and 52 species in the shrub layer. The average height of the canopy layer was 4 ± 1 m, and the average DBH was 18 ± 11 cm. The subcanopy layer had an average height of 2 m and a DBH of 9 ± 6 cm. The planting density was 0.202 tree/m² for the canopy and subcanopy layers and 0.334 shrubs/m² for the shrub layer. At Nowije, the total area of planting units was 1170 m², the largest among the sites, and the actual planted area was 997 m², representing 84% of the total site area. Sixteen planting units were identified, and the planting structure was found to correspond to the “Tree and Shrub” type.

(2)

Network Centrality Analysis

The four commercially linked and artificially managed sites exhibited variations in species distribution according to climate zones. Three sites located in Daejeon belong to the temperate central climatic zone, while one site in Ulsan is classified within the temperate southern climatic zone. A network centrality analysis was conducted to examine interspecies relationships within the planting compositions, as shown in

Table 5. Analysis of network centrality in gardens under commercial-linked artificial management.

Site	Pine Landscape				Walden				There				Integration (Daejeon Area)				Nowije
ID	1	2	3	4	1	2	3	4	1	2	3	4	1	2	3	4	1	2	3	4
D. kaki	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	43	0.41	0.59	676
Diospyros lotus	-	-	-	-	-	-	-	-	7	0.24	0.87	0	-	-	-	-	-	-	-
P. thunbergii	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	43	0.41	0.55	406
A. palmatum	-	-	-	-	-	-	-	-	15	0.52	0.56	59	16	0.34	0.43	130	25	0.24	0.44	225
C. japonica	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	17	0.16	0.33	160
Pinus rigida	-	-	-	-	-	-	-	-	6	0.20	0.4	12	-	-	-	-	-	-	-
Prunus mume	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	18	0.17	0.49	45
Pseudocydonia sinensis	-	-	-	-	7	0.50	0.80	10	-	-	-	-	-	-	-	-	15	0.14	0.43	114
L. indica	9	0.69	0.83	8	-	-	-	-	8	0.27	0.63	32	17	0.36	0.51	261	19	0.18	0.41	157
M. denudata	-	-	-	-	-	-	-	-	13	0.45	1	24	13	0.28	0.37	55	-	-	-	-
Pruns serrulata	-	-	-	-	-	-	-	-	13	0.45	0.78	15	13	0.28	0.43	35	20	0.19	0.44	175
Styphnolobium japonicum	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	20	0.19	0.46	0
P. granatum	6	0.46	0.56	0	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
P. densiflora	4	0.30	1	0.5	-	-	-	-	14	0.48	0.49	54	17	0.36	0.42	149	54	0.52	0.60	936
P. orientalis	-	-	-	-	10	0.71	0.73	24	-	-	-	-	-	-	-	-	28	0.27	0.42	581
J. chinensis	10	0.77	0.77	14	-	-	-	-	-	-	-	-	12	0.26	0.44	124	-	-	-	-
E. japonicus	-	-	-	-	5	0.36	0.47	10	-	-	-	-	13	0.28	0.24	125	-	-	-	-
Rhododendron schlippenbachii	6	0.46	0	0	-	-	-	-	8	0.27	0	0	-	-	-	-	-	-	-	-
B. sinica	5	0.38	0	0	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-

Note: ID—Degree (1), Degree centrality (2), Closeness centrality (3), Betweenness centrality (4).

and illustrated in Figure 6. The centrality indices including degree centrality, closeness centrality, and betweenness centrality were calculated for each site.

At Pine Landscape Garden, seven planting units were analyzed, consisting of two units of Type 1 and five units of Type 2. Based on these units, a species interaction network was derived, comprising 14 nodes and 29 edges. Species with the highest degree centrality included J. chinensis, L. indica, Punica granatum, and Rhododendron (spp.). J. chinensis exhibited the highest connectivity, with a degree of 10 and a degree centrality of 0.77, functioning as a key species with active interactions throughout the network. L. indica followed with a degree of 9 and a degree centrality of 0.69. This species also recorded high closeness centrality, indicating its strong influence on the disseminating of interactions. P. granatum and Rhododendron (spp.) each had a degree of 6 and a degree centrality of 0.46. Both J. chinensis and L. indica showed high betweenness centrality, emphasizing their role as bridge species within the network.

At Walden Garden, five planting units were analyzed, consisting of one unit of Type 1 and four units of Type 2. The resulting network was composed of 15 nodes and 21 edges. Species with the highest degree centrality included Thuja orientalis, Chaenomeles sinensis, and E. japonicus. T. orientalis recorded the highest connectivity with a degree of 10 and a degree centrality of 0.71, and it also showed high betweenness centrality, indicating its role as an intermediary facilitating accessibility within the network. C. sinensis followed with a degree of 7 and a degree centrality of 0.50, and E. japonicus ranked third with a degree of 5 and a centrality of 0.36.

At There Garden, six planting units were analyzed, consisting of thirteen units of Type 1 and six units of Type 2. The network was composed of 30 nodes and 63 edges. Dominant species included A. palmatum, P. densiflora, Magnolia denudata, and P. yedoensis, A. palmatum had the highest connectivity, with a degree of 15 and a degree centrality of 0.52, indicating its central role within the network. P. densiflora followed with a degree of 14 and a centrality of 0.48. M. denudata and P. yedoensis each had a degree of 13 and a centrality of 0.45. M. denudata additionally recorded high closeness centrality, reflecting its wide influence across the network. Rhododendron (spp.) were found to be connected with all canopy species, indicating a distinct planting structure that combines canopy and shrub layers. A synthesis of the gardens located in the temperate central climatic zone revealed an integrated network composed of 48 nodes and 109 edges. L. indica and P. densiflora both recorded the highest connectivity, each with a degree of 17 and a degree centrality of 0.36. A. palmatum followed with a degree of 16 and a centrality of 0.34. E. japonicus, M. denudata, and P. yedoensis each had a degree of 13 and a centrality of 0.28. L. indica was particularly notable as a species that consistently appeared across multiple sites, underscoring its ecological function and cultural presence in the region.

At Nowije garden, which is located in the temperate southern climatic zone, a total of 58 planting units were analyzed, including 48 units of Type 1 and 10 units of Type 2. The network was extensive, consisting of 105 nodes and 329 edges. P. densiflora showed the highest connectivity, with a degree of 54 and a degree centrality of 0.52, serving as a central species that interacted actively with a wide range of others. P. thunbergii and D. kaki each had a degree of 43 and a degree centrality of 0.41, followed by T. orientalis with a degree of 28 and a centrality of 0.27. P. densiflora and D. kaki also exhibited high closeness and betweenness centrality, highlighting their role in maintaining network cohesion and acting as bridges connecting diverse species.

4. Discussion

This study analyzed the planting structures and spatial patterns of public-type private gardens registered with the Korea Forest Service to examine their public value and potential uses. The analysis yielded three principal findings.

First, public-type private gardens are officially registered through a government-prescribed procedure, thereby acquiring representative status as gardens. On the basis of this institutional standing, registered private gardens can function not merely as privately owned spaces but also as model gardens embodying public value. To examine their planting structures and patterns, this study introduced the concepts of planting units and analysis units, and systematically analyzed the planting characteristics within these frameworks.

As a result, common planting compositions and structural patterns were identified, suggesting the potential of these gardens to serve as reference models for the design and development of other private gardens. The five public-oriented private gardens analyzed in this study were categorized into one site under mixed artificial and natural management and four sites under artificial management with commercial linkage. Their planting structures were primarily composed of both canopy and shrub layers, or canopy layers alone. In the mixed-managed garden, species such as P. thunbergii, P. densiflora, and P. yedoensis were identified as central species through network analysis. Naturally regenerated vegetation was also observed surrounding the major species, indicating a hybrid planting pattern that combined artificial and natural elements. In the commercially linked gardens, L. indica, P. densiflora, and E. japonicus were dominant in the temperate central region, while P. densiflora and D. kaki were central in the southern region.

Despite differences in climatic zones, several common plant species were identified, and planting pattern analysis suggests that these results can serve as baseline reference data for unregistered private gardens.

These findings signify more than mere planting patterns, insofar as public-type private gardens operate within a legal institutional framework. Grounded in this institutional basis, public-type private gardens can function as core spaces that foster community participation and sustain garden culture [10]. In practice, some gardens have further expanded accessibility and public character by offering free admission. This process can be understood as maximizing the potential of private gardens and laying the groundwork for citizens to directly experience diverse types of private gardens and benefit from them [28].

Nevertheless, despite growing registrations and various institutional initiatives, quantitative data on planting characteristics remains limited. Given that private gardens now carry a public mandate, it is imperative to objectively evaluate their ecological and cultural value and to systematically collect and analyze planting data. As the number of public-type private gardens continues to increase, and as institutional efforts such as nationwide competitions to identify exemplary cases advance in tandem with data accumulation, a more scientific and sustainable garden management system can be established.

Second, public-type private gardens were found to have a higher average tree density (0.176 trees/m²) than neighborhood parks (0.089 trees/m²). These findings are consistent with prior research indicating that private gardens, compared with urban parks, harbor greater plant species richness and exhibit greater structural complexity in stem density patterns [35].

While previous studies [36] emphasized the limited planting area on rooftops and hardened urban surfaces, the private gardens analyzed in this study demonstrated relatively large and dense vegetation coverage despite their urban locations. Notably, the Guam Garden, a rooftop garden, was found to have a high proportion of trees and shrubs, contrary to earlier findings that rooftop gardens are typically dominated by herbaceous species (80–90%). This suggests that even small-scale gardens can provide meaningful ecological functions through dense and diverse plantings [19,37,38,39,40].

Even small urban gardens can support invertebrate assemblages at levels comparable to those in large suburban gardens; in particular, gardens with favorable solar exposure and abundant flowering can sustain diverse pollinator assemblages [41].

With respect to the green area ratio, Cameron [42] proposed that more than 50% of garden space should be vegetated for environmental policy to be effective. Among the five public-type private gardens surveyed in this study, all except Walden Garden (45%) exhibited a green area ratio of at least 50%, reflecting the government’s registration criterion requiring that at least 40% of the total site area be green for such gardens. Taken together, these results indicate that public-type private gardens not only satisfy institutional requirements but also align with prior research underscoring the need to secure sufficient vegetation.

These vegetation structures align with the findings of Humaida et al. [43], who reported that urban gardens can be effective in mitigating the urban heat island (UHI) effect, particularly in areas with elevated surface temperatures.

Furthermore, garden vegetation generally reduces air temperature by approximately 2–3 °C and surface temperature by around 10 °C [44,45]. In this context, the high tree density of public-type private gardens can be interpreted as a potential basis for urban temperature reduction and energy conservation. In addition, as components of the urban green-space network, gardens connect with other green areas to provide habitat continuity and corridors for species movement, underscoring the need for integrated management that links private gardens with public urban green spaces [3]. Moreover, as readily accessible green spaces for urban residents, gardens promote physical activity, mental health, and contact with nature, and therefore hold significant value in urban planning [2,46]. Regardless of whether people prefer private gardens, it is important that private and public green spaces share access and benefits [47].

However, purely private green spaces such as home gardens face constraints on public use because ownership is dispersed among many individuals [48]. Even so, Korea’s private-garden system is significant in that it partially opens private spaces, imbuing them with a public character and establishing a basis for their functioning as urban green infrastructure contributing to climate mitigation, biodiversity enhancement, and improved access to nature. Looking ahead, as cities seek to alleviate green-space fragmentation driven by urbanization and to lower temperatures by increasing planting density within existing green spaces, private gardens are poised to serve as key sites for realizing this potential.

Third, Guam Garden can be regarded as an example of an artificially and naturally integrated urban garden, indicating the potential for developing sustainable ecosystems in densely built environments. This rooftop garden, established 20 years ago on the 7th floor of the building, spans 670 m². Field surveys and owner interviews revealed that 17 planted tree species and 24 spontaneously germinated species coexisted, forming a distinctive vegetation structure.

Network analysis using Gephi confirmed that P. thunbergii and P. densiflora functioned as core nodes, interacted with naturally regenerated species. This hybrid planting suggests possible interactions between artificial design and natural ecological processes. The garden also applies a set of sustainable management practices, including a closed loop water system incorporating rainwater and greywater, a rooftop pond for natural irrigation, food waste recycling through backyard poultry, and solar panels for renewable energy production. These strategies align with the sustainable landscape principles proposed by Alkaisi et al. [25] and reflect approaches similar to those adopted in the Longwood Gardens’ sustainability system as outlined by Turner-Skoff et al. [23]. The Royal Horticultural Society (RHS) has likewise identified escalating water demand as a key challenge for future urban garden management [49]. In this context, Guam Garden may provide useful insights for enhancing the sustainability of urban garden management.

5. Conclusions

This study aimed to analyze the planting structures and patterns of public-oriented private gardens in order to provide foundational data for the systematic management of public gardens. To this end, the concepts of planting units and analysis units were established and applied to systematically examine planting characteristics.

First, the study sites are gardens formally registered in the national registry through a government-prescribed procedure and therefore embody state-recognized public value. On this legal institutional footing, public-type private gardens can function not merely as privately owned spaces but as shared public resources. Moreover, the public mandate conferred by the registration scheme gives these gardens significance as reference planting models that can inform the creation of yet unregistered private gardens.

Second, public-type private gardens exhibited a higher tree density than that of neighborhood parks, confirming that high density planting is feasible within the urban fabric. The rooftop garden case suggests that even small spaces, when planted primarily with trees and shrubs, can perform meaningful ecological functions. Collectively, these results support the role of public-type gardens as urban green infrastructure and indicate that expanding the public use of private spaces through the registration scheme can contribute to enhancing biodiversity.

Third, Guam Garden demonstrates through the coexistence of artificially planted vegetation and spontaneous recruits that sustainable ecosystem formation is attainable even in compact urban contexts. Despite being a rooftop garden, it maintains diverse species and complementary planting patterns, and employs integrated management (rainwater reuse, renewable energy, etc.), indicating potential as a sustainable management model for urban gardens.

This study is limited to five public-type private gardens located in urban areas; accordingly, it cannot on its own yield generalizable planting guidelines. Looking ahead, urban gardens should move beyond simple greening toward effective new forms that can constitute a substantive category of urban green space. In Korea, multiple garden policies for national gardens, local gardens, and public-type private gardens are being advanced; evaluating how public policy integrates and promotes these forms remains an important task. Equally important, the expansion of public-type private gardens and their systematic management and support as a category of public green space should be underpinned by resident collaboration and community engagement. Future research should expand the scope of study sites, accompanied by verification studies on the environmental improvement functions of gardens, including urban heat island mitigation (based on field measurements of soil and temperature) and wildlife monitoring. In addition, conducting high-quality research on public awareness of public-type gardens and their social, cultural, aesthetic, and public health values is essential.