1. Introduction
The area of vacant and abandoned houses is expanding around the original downtown locations in rural and metropolitan areas in South Korea. This is a result of the natural population decline due to the low birth rate, aging population, and movement of the population from city centers to the suburbs [
1,
2,
3]. Neglected and unmanaged vacant houses can cause social problems, such as crime, deterioration of the surrounding landscape, and creating a cycle of producing other vacant houses [
1,
4,
5,
6].
South Korea has been implementing various measures to mitigate these problems. The Fifth Comprehensive Land Development Plan (2020–2040) and the Fifth Comprehensive National Environmental Management Plan (2020–2040), which provide the fundamental direction for national land and environmental management, propose strengthening the small-scale maintenance of old abandoned buildings, monitoring vacant houses, and implementing measures to prevent their transition into further vacancy. These plans are aimed at the management and utilization of old buildings and vacant houses, emphasizing sustainable smart shrinking strategies, such as the removal of vacant houses with low development demand or those that are difficult to reuse, in favor of preserving them as green spaces instead of replacing them with artificial structures.
Strategies to preserve vacant houses and idle spaces in city centers as green spaces can have various effects on declining downtown metropolitan areas. Large cities in South Korea developed rapidly during a period of high economic growth since the 1970s, resulting in overcrowded buildings and a scarcity of green spaces [
7]. Busan Metropolitan City, the second largest city in South Korea, lacks sufficient land for development because slopes with an inclination of 10° or more occupy more than 50% of its total area [
8]. After the Korean War, densely built residential areas for low-income workers were developed on these steep slopes, forming the original downtown residential area with narrow individual building sites and highly inadequate infrastructure, including parks and green areas [
9]. Moreover, Busan has become a highly vulnerable region where the amount of runoff during intensive rainfall increases rapidly due to the steep slopes, causing considerable flood damage. In addition, the concentration of houses hinders air circulation, exacerbating heat during summer. If green spaces are provided in these areas, they can offer not only resting and leisure spaces for citizens but also climate change response effects such as flood risk and CO
2 mitigation, which are currently being emphasized. However, the reality for local governments is that vacant houses are often neglected for a long time due to a lack of financial resources for maintenance and owners’ non-agreement. Although vacant houses may be demolished through vacant-house maintenance, they are typically left as barren sites with concrete pavements [
10].
The effectiveness of the greening strategies emphasized in this policy, particularly concerning the demolition of persistently vacant houses and the management of barren sites left after demolition, requires thorough evaluation prior to their practical application and widespread adoption. Data-driven quantitative assessments can serve as a robust foundation for policy introduction and implementation. The benefits of greening vacant houses and idle spaces have been evaluated from various perspectives in other countries, including increased real estate prices [
11], crime suppression [
12,
13,
14], health improvements [
12,
15], enhanced access to green spaces [
16,
17], ecosystem connectivity [
18], and fire protection [
19]. Only Chae et al. [
20] and Lee [
7] have investigated the effects of greening vacant houses in South Korea. Chae et al. [
20] examined the impact of demolishing vacant houses and creating vegetable gardens within a crime-safe village pilot project in Cheonan, Chungcheongnam-do, through a resident survey. Their findings suggest that demolishing vacant houses and engaging in gardening significantly reduced crime-related fears and fostered a sense of community among residents. Lee [
7] conducted a simulation to assess the thermal environment improvement from creating green spaces in a densely populated residential area with vacant houses as part of an urban regeneration project in Daegu City. The study revealed that the formation of small-scale green spaces, such as gardens, had a minimal effect on temperature reduction, but the creation of a relatively large park led to a temperature decrease of 1.1–3.7 °C. However, little empirical research has been conducted on the effects of greening strategies in South Korea, partly due to the few instances of greening vacant houses and idle spaces, and the small areas these sites typically cover. Therefore, this study employs simulation analyses to investigate the planning process, focusing on the climate change response effects of green spaces, an aspect that has recently been gaining prominence. The response to climate change can be categorized into mitigation and adaptation strategies [
21]. Mitigation strategies involve the reduction of greenhouse gases, such as carbon dioxide, which contribute to global warming. On the other hand, adaptation strategies focus on alleviating the impacts of climate-induced disasters, such as flood risk reduction, addressing damages caused by the changing climate. Given that the greening strategy aims to provide co-benefits in both mitigation and adaptation, this study examines the greening effects of vacant houses and demolished sites in terms of their potential for flood risk and CO
2 mitigation effects.
While no previous studies have directly linked flood risk and CO
2 mitigation to the greening of vacant houses, research on land cover changes related to these mitigation effects has been conducted. Studies on the flood risk mitigation effect of green spaces in urban planning have provided insights into land use, environmental, and disaster prevention planning. A study conducted by Nagano et al. [
22] confirmed that the precipitation interception effect of tree canopies for rainfall exceeding a total of 20 mm ranges from 10% to 26%, with some variation depending on the type of trees. In their study, Choi et al. [
23] analyzed surface runoff using simulations for urban areas on sloped terrain when employing low-impact development (LID) techniques such as planting and permeable pavement. The results revealed that with a reduction in impervious surfaces of approximately 1.5%, an expected annual decrease of about 7.4% in surface runoff could be achieved. In their simulation-based assessment of the impact of urban greening on runoff reduction for a short-term rainfall event exceeding 100 mm, Yang and Lee [
24] revealed that greening with trees is more effective than greening with grass alone.
Studies on the CO
2 mitigation effect of green spaces falls into two categories: those calculating carbon fixation of a unit of green space through direct measurement [
25,
26], and those determining carbon fixation for a target area using the estimated carbon fixation of a unit of green space [
27,
28]. The amount of CO
2 fixation by trees varies significantly based on the tree species, growth stage, and environmental conditions. Many previous studies have focused on forest trees, determining carbon content from the dry weight of tree biomass. These studies also estimate equations for calculating annual carbon dioxide fixation based on tree attributes such as diameter at breast height and height [
25]. In a study conducted by Jo and Cho [
26], which measured the annual CO
2 fixation for trees planted in urban areas in Korea, it was revealed that trees fix between 0.5 and 19.0 kg of CO
2 per tree. Based on the results of these previous studies, research has been conducted to calculate the annual carbon sequestration by green spaces in urban and specific regions. In a study by Jo and Ahn [
27] that calculated the comprehensive carbon reduction effects of trees in South Korean single-family residential areas, the total carbon sequestration per house ranged from 59 to 110 kg annually. The study further concluded that indirect effects, such as the reduction in heating and cooling energy consumption due to shade provided by trees, accounted for more than 70% of the overall carbon reduction effects. Lee [
28] focused on typical multi-family residential buildings in South Korea, where green spaces accounted for 28% of the building volume. In this scenario, it was estimated that plants would take approximately 44 years to absorb all the CO
2 emitted during the construction of the residential buildings.
In summary, this study examines the status of vacant houses and demolished sites in Ami-dong and Chojang-dong, Seo-gu, Busan Metropolitan City; these areas are deteriorated, high-density sloped residential areas in old downtown metropolitan areas with a significant vacancy problem. In addition, this study evaluates the flood risk mitigation and carbon fixation effects of implementing greening policies. To assess the effectiveness of flood risk mitigation, a greening scenario was constructed, utilizing K-LIDM, a runoff analysis model. CO2 fixation effects were evaluated using the annual carbon sequestration formula from previous research. The findings are expected to serve as a reference for utilizing vacant houses as space resources for green spaces in downtown locations, where securing such spaces is challenging. In addition, by quantitatively demonstrating the flood risk mitigation and carbon fixation effects of greening vacant houses, this research aims to support the advancement of greening policies for vacant houses, contributing to the realization of sustainable cities.
3. Results
3.1. Current State of Persistently Vacant Houses and Demolished Building Sites in the Study Area
An analysis of changes in vacant houses from 2017 to 2020 reveals that there were 340 vacant houses out of 3787 detached houses in 2017. The number of persistently vacant houses in 2020 was 246, accounting for approximately 72.35% of all vacant houses in 2017. In 2020, 498 of 3675 detached houses were vacant, with 252 new vacant houses. These results suggest that once houses become vacant, they are unlikely to be reused. In the study area, the average annual increase rate of vacant houses was 13.57%, confirming the persistent existence of vacant houses throughout the area.
Table 3 summarizes the surface states of the demolished building sites as of 2020. In the study area, there were 313 demolished building sites. Although the previous uses or reasons for demolition of the buildings could not be determined, it is presumed that most of these sites were detached houses, given the characteristics of the area. Moreover, many vacant houses were demolished as part of a vacant-house maintenance project carried out by the local government; these must also be considered.
A total of 213 sites were left as barren sites after demolition, accounting for 68.05% of all demolished sites. Among them, 177 sites (83.10%) featured impermeable concrete pavement surfaces (
Table 3). In this region, the prevalence of concrete pavements was due to post-demolition management challenges. For the barren sites situated on permeable surfaces, there were seven cases where gravel was spread to inhibit plant propagation. Most permeable barren sites were neglected and overgrown with weeds.
Among the demolished sites, 100 sites were reused as parking lots. A review of the usage status of these maintenance sites showed that 33 sites were maintained as roads or sidewalks, 29 were used as communal spaces such as parks or sports facilities, 24 served as vegetable gardens, and 5 were used as parking lots.
Figure 4 depicts the locations of persistently vacant houses and barren sites targeted for greening. Additionally, the positions of maintenance sites are included for reference.
3.2. Deriving Greening Scenarios
Table 4 summarizes the area and percentage of land cover for each scenario. Impermeable surfaces were classified into three types: building, concrete, and asphalt. Permeable surfaces were classified into three types: barren, ground cover plants, and trees. The area calculations were based on field survey results, a road name address electronic map, and a land cover map. Initially, ground surfaces identified as roads in the land cover map were classified as asphalt, and other urban/dry areas were classified as concrete. Building areas were calculated based on the road name address electronic map. The ground surfaces of barren and maintenance sites from field survey data were then incorporated. Forest areas in the land cover map were classified as trees, grassland and agricultural areas as ground cover plants, and barren sites as barren. In the current state (S1), impermeable surfaces in the study area accounted for 53.80% of the total area, whereas buildings accounted for 24.42% of the total area. The study area is surrounded by Mount Ami and Mount Cheonma, resulting in trees covering 33.50% of the total area. The total area of persistently vacant houses and barren sites slated for demolition measured 19,764 m
2. The total area being smaller than the administrative district’s area (1.169 km
2) is due to the exclusion of areas where large-scale construction projects were underway during the survey period for vacant houses and vacant lots from the analysis.
Figure 5 illustrates the land cover for each scenario.
3.3. Flood Risk Mitigation Effect
Rainfall runoff analysis was conducted for each scenario by using K-LIDM to assess the flood risk mitigation effect of greening vacant houses and barren sites. Runoff flow was examined with both long- and short-term runoff models. In the case of the short-term runoff model, data for 30-year and 2-year rainfall events were analyzed over a 180 min period.
For the runoff analysis under the long-term runoff model, the annual average flow rate of rainwater from the target site for each scenario was calculated and compared based on scenario (
Table 5). The flow rate increased by 0.16% in S2 compared to that in S1, where both vacant houses and barren sites were paved with concrete. Conversely, the flow rate decreased by 0.91% and 1.71% in S3 and S4, respectively. When compared to S2, which had the highest flow rate, the flow rates in S3 and S4 decreased by 1.07% and 1.86%, respectively.
The results of the short-term rainfall runoff analysis are shown as runoff curves for both 2-year and 30-year rainfall scenarios. In the case of the 2-year rainfall scenario (
Figure 6a), S1 and S2 exhibited similar curves, while S3 and S4 showed overlapping curves. Similarly, the 30-year rainfall curves (
Figure 6b) showed similarity between S1 and S2, but a slight difference was observed between S3 and S4. Notably, S4 had the smallest flow rate, and the runoff curves did not show variations in peak time across all scenarios.
Table 6 summarizes the peak flow rates for each rainfall scenario. In the 2-year rainfall case, the peak flow rates of S3 and S4 were the same, and the peak flow rate decreased by 0.26 CMS (3.06%) compared with that of S1. The peak flow rate of S2, which was the highest flow rate, increased by 0.02 CMS (0.26%) compared to that of S1. By contrast, the flow reduction rate of S3 and S4 was 3.31%. In the 30-year rainfall case, the peak flow rate decreased in S3 and S4. Compared with that of S1, the peak flow rate of S3 and S4 decreased by 0.20 CMS (0.99%) and 0.48 CMS (2.38%), respectively. The peak flow rate of S2, which was the highest flow rate, increased by 0.01 CMS (0.01%) compared to that of S1. By contrast, the peak flow rate of S3 and S4 decreased by 0.21 CMS (1.04%) and 0.49 CMS (2.43%), respectively.
The decrease in the peak flow rate in the greening scenario was larger for the 2-year rainfall event than for the 30-year rainfall event. In addition, the effect of greening was more prominent during lighter rainfall. Moreover, in the case of the 30-year rainfall event, there was a distinction in the peak flow rates between S3 and S4. This suggests that, during heavy rainfall, S4 exhibits a greater flood risk mitigation effect than S3.
3.4. CO2 Mitigation Effect
The selection of trees suitable for achieving the CO2 mitigation effect was the first step. Subsequently, the resulting CO2 mitigation effect achievable by planting these trees was calculated using an equation based on previous studies.
The following four conditions were established for the trees chosen for greening the target sites: (1) trees of a size suitable for planting in vacant houses and barren sites, (2) landscape trees commonly planted in urban areas in South Korea, (3) trees with a relatively slow growth rate that require infrequent pruning, and (4) deciduous broadleaf trees that effectively block sunlight in summer and allow solar radiation in winter. Condition (3) can be considered a criterion for tree selection from a climate change mitigation perspective.
Before considering condition (1), the average area per site was calculated. Based on the road name address electronic map, the average area per persistently vacant house and barren site slated for demolition was determined to be 38.89 m
2. Therefore, it was assumed that the appropriate crown size for trees that could be planted on vacant houses and barren sites after the removal of vacant houses should be within 7.0 m in width. Regarding conditions (2)–(4), it was assumed that deciduous broadleaf trees, such as Chinese maple and red maple, were planted based on research by Lee [
38], which established quality evaluation criteria by standardizing the measurements for 15 tree species, considering the primary planting functions of landscape trees. The average diameter at breast height (DBH), mean crown width (W), and mean crown area of Chinese and red maples, as measured by Lee [
38], were 13.0 cm, 2.8 m, and 6.2 m
2, respectively.
In this study, we utilized equation 1 proposed by Lee [
28] to calculate the annual carbon fixation of broadleaf trees for evaluating the CO
2 mitigation effect of greening. This equation was developed based on research on Korean trees. It is important to note that the carbon fixation rates of trees vary among different species; even within the same species, they can differ based on factors such as forest type, forest age, growth conditions, stand density, and climate [
35]. Consequently, the equation was considered suitable for application in this study.
The annual carbon fixation per deciduous broadleaf tree planted in the study area can be computed as 19.35 CO2 kg/tree·y by substituting the average value of the DBH as shown in Equation (1). The annual carbon fixation resulting from greening the demolished vacant houses and barren sites (S4) can be determined by multiplying the annual carbon fixation per tree by the number of trees to be planted. The number of trees that can be planted was calculated by dividing the total area of vacant houses and barren sites by the canopy area of the trees (6.2 m2). Therefore, 3211 trees could be planted. Therefore, the annual carbon fixation resulting from the greening of vacant houses and barren sites in the study area with trees was calculated to be 62,123.78 CO2 kg/y.
5. Conclusions
This study aimed to evaluate the climate change response effects of greening vacant houses and barren sites in deteriorated high-density residential areas within old downtown metropolitan areas. The current status of demolished building sites was identified in the study area of Ami-dong and Chojang-dong, Seo-gu, Busan Metropolitan City. Furthermore, the effects of greening were quantitatively evaluated in terms of flood risk and CO2 mitigation effects.
The analysis of flood risk mitigation effects involved constructing four greening scenarios for vacant houses and barren sites. A comparison of runoff quantities for each scenario was conducted using the K-LIDM runoff analysis model, encompassing both short-term and long-term runoff models. The CO2 reduction effects were assessed by assuming the greening of all vacant houses and barren sites in the study area with trees and calculating the annual CO2 fixation amount using a previous CO2 fixation estimation formula for trees.
This study’s results can be summarized as follows. Firstly, in the study area, there is a tendency for vacant houses to remain unoccupied for extended periods once they become vacant. Furthermore, when these houses are demolished, they are frequently left neglected with concrete pavement, highlighting the insufficiency of efforts in securing green spaces in vacant-house management. Secondly, the flood risk mitigation effect of greening long-neglected vacant houses and barren sites is most effective when utilizing trees for greening. Compared with the current state in the long-term runoff model, a reduction of 1.71% in annual runoff volume was observed. In the short-term runoff model, a reduction of 3.06% in peak runoff volume for a two-year rainfall event and 2.38% for a 30-year rainfall event was anticipated. Lastly, greening strategies with trees resulted in a CO2 mitigation effect, fixing an annual amount of 62,123.78 kg of CO2. In conclusion, greening long-neglected vacant houses and barren sites within high-density residential areas with trees proves effective in reducing regional flood risk caused by climate change and mitigating greenhouse gas emissions.
This study is significant because it simulated and quantitatively analyzed the effects of greening vacant houses in terms of climate change response, a subject area with limited prior research. The findings of this study are expected to change the perceptions of residents and local governments about the greening of vacant houses. However, this study focused only on a few benefits of green spaces, such as flood risk mitigation and carbon fixation, and was confined to an area-specific analysis, limiting the generalizability of its results. Future research should explore a broader range of greening effects, including its impact on the thermal environment and application across various regions. Furthermore, for the runoff analysis, it is imperative to conduct analyses that reflect the latest rainfall patterns along with updates to precipitation data.