Cumulative Environmental Impacts of Wind Power Complex Construction in Mountain Forests: An Ecological Restoration Perspective Through Avian Diversity

Abstract

Countless studies have been conducted on cumulative environmental impacts, but not many have been performed on specific cases of the presence or absence of actual cumulative environmental impacts and restoration, and there is no standardized method for judging cumulative environmental impacts. With our study, we aimed to fill this research gap. We methodologically propose that environmental impacts and restoration can be intuitively expressed by measuring alpha diversity, beta diversity, and gamma diversity in avian fauna. We hope that our proposal will enable many researchers to apply these measurements not only to wind power projects but also various others. This will let them move away from data-less claims and rather track changes in environmental impacts with objective data, considering cumulative environmental impacts and recovery of impacts over time. The results showed that there were changes in the three types of diversity due to the development of wind power complexes on mountain ridges during construction, and the environment was restored over several years. In conclusion, cumulative environmental impacts due to the development of wind power complexes on mountain ridges were not found when limited to avifauna, and impacts due to construction showed a pattern of restoration after construction. There were also cases where restoration occurred even during construction.

1. Introduction

Since the 1990s, concepts such as sustainable development, environmental impact assessment, and cumulative environmental impact have been actively studied [1,2,3]. The use of all of these concepts intends to prevent unnecessary development, minimize environmental destruction or pollution, and record environmental changes in three stages before, during, and after construction from various development projects to provide clearer standards for similar projects in the future. In Korea, after reviewing the results of the environmental impact assessment, government agencies instructed businesses to take mitigation measures to minimize environmental destruction. Records of various projects have been accumulated as public data, and there are cases where various projects have been studied to determine the effects of implementing mitigation measures based on this [4].

Environmental impact assessments have been conducted for decades. Institutionally, environmental impact assessments are reviewed by government agencies, meaning that they are mainly conducted by civil servants. In Korea, civil servants generally change jobs once every two years, so the review opinions presented for similar or identical environmental impact assessment results also change every few years.

For this reason, studying the methodologies that are used to assess environmental impacts is extremely important, and many researchers are highly focused on generating appropriate methodologies [5,6,7,8]. The reason why practitioners who review the results of environmental impact assessments review the ecological part differently every time is because the results of ecology monitoring are merely listed when the process is carried out by evaluators. These differences in evaluation arise because the changes across the three stages—before, during, and after construction—are not intuitively presented. A report might state that species a, b, c, and d appeared before construction, while species b, c, d, and e appeared during construction, and species c, d, e, and f appeared after construction. In such cases, practitioners may focus their review on the species they consider more important or evaluate only the species c and d that appeared together. Most reviewers are not proficient in statistics or are too busy to perform statistical processing, making it impossible to evaluate the data presented in the report. Consequently, practitioners must make inferences based on unprocessed data, and this data is reviewed differently depending on individual abilities.

The concept of cumulative environmental impact begins with the assumption that construction projects can have cumulative environmental impacts. For about 20 years, there has been no clear international agreement on the definition of cumulative environmental impact, and there have been no established standards for how to measure cumulative environmental impact [9]. A cumulative impact can occur from a single construction project carried out over a very long period of time, or when multiple construction projects are being carried out in one area. When multiple construction projects are underway over a very large area, monitoring of all construction projects is necessary to analyze their cumulative environmental impacts. The project owner who initiates the first construction project only needs to monitor the scope of their own project. However, the project owner of the second construction project must track the impacts of the first project. As the project progresses to the third and fourth phases, monitoring must also include the scope of the earlier construction projects. Ultimately, assessing cumulative environmental impacts incurs greater monitoring costs for the project owner who initiates later projects. In some cases, this can place an unrealistic burden on project operators who are late bidders, so it is very important to accumulate data on cumulative environmental impact and gain experience. First of all, it is essential to continuously accumulate data obtained from the same method over three stages: before, during, and after construction. Second, we need to provide easy-to-read analysis results so that practitioners can intuitively understand the environmental impact caused by construction and the time it takes for the environment to restore. At the very least, each construction project should assess its cumulative environmental impact. The time it takes for the environment to recover should be clearly documented. Furthermore, everyone should be able to draw the same conclusions without ambiguity. This will help reduce unnecessary costs.

In the assessment of environmental impacts and cumulative environmental impacts, the intuitive presentation of monitoring results for highly mobile animal species presents a significant challenge. This difficulty is particularly pronounced when dealing with species that exhibit extensive home ranges, appear only during specific periods, are observed intermittently, or occupy unique ecological niches. In such cases, the interpretation of results can vary considerably among practitioners, depending on their specific interests and perspectives. To mitigate these discrepancies in interpretation, it is crucial to acquire a substantial volume of monitoring data collected across all seasons. Such abundant data facilitate the tracking of seasonal variations occurring across the three distinct phases—pre-construction, during construction, and post-construction—or enable the analysis of annual changes by integrating comprehensive yearly datasets. Given that the population size of animals within a monitoring area is subject to temporal fluctuations, it is imperative to process the data according to consistent criteria to ensure the generation of intuitive and comparable results.

The three indices that can intuitively track ecological changes—alpha diversity, beta diversity, and gamma diversity—were proposed a long time ago by the great ecologist Robert H. Whittaker. Since then, many ecologists have applied these concepts [10,11,12,13]. If environmental impact assessors make good use of these three indicators to intuitively represent environmental impacts and clearly present their interpretations, they can prevent inferences that vary depending on the individual capabilities of practitioners reviewing environmental impact assessments. These diversity indicators are based on quantitative data and can therefore be used to analyze fauna where quantitative data are recorded.

Although active research on environmental impact and cumulative environmental impact began about 30 years ago, the concept of cumulative environmental impact has not yet widely spread in Korea. In the mid-2000s, most reports were published without peer review by selecting several overseas cases, and very few were published in academic journals [14]. Research that is actually applicable to the Korean situation has not been conducted for a long time. Although papers related to cumulative impact have been published since 2019 [15,16], they are focused on spatial research; research on cumulative environmental impact focused on animal ecology is still not being conducted. In the neighboring countries of China and Japan, there has been some mention of cumulative environmental impacts, but there have been no in-depth studies on the impacts on animals and living things and the recovery process [17,18,19].

We deem Yeongyang-gun County in Korea an appropriate area to study the cumulative environmental impact of constructing wind farms. Currently, five wind farms are located within a radius of about 5 km, and each wind farm was constructed at different times. Each wind farm was developed linearly along the ridge of the mountain, and roads were developed to access the wind farm. The mountain forest was removed to create the power farm site and to allow an access road to be constructed, creating flat areas without trees. Since this construction did not completely remove the mountain but rather developed a part of the ridge, there are many spaces left for animals to use. Due to the construction in this area, it can be very advantageous to track the environmental impact on animals here and the degree of their ecological restoration. Another advantage is that the results of monitoring three out of five wind farms before, during, and after construction in terms of the environmental changes caused by such developments have been accumulated, allowing us to compare the ecological changes in animals in each development complex.

The purpose of this study is to analyze the cumulative environmental impacts of construction on animals and to suggest standards. Frequently, many studies focus on something that is accumulated too closely, so they present limited results on whether or not there is environmental restoration and its degree. Therefore, to fill this research gap, we aim to present both aspects of environmental impact and restoration objectively. Since quantitative data are not recorded for all animals, they cannot represent the ecological restoration of all animals. However, our study methodologically provides an intuitive way to present changes in the ecology of animals and their degree of restoration.

2. Materials and Methods

2.1. Study Sites of Cumulative Environmental Impact

Our study targets wind power plant projects implemented in Samui-ri and Changsu-ri (village), Yeongyang-gun (county), and Gyeongsangbuk-do (province) in South Korea. We collected the records of three wind farms in Yeongyang-gun County after obtaining permission from each operator. The details of the wind farms in the study area are shown in

Table 1. List of sites with cumulative environmental impact study.

Sites	Monitoring Year	Construction Status
Yeongyang Wind Power Complex (YY)	2009	Before construction
	2010–2013F a	Under construction
	2013S b–2016	After construction
Muchang Wind Power Complex (MC)	2015	Before construction
	2017–2018F a	Under construction
	2018S b–2019	After construction
Yeongyang 2nd Wind Power Complex (GS-2nd)	2017, 2019, 2020	Before construction
	2022–2023F a	Under construction
	2023S b–2024	After construction

^a F, first half; ^b S, second half.

and Figure 1 and Figure 2.

2.2. Avifauna and Diversity

Quantitative records pertaining to avian populations were manually extracted from a diverse collection of existing reports and subsequently compiled into a single, unified text file; this then served as the primary dataset for the subsequent analysis. The data organized into a single file were preprocessed using various methods and analyzed using R version 4.3.3 and R Studio version 2025.03 [20,21,22,23,24,25,26,27,28]. The tidyverse (version 2.0.0) package was used for data transformation, and the tibble (version 3.2.1) package was used for row matrix transformation [29,30,31,32]. The DPLYR (version 1.1.4) package was used to filter and reformat the data frame [33,34]; the traditional method was used for each diversity measure (Shannon and Chao1 index for alpha diversity; Bray–Curtis distance for beta diversity; and gamma diversity) [10,11,12,13]; and the ggplot2 (version 3.5.1) package was used for the graphical representation [35,36,37,38].

3. Results

3.1. Alpha Diversity

To measure the alpha diversity of the three wind power complexes, YY, MC, and GS, we used two indices: the Shannon index and Chao1 index. Though not found with the Shannon index for GS, alpha diversity showed a pattern of increasing with construction at all sites. In addition, alpha diversity decreased after construction was completed at two of the sites, YY and GS, and increased over time (Figure 3).

3.2. Beta Diversity

The dissimilarity plots, made using Bray–Curtis distance, are shown in Figure 4. YY11 and MC18S, which are located the furthest from each other in MDS1, showed a difference of 49.2%, and GS23S and GS22, which are located the furthest from each other in MDS2, showed a difference of 14.9%.

In terms of restoration, YY10, when construction began, showed the effect of the construction by changing close to -1.0 on the MDS2 axis; this restored slightly in YY11, one year later, and then it showed a relatively large restoration trend in YY12, the second year of construction, and in YY13F and YY13S, the third year of construction. Due to the start of the construction of the GS–wind power complex (GS-1st) in the adjacent northern area, YY14 changed to -0.5 on the MDS2 axis, showing a pattern of being affected by about half compared to direct construction; this recovered slightly in YY15, one year later, and was restored relatively significantly in the second year of construction. The MC site was greatly affected in MC17, at the start of construction, though it was restored significantly in MC18F and MC18S, one year later, and showed restoration after only two years. The GS site was only greatly affected in the first year of construction, after which it was restored.

3.3. Gamma Diversity

In the case of YY, gamma diversity increased during construction compared to before construction, and as construction completed, from 2010 to the first half of 2013, the gamma diversity here immediately restored to a level similar to that found before construction in the second half of 2013. However, gamma diversity decreased in 2014 and 2015 due to the construction of another wind power complex to the north. After this, it was restored to a level similar to that found before construction in 2016. In the case of MC, the level of gamma diversity decreased during construction compared to that found before construction, and the largest decrease in gamma diversity was observed in the first half of 2018 when construction was being finalized. After construction was completed, it began to restore immediately in the second half of 2018, and gamma diversity increased, showing a higher level of gamma diversity than that found before construction commenced in the second year after completion. In the case of GS, gamma diversity decreased during construction. However, in the first half of 2023, when the construction work was complete, gamma diversity increased; it showed similar values to those found during construction in the second half of 2023; and in 2024, a slight increase in gamma diversity was observed.

4. Discussion

Usually, environmental impact assessment reports only record the number and population of bird species, making it difficult for viewers to compare different species. However, our research results demonstrate comparisons across three different species diversity (alpha diversity, beta diversity and gamma diversity).

In order for data to be shared and applied by everyone, data preprocessing or standardization through statistical methods is required [39,40,41,42]. Use of the Shannon index, a type of standardization index and part of alpha diversity, generates slightly better reports. However, only the numerical results from simple measurements are presented, making it difficult to interpret what the numbers actually mean. In order to make a correct comparison and interpretation, several concepts must be clearly understood before analysis commences. Firstly, in environmental impact assessments, all developments are assigned animal monitoring ranges that are proportional to the type or scope of the construction. This means that the probability of animals appearing in different monitoring ranges may be different. In fact, it should be assumed that the data collected at all monitoring stations have different standards. Even when referring to previous studies, it should be recognized that it is difficult to compare simply by listing the number and population of species. Accordingly, when predicting environmental impact on species due to construction before the construction has begun, a comparison should be made using standardized data, not by merely comparing the number of species and their population. Secondly, we need to analyze standardized data from various angles. Since there are differences in perspectives depending on each standardization method, multiple complementary indicators need to be used in order for comparisons to be made. Thirdly, we should not assume that cumulative environmental impacts always exist; they may or may not exist depending on the situation. The introduction of a new concept to the academic world does not necessarily result in that concept being applicable to all cases. Therefore, we need to think more flexibly, understand the reasons why the concept arose, and apply it to the cases we encounter directly. The concept of cumulative environmental impact means that damage to the environment caused by construction does not result in the environment being restored to its original state; rather, the damage accumulates.

Our study analyzed whether cumulative environmental impact really occurs during the construction process of wind farms, which develop a linearly limited range along the ridge of the mountain forest, by narrowing down the target to the most mobile animals: bird communities. We analyzed the changes before, during, and after the construction of three wind power complexes with different development ranges and avifauna monitoring ranges using the Shannon and Chao1 indices, that belong to alpha diversity, and presented intuitive changes using the Bray–Curtis dissimilarity method, which is a detailed methods of beta diversity that allows for the most intuitive expression. We also measured gamma diversity, which reflects both alpha and beta diversity, to show the overall changes and restoration process of avifauna. The Shannon index is the most widely used and preferred index; this is the best index for moderately analyzing community diversity [43]. In the case of the Chao1 index, it is beneficial to use it together with the Shannon index because it undergoes a correction process for rare species that can sometimes be ignored in the Shannon index [44]. The Bray–Curtis dissimilarity method is the most widely used method to aid understanding because it intuitively reflects actual ecology very well [45]; therefore, if used correctly, it helps researchers to understand ecology well. Since there may be cases where the results of alpha diversity and beta diversity are somewhat different, measuring gamma diversity, which reflects both, is very helpful for understanding the overall flow; the accuracy and advantages of gamma positivity are widely reported [46].

When understanding alpha diversity, beta diversity, and gamma diversity, the original state before construction began must be taken as the standard. Increased diversity does not necessarily mean ecological health. However, many people mistakenly believe that increased diversity means improved ecological health. This is especially true for birds with large mobility and a very wide range of activities. A stable diversity value is measured in an ecologically stable environment, i.e., if there is an environment where food is readily available, where birds can rest and sleep well, the likelihood that bird species have adapted to this environment will increase. On the other hand, if the environment is poor, species that are observed in passing will be observed probabilistically. In our research results, the Chao1 value of all three wind power complexes increased as soon as construction began, meaning that diversity increased. This is because the number of species that live in stable groups decreased and the number of species that are intermittently observed increased. When the dominance of a specific species decreases, the diversity value increases. Ideally, the highest diversity value is produced when each species is represented by only one individual. With this in mind, when the diversity value increases significantly, we should first think that species that were stably dominant have decreased. From this, the Shannon index, which reflects less rare species and more dominant species, did not show much change in terms of diversity of the three wind power complexes, and it was very difficult to judge whether there was any cumulative damage. In the Chao1 index, which reflects many rare species, diversity generally increased at the start of construction, decreased the following year, and then increased again after completion of the construction, meaning that it was also difficult to judge cumulative damage using this index. In the case of YY, which has the widest construction scope, construction was directly carried out within the scope in 2010, and the construction of another wind power complex was carried out outside the scope in 2014, but the Shannon index repeatedly increased and decreased and did not show any cumulative changes in value. This suggests that the construction did not remove the entire mountain; rather, species that appear stably were maintained due to the characteristics of the construction, meaning that a considerable amount of forest remains. Even in the case of GS, where the straight-line length of the development area is about half that of YY, development was carried out on the mountain directly west of YY, but the Shannon index showed repeated slight increases and decreases, failing to support evidence for cumulative environmental impact.

Regarding alpha diversity, no cumulative environmental impact of the development of the mountain wind power complex on avifauna was found in terms of beta diversity. As shown in Figure 4, due to the start of direct construction within the wind power complex in YY, there was a change of about 5% in MDS1 and about 7% in MDS2 compared to the values found before construction (YY10). After one year, MDS1 restored to its original state by about 1.5% and about 1.2% in MDS2 (YY11). After another year, it restored to the original state by about 2% in MDS1 and about 5.2% in MDS2; after about two years (YY12), it resorted to a level very similar to that found before construction. After half a year, it changed in the opposite direction due to the impact of the construction (YY13F), but after one year (YY13S), it became similar to that found before construction. This suggests that some time is needed for stabilization to be achieved after construction is completed. The following year, indirect impacts occurred due to the construction of another wind power complex located north of YY, and the changes were approximately 7.6% for MDS1 and 5.5% for MDS2 from YY13S, indicating that the indirect impacts from adjacent development projects were not significantly different from the direct impacts from direct development projects. These two wind power complexes were located on mountain ranges connected in the north and south, which resulted in these findings. From these results, we were able to clearly and intuitively demonstrate that indirect development impacts, similar to direct development impacts, would occur in adjacent areas in the future if construction was to be carried out in areas where mountain ranges are connected. Here, after one year, the impacts restored by approximately 0.8% for MDS1 and approximately 0.85% for MDS2, and again, after one year, they changed in the opposite direction to the impacts from the construction, showing high similarity to YY13S. Although there is a limitation in the absence of monitoring results after this, the conclusion from all of these facts is that there is a change of approximately 5–8% during construction, a slight restoration after one year, and a very large restoration after two years. In GS, where only 10 wind turbines were installed compared to the 41 installed in YY, the impact was significant one year after construction, but the impact was restored to a level similar to that before construction the following year. In addition, the impact was somewhat evident during the period when many management personnel were deployed for the initial operation after the completion of construction (GS23S), but the impact was restored the following year. In MC, where five wind turbines were installed, the impact was significant at the start of construction, but the impact was significant one year later, and the change at the time of completion was half that during construction. In the case of MC, the impact was relatively small compared to other wind farms, and the monitoring range was not wide, so the number of birds and species counted was very small, which may reduce the reliability of the similarity, but the tendency to restore one year after construction is clearly evident.

The gamma diversity graph in Figure 5, which shows the combination of alpha and beta diversity, also shows changes and restoration. The diversity of the three wind power complexes changes during construction, and the other two wind power complexes, except for MC, show a process of restoration during construction. After construction, a restoration trend was observed in all three complexes. In YY, the diversity increases during direct construction and decreases when indirectly affected by construction. This suggests that each type of construction may have different effects, but more cases must be assessed to analyze the clear cause.

5. Conclusions

This study is the first to reveal the cumulative environmental impacts and environmental restoration resulting from the development of mountain wind farms. The results of this study provide objective data, rather than proposing or opposing development. This approach can help practitioners unfamiliar with statistics or ecology reduce the potential problems of subjective judgments based solely on population data when conducting environmental impact assessments. Our proposed method allows anyone to easily visualize the process of environmental change and restoration due to construction, allowing anyone to objectively assess the environmental impacts.

Based on the methodology presented in this study, future environmental impact assessors and researchers will be able to think more flexibly and conduct intuitive research on other taxa.

The methodology proposed in this study adopts a community-ecological or macro-ecological perspective, differing from existing methods that closely monitor individual species. However, the two methodologies are complementary and, when implemented together, can provide a broader perspective.