Evaluation of the 20-Year Restoration Process in an Air-Pollution-Damaged Forest near the Ulsan Industrial Complex, Korea

Abstract

A study was conducted to evaluate the effects of restoration practices in a forest ecosystem near the Ulsan Industrial Complex in southeastern Korea. The calcium and magnesium contents in the soil, as well as the soil pH, increased after the application of a soil ameliorator but decreased again after 20 years. Meanwhile, the aluminum content presented the opposite trend. After restoration, the species composition and diversity of vegetation tended to differ from that of the non-restored site over time while continuously becoming more similar to that of the reference site. The ratio of exotic plant species was lower than that at the non-restored site but higher than that at the reference site. The frequency distribution for the diameter class of oaks established through restoration presented a reverse J-shaped pattern, and thus, they can be maintained continuously; similar results were obtained for the reference site. In sum, the forest ecosystem near the industrial park—which had been severely degraded due to air pollution and soil acidification—was restored to a forest close to natural conditions through restorative treatments, including the neutralization of acidic soil and the introduction of tolerant species.

1. Introduction

Significant material and energy resources are used in human activities, especially in industry. As a result, such activities tend to generate a significant amount of pollutants, often exceeding the limit of the buffering capacity of the location where the activity takes place and causing damage to the vegetation and soil of the area [1,2,3]. If the imbalance between the pollution source and the buffering capacity worsens, the vegetation will decline, and if it worsens further, desertification may finally occur. Such a phenomenon has been observed in the forests near the Ulsan Industrial Complex, where this study was conducted [2]. If such conditions are left unaddressed, their impacts may lead to significant damage to other biological communities (e.g., animal and micro-organism communities) and may spread to a wider area [4,5,6]. In addition, the function of carbon-absorption sources (which absorb and remove the carbon dioxide (CO₂) emitted by industrial activities) may deteriorate, resulting in exaggerated differences between the emission and absorption of carbon [7]. Therefore, it is necessary to restore damaged ecosystems in order to prevent the spread of such damages [1,5,8].

Ecological restoration is an ecological technology that aims to restore the disrupted ecological balance by healing the damage caused to nature and restoring its natural buffering function. One process through which a disturbed ecosystem recovers by itself is succession. Ecological restoration is a process that promotes succession through human aid, imitating an intact natural system and thus providing habitats for various organisms. In such a way, ecological restoration seeks to secure the future environment for humankind [9,10,11].

Ecological restoration is carried out through a series of procedures, including diagnostic evaluation, the collection of reference information, the preparation of a restoration plan (based on both types of information), restoration practice, the monitoring of the restoration process, the analysis of monitoring results, adaptive management based on the results, and the evaluation of the restoration effect [10,11,12]. The implementation of restoration is the starting point, while the process of ultimately checking the performance of the restoration execution—including the processes that proceed thereafter—allows for evaluation of the restoration effect(s) [10,11,13]. Restoration is a type of work that seeks to shorten the period of succession—the process of restoring the natural state damaged by disturbances from nature itself—which usually takes a long time. Numerous restoration projects have been promoted; however, the evaluation of the restoration effect is often neglected [14,15,16]. If a project is not evaluated, information on its performance cannot be obtained. Projects cannot evolve without help from past knowledge regarding what has and has not worked and how these achievements differ depending on the project. Small investments into systematic monitoring, information delivery, and evidence-based evaluations will greatly help in the planning of future projects, thereby significantly improving them. For example, policy-making tools have developed with accurate information, cost, and spatial dependencies regarding the likelihood that successful restoration can help to determine which restoration activities to prioritize [17]. Such investments into evaluation also contribute to increasing the relevance of restoration and ecological research, as well as improving the applicability of the results [18,19]. In this respect, evaluating the effects of restoration is an essential task for the development of ecological restoration.

Evaluating the effect of restoration is not simple, and there have been extensive debates on the characteristics of a successful restoration and how to measure them [20]. Ecological restoration is carried out with the aim of restoring the damaged structure and function of nature, thus further securing a system that can be self-maintained. As such, the restoration effect can be evaluated by comparing the structure and function of the restored system with reference information obtained from an ecosystem or landscape with an intact natural system [13].

The Society for Ecological Restoration (SER) Primer is a valuable contribution to the discipline, providing a practical overview of ecological restoration and outlining nine key attributes for judging restoration success [12]. These attributes encompass three general ecological outcomes, vegetation structure, species diversity, and abundance, which are commonly used as indicators of ecosystem conditions [21,22,23]. The “recovery wheel” proposed by McDonald [10] is a system for evaluating the progress of a restored ecosystem along its recovery trajectory. It covers three general ecological outcomes: vegetation structure, species diversity and abundance, and ecological processes [24]. These indicators are frequently utilized in the literature to assess ecosystem conditions [21,22,23].

This paper focuses on evaluating the effects of an ecological restoration practice conducted to enhance the overall system functions of a forest near the Ulsan Industrial Complex in southeastern Korea. The aim was to restore the forest, which was degraded significantly due to severe air pollution and soil acidification, to a self-sustaining ecosystem. The restorative treatment involved soil amelioration and the introduction of tolerant plant species. The implementation of the restoration took place in 1998, and its effects were evaluated in the year of restoration, the 4th and 20th years after the restoration.

2. Materials and Methods

2.1. The Study Area

Ulsan is located on the southeastern coast of the Korean Peninsula and faces the East Sea (126°46′15″ to 127°11′15″ E longitude, 37°25′50″ to 37°41′45″ N latitude; Figure 1). The topography of Ulsan is a concave plain surrounded by high mountains to the west and east. The elevation of the study area, Mt. Dotjil, ranges from 20 to 89 m above sea level. The parent rock of Ulsan is usually composed of sedimentary rock, and the flat land beside rivers and streams consists of alluvium. The soil in these areas was classified into the Eutrudepts great group of the Udepts suborder in the Inceptisols order, which developed on sedimentary bedrock. Based on the soil texture, the soil was silty loam. Due to the poor development of vegetation, the thickness of the litter layer was thin, within 1 cm.

The climate of Ulsan is continental, with warm and moist summers and cold and dry winters. From 1991 to 2020, the mean annual temperature was 13.95 °C, and the mean annual precipitation was 127.0 cm [25].

The Ulsan Industrial Complex serves as a representative example of industrial complexes in Korea. The construction of the complex began in the late 1960s and continues to expand. The industrial activities in this complex primarily consist of heavy and chemical industries, including the petrochemical industry, which emits approximately 60,000 tons of sulfur dioxide (SO₂) per year. These emissions have led to a mean concentration of 0.03 ppm of SO₂ in the local atmosphere, with a maximum daily average of 0.10 ppm [26]. Such severe air pollution and the resulting soil acidification have severely degraded the surrounding vegetation, forming a wide range of grasslands covered with Miscanthus sinensis mats and, in severe cases, bare ground (i.e., not covered with vegetation) (Figure 1). The mean soil pH was 4.1, lower than the pH of 4.8 in non-polluted areas 20 km to the south [27,28]. As a result, the physico-chemical properties of the soil differed greatly from those of the reference area. In particular, the pH and the Ca²⁺ and Mg²⁺ contents in the soil (which are essential ions) were lower than those in the reference area, while the Al³⁺ content was higher [27,28].

Mt. Dotjil, selected as a restoration target site for this study, is the area that has suffered the most ecological damage in the Ulsan Industrial Complex. This mountain is an isolated island surrounded by industrial facilities and urban areas. As a result, it has suffered from severe environmental pollution since the operation of industrial facilities began in the late 1960s. Vegetation patches are interspersed with bare soil, consisting of grassland and several tree plantations (Figure 2). Our restoration focused on improving the acidified soil in this area and introducing plants that are tolerant to polluted environments in order to achieve an integrated forest ecosystem. Land with vegetation coverage below 20% was categorized as bare ground (hereafter BG) and divided into two groups: one dominated by the grass M. sinensis, and the other by the forb Pueraria thunbergiana. Grassland was also divided into two types: one dominated by grass, M. sinensis (hereafter GG), and the other by forb, P. thunbergiana (grassland forb, hereafter GF). M. sinensis maintains dense coverage, inhibiting other plants from entering and establishing in the area. On the other hand, in the plots with a vegetation coverage of less than 20%, the plots with a vegetation coverage increased to more than 20%, which was expressed as a former bare ground (hereafter FB). Natural reference sites were selected in an unpolluted area located more than 20 km south of the Ulsan Industrial Complex (Figure 1). The concentration of SO₂ in the air in this area was less than half of that in the polluted area [2]. The reference area exhibited better environmental conditions compared to the polluted areas, including the physicochemical properties of the soil mentioned earlier. The vegetation of the reference site is dominated by oaks, including Quercus serrata, but a number of Pinus thunbergii have also been established because it is located on the coastal area. The trees that dominate the forest are about 15 m high and more than 20 cm in diameter and thus form a mature forest. In addition, afforestation species such as Alnus firma, Robinia pseudoacia, and Pinus rigida have also invaded under the influence of afforestation created around them.

2.2. Restoration Practice

The planting bed soil was improved by applying a 1:1 v/v mixture of dolomite and sludge at a rate of 1000 kg ha⁻¹, following the method described by Kim et al. [27]. For experimental restoration, we installed five 5 m × 20 m plots. The restoration was carried out by planting three-year-old seedlings (nine each for Quercus serrata, Alnus firma, and Ligustrum japonicum for a subplot with a size of 5 m × 5 m) on both grassland and bare ground in 1998. Seedlings were planted at 0.5–1.0 m intervals. To restore this area, which was close to bare ground, plant species comprising tree, understory tree, and shrub layers were introduced as plants with high tolerance levels [28]. Q. serrata, A. firma, and L. japonicum were introduced as the plant species that would form the tree, sub-tree, and shrub layers, respectively, after restoration. Furthermore, the inclusion of A. firma was based on its ability to fix atmospheric nitrogen through a symbiotic relationship with actinomycetic fungi [2].

2.3. Soil Analysis

Soil samples were collected from each plot, with five points sampled per plot. These individual samples were then combined to create a composite sample. The soil pH, as well as the contents of Ca²⁺, Mg²⁺, and Al³⁺, were measured since they serve as sensitive indicators of soil acidification. The pH measurement was conducted using a pH meter. To do this, a soil sample was mixed with distilled water at a ratio of 1:5 (weight per volume). The mixture was then filtered through filter paper. The levels of exchangeable Ca²⁺, Mg²⁺, and Al³⁺ were measured using an inductively coupled plasma (ICP) atomic absorption spectrophotometer (Shimadzu ICPQ-1000, Tokyo, Japan). Extraction of the ions was performed using 1N ammonium acetate solution, with a pH of 7.0 for Ca²⁺ and Mg²⁺ and a pH of 4.0 for Al³⁺. The extraction method used was described by Allen et al. [29].

2.4. Vegetation Analysis

To compare the restored plots at Mt. Dotjil with relatively undisturbed areas, five reference plots measuring 20 m × 20 m were established in a forest located approximately 20 km south of Mt. Dotjil (Figure 1). No soil amelioration or planting activities were conducted in the reference plots.

Vegetation data were collected three times at the restored stands in Mt. Dotjil, in 1998, 2002, and 2018. For the reference plots, data collection occurred only in 2002. Sixteen plots for bare ground; ten plots for GF, GG, and FB; and five plots for the restored and the reference plots were surveyed.

All plant species were identified following Lee [30] and the Korea Plant Name Index [31]. The cover of each species was assessed using the cover class system of Braun-Blanquet [32]. The ordinal scales were then converted to the median value within each cover class to quantify the percent coverage range. The importance value of each species was determined based on its relative coverage. The relative coverage was calculated by dividing the cover fraction of each species by the summed cover of all species in each plot and multiplying the result by 100. An importance value matrix for all species in all plots was constructed and used as data for detrended correspondence analysis (DCA) ordination [31]. To compare species diversity and dominance, a rank abundance curve [32,33,34] was created, and the Shannon–Wiener Diversity Index (H′) was calculated for each stand type.

For woody species in each quadrant, the diameter at breast height (DBH) was measured. Woody plants with a height of less than 2 m were measured for their diameter at ground surface (D₁₀). A diameter class distribution diagram was created based on the frequency distribution within each class, which was divided at 5 cm intervals.

2.5. Evaluation of Restoration Effect Based on Chemical Properties of Soil

The effects of restoration on the chemical properties of the soil were evaluated in terms of how effectively the soil ameliorator treatment helped to recover healthier soil properties. Soil characteristics measured in the year of restoration and the 20th year after restoration were compared with the soil characteristics at the reference site.

A one-way analysis of variance (ANOVA) was performed to compare the differences in soil characteristics (i.e., pH, Ca²⁺, Mg²⁺, and Al³⁺) among the sites. A post hoc analysis was conducted using Scheffe’s test. The data analysis was carried out using SPSS (version 24, SPSS Inc., Chicago, IL, USA) and R software version 4.2.3 (R Project for Statistical Computing).

2.6. Evaluation of Restoration Effect Based on Vegetation Structure

The effects of restoration on vegetation were evaluated based on the changes in species composition and diversity, the ratio of the number of exotic species to total species, and vegetation dynamics. Changes in species composition were analyzed by applying DCA ordination. The species rank—dominance curve and the Shannon–Wiener Diversity Index (H′), were employed to assess changes in species diversity and species dominance [34,35,36]. The effects of restoration on vegetation dynamics were evaluated using the diameter class distribution diagram.

3. Results

3.1. Soil Properties

After treatment with the soil ameliorator, the soil pH, Ca²⁺, and Mg²⁺ contents increased but then decreased 20 years after the treatment. The Al³⁺ content showed the opposite trend—decreasing after treatment but increasing again 20 years after the treatment (Figure 2).

3.2. Species Composition

The DCA ordination analysis of vegetation data surveyed in the year of restoration revealed that the restored sites were positioned between the non-restored sites and the reference sites, indicating a positive restoration effect in terms of species composition. However, the restored sites exhibited distinct species compositions compared to the reference sites (Figure 3).

Similarly, the ordination results based on data collected in the fourth year after restoration displayed a similar trend to the year of restoration. However, it was observed that the restored sites approached the location of the reference sites more closely, and the species composition of the restored sites became more similar to that of the reference sites (Figure 3).

In the DCA ordination analysis using vegetation data surveyed in the 20th year after restoration, the restored sites tended to cluster in nearly the same location as the reference sites (Figure 3).

3.3. Coverage of Miscanthus sinensis

The coverage of M. sinensis—which flourishes in destroyed ecosystems (such as the one at the study site) and inhibits the establishment of the other plants—continuously decreased over time after restoration, matching the level at the reference site 20 years after restoration (Figure 4).

3.4. Species Diversity

Regarding the change in species diversity, analyzed according to the species rank–dominance curves, the restored sites showed no difference from the non-restored sites in the year of restoration, but the species richness doubled in the 4th year after restoration and was nearly four times higher in the 20th year after restoration (Figure 5).

3.5. Exotic Species

The exotic plant species ratio in the restored stand was lower than that at the non-restored sites but higher than that at the reference site (Figure 6).

3.6. Vegetation Dynamics

The frequency distribution of the diameter class of oaks—the trees introduced for restoration—presented a reverse J-shaped pattern, while that for the reference site was observed as a pattern in which a reverse J-shaped pattern and normal distribution overlapped (Figure 7).

4. Discussion

4.1. Restoration Effects Confirmed from Soil Physico-Chemical Properties

The application of a soil ameliorator resulted in the neutralization of the acidified soil. Furthermore, it increased the Ca²⁺ and Mg²⁺ contents of the soil while reducing the Al³⁺ content. The neutralization of the acidified soil and increases in Ca²⁺ and Mg²⁺ contents can be attributed to the effects of dolomite [2,37,38]; on the other hand, the reduction in Al³⁺ in the ameliorated plots suggests that the sludge applied as a soil ameliorator functioned as a chelating agent for Al³⁺ [2,39,40].

Many tree species are sensitive to the Al³⁺ content in the soil solution [41,42,43]. Aluminum toxicity inhibits root growth by impeding both cell division and elongation [42,44,45]. Consequently, it reduces the volume of the root system, damages the roots, and disrupts the uptake of ions such as calcium and phosphate [46,47,48]. Meanwhile, a lack of soil nutrients such as Ca²⁺ and Mg²⁺ restricts root growth and aggravates the issue of inefficient nutrient uptake caused by root damage [45,49,50].

The supply of plant residue compost, urban waste compost, animal manure, and coal-derived organic products to acidified soil often increases soil pH, reduces Al³⁺ saturation, and improves the soil to conditions suitable for plant growth [51,52,53]. Recycling these wastes for soil amelioration offers environmental and economic benefits as long as they do not contain environmentally harmful substances. These organic materials provide metal-binding and pH-buffering capacities, which play crucial roles in determining the pH of the treated soil [53,54,55]. Although the use of dolomite and sludge mixture contributed to improving the acidified soil, it is important to note that its application raises concerns due to potential groundwater contamination and eutrophication [4,56,57]. By facilitating the mineralization of soil organic matter, dolomitic liming increases nitrate release from the soil, leading to groundwater pollution [40,56,58]. Therefore, we recommend the restricted use of these soil ameliorators.

4.2. Vegetation Development

The results of stand ordination indicated that the species composition of the restored stands differed from that of the non-restored stands and was similar to that of the reference stand. This trend became more pronounced over time, and the species composition of the restored stands 20 years after restoration showed little difference from that of the reference stands (Figure 3). The establishment of Q. serrata, L. japonicum, and A. firma—which were introduced for restoration—reduced the coverage of M. sinensis (Figure 4). The reduction in M. sinensis coverage can create a safe site for other species, encouraging their recruitment and establishment [59].

In the forest of the polluted area, the trees that make up the canopy layer are usually affected first and removed. As the trees that make up the canopy layer decline, plants of the shrub layer are impacted next, followed by the plants of the herb layer. The result of these sequential effects in the vertical layers of vegetation led to a homogeneous M. sinensis mat. This sequential decline in the vertical layer of vegetation is known as the “peeling off” or “layered vegetation” effect [60]. This grass mat also appears during the recovery process of ground made bare due to severe pollution damage [2]. These dense mats are usually composed of pollution-tolerant plants [5], including M. sinensis itself [3]. Such vegetation mats monopolize resources such as light, nutrients, and water, hindering the progression of succession and the establishment of other species [61], as has been demonstrated in the inhibition model of Connell and Slatyer [62]. The decline of such vegetation mats indicates the resumption of succession [63,64,65] and can be considered a positive indicator of successful restoration [1,66,67].

As shown in Figure 3, vegetation development through passive restoration did not progress at all or rarely progressed in the non-restored sites. On the other hand, the restored sites presented a noticeable change toward the reference vegetation, along with increased species diversity (Figure 5).

The species rank–dominance curve (Figure 5) provides insights into species diversity, including species richness and species evenness. In response to the restorative treatment, species richness increased, indicating the addition of more species to the restored sites. The slope of the curve, which represents species evenness, also increased, indicates a more balanced distribution of dominant species across the community. Consequently, both species richness and evenness showed an increasing trend with time after restoration. The Shannon–Wiener index, which reflects species diversity, exhibited a similar pattern of increase over time, supporting the observations from the species rank–dominance relationship (Figure 5).

The size distribution of trees is a valuable indicator of changes in the tree population structure [68,69,70]. Ecologists have long focused on these measurements to understand species coexistence, competition, and forest management [71]. In a plant population’s diameter-class distribution histogram, the pattern of frequency distribution across diameter classes provides insights into the potential changes within the plant community. A reverse J-shaped distribution pattern, with more young individuals and fewer mature individuals, suggests a population that can be maintained continuously [71,72]. On the other hand, a normal distribution pattern, with fewer young individuals than mature individuals, often indicates a population that may be replaced in the future [66]. However, bimodal patterns may occur in populations regenerated by periodic disturbances, deviating from the typical patterns [72,73].

The frequency distribution of the diameter class for the introduced oak population displayed a reverse J-shaped distribution pattern, suggesting the potential for the continuous maintenance of the oak community established through the restoration project [74,75,76]. The diameter class distribution of the oak population from the reference site exhibited a mixture of a normal distribution type and a reverse J-shaped distribution. This indicates ongoing recruitment of young individuals, such as seedlings and saplings, and suggests the possibility of continuously maintaining the oak population [74,75,76].

4.3. Percentage of Exotic Species

The proportion of exotic species at the restored site was lower than that in the non-restored site but was higher than that at the reference site (Figure 6). The introduction and spread of exotic species is a global phenomenon, occurring both intentionally and unintentionally [77,78,79]. These exotic species often expand their habitat range beyond early settlements through favorable life history strategies [78,79]. It is well-known that disturbed sites provide microsites that favor exotic species with opportunistic or ruderal life history strategies [3,78,79]. Although all ecosystems have the potential to be invaded by exotic species, certain ecosystems, such as periodically disturbed areas such as industrial sites, are particularly vulnerable to invasion [80,81]. Additionally, evolutionarily and geographically isolated ecosystems, such as oceanic islands, are also at higher risk. Human disturbances further enhance the vulnerability of ecosystems to invasion [82,83,84]. The invasion of exotic plants begins in places disturbed by human influence, which also leads to their spread. The considered study area is surrounded by an industrial complex and urban areas and, thus, is similar to a real island. Moreover, the area where this study was conducted is in a state where the vegetation has been damaged to the level of grasslands or lowlands due to extreme air pollution and soil pollution, thus providing favorable conditions for the invasion of exotic species. In this respect, it can be seen that it is exposed to the invasion of exotic species in a defenseless state. Under these conditions, ecological restoration serves as a means of lowering the ratio of exotic species and exerting the restoration effect (Figure 6).

4.4. Evaluation of Restoration Effect

Continued population growth and the spread of industrial land uses lead to declines in environmental quality. Furthermore, the natural landscape, which provides various ecosystem services, is rapidly transforming into agricultural, industrial, and urban areas, even degenerating into wasteland. Consequently, the balance between the emission source of environmental stress and its absorption source is being tipped [81,83]. As shown in our results (Figure 3), natural succession was delayed in the study area. Thus, restoration activities to promote natural recovery and passive development along desirable trajectories were required.

Soil ameliorators mixed with dolomite and sewage sludge can be applied to restore degraded ecosystems, particularly those degraded by air and soil pollution [59,85,86,87]. These soil ameliorators were effective in improving the polluted soil, thus achieving successful restoration [59]. However, they may also introduce other issues, including groundwater contamination and eutrophication [4,56,58]. In light of these concerns, we also recommend planting tolerant plants, rather than applying soil ameliorators as the sole restorative treatment, in all cases. The trajectory of a restoration project can be observed by assessing both the structure and function of the ecosystem [88,89]. Similarly, the degradation of an ecosystem involves changes in both dimensions. The ultimate goal of restoration is to bring a habitat or ecosystem back to a state similar to its pre-degraded condition. While complete restoration implies a return to that specific state, partial restoration or alternative trajectories may also occur through the establishment of alternative systems [9,90,91].

To effectively restore degraded areas or protect existing high-quality areas, it is important to define attributes based on the characteristics of “normal” and “healthy” habitats [40,92,93]. One approach to setting such criteria and assessing the restoration effect is to define the biological integrity of the system and then measure the deviation from it. Integrity refers to the state of being non-damaged, complete, or non-fragmented. Biological integrity has been defined as “the ability to support and maintain a balanced, integrated, adaptive biological system with the full range of elements and processes expected in the natural habitat of a region” [94,95]. To assess the restored site in this study, the ecological characteristics of the site were compared with those of an “undisturbed” natural reference site. In our study, we decided to compare the species composition, biodiversity, the control of a plant-inhibiting vegetation development, and the ratio of exotic plant species in a restored sites with those at the reference (natural oak forest) and non-restored sites. In addition, we evaluated the sustainability of the restored vegetation by analyzing the frequency distribution of diameter classes in major plant populations.

The species composition of the restored site resembled that of a natural oak forest (Figure 3), and the diversity increased to the extent that it was close to that at the reference sites (Figure 4). In contrast, the sites without any restorative treatment not only showed different species composition from that at the reference sites (Figure 3) but also had a lower biodiversity (Figure 4). Therefore, the restorative treatment approached the restoration goal by increasing both biological integrity and ecological stability.

Biodiversity was significantly increased at the restored site after restoration (Figure 4). The importance of biodiversity is based on a variety of values involving various ecological functions leading to environmental stability [96,97]. Biodiversity is based on the heterogeneity of habitats (or eco-diversity). High biodiversity also derives from the integrity of the environment, that is, a healthy environment with all its components [98,99,100]. High biodiversity in the restored site resulted from suppressing the M. sinensis by restoring taller woody plants. In addition, this result was also promoted through soil amelioration [101].

On the other hand, the restorative treatment contributed to lowering the dominance of succession-inhibiting plants (Figure 5) and the ratio of exotic species (Figure 6), thus securing the sustainability of the introduced vegetation (Figure 7). Consequently, the restorative treatment increased both biological integrity and ecological stability and thereby met the restoration goal [9,94,95].

5. Conclusions

The forest ecosystem surrounding the Ulsan Industrial Complex (the largest industrial complex in Korea) has been severely damaged by extreme air pollution and subsequent soil acidification. The forest vegetation was gradually destroyed through the “peeling off” effect and was degraded to grasslands; many places were degraded further to bare ground.

This study was conducted in Mt. Dotjil, where the vegetation had been severely damaged and soil acidification had severely progressed. This study was conducted with the aim of restoring a damaged forest ecosystem by ameliorating the acidified soil and introducing plant species that are tolerant to air and soil pollution. The restoration practice was performed in 1998, and its effect was evaluated in the year of restoration, 4 years later, and 20 years later. The restoration effect was evaluated based on the physico-chemical characteristics of the soil and the species composition, species diversity, dynamics, and exotic species ratio of vegetation.

As a result of the evaluation of the restoration effect, the physico-chemical properties of the soil improved significantly after treatment with the soil ameliorator (a mixture of dolomite and sludge), but indicated that they acidified again after 20 years. These results were interpreted to be due to the continued emission of pollutants from industrial facilities, although the environmental condition had improved significantly since the early days of industrial activity in the area. However, the vegetation achieved by introducing tolerant species to pollutants was well-established, and 20 years later, a species composition similar to that at the reference site was achieved and species diversity was also restored to a level similar to that at the reference site. It is believed that the plants introduced for restoration played an important role in pioneering an open habitat by suppressing M. sinensis, which served as an obstacle to the invasion of new vegetation.

Based on the results of this study, it was concluded that the selection and introduction of tolerant species, along with soil amelioration, are important steps toward restoring an ecosystem destroyed by pollutants emitted from industrial facilities. In addition, in order to maintain the establishment and the development of vegetation introduced for restoration, it is important to create a safe site by suppressing aggressive plant species that thrive in the destroyed site. On the other hand, it was judged that soil improvement was important for the establishment of introduced plants in the early stages of restoration but did not have a significant impact after their settlement.