Patches are recognized as ecotones as transition zone between adjacent patches that exhibit heterogeneity due to differences in vegetation conditions. Ecotones play an important role in environmental ecology by providing high biodiversity, ecosystem c...
Patches are recognized as ecotones as transition zone between adjacent patches that exhibit heterogeneity due to differences in vegetation conditions. Ecotones play an important role in environmental ecology by providing high biodiversity, ecosystem connectivity, and diverse habitat environments. Since South Korea has experienced spatial changes in patches due to rapid industrialization and urbanization, preservation of ecotone depends on confirming the impact of human activities on the natural environment. Therefore, in order to devise sustainable management measures, we tried to monitor the ecotone vegetation dynamics that change due to urbanization and evaluate the extent of impact. This study proposed an impact assessment tool to predict and quantify the range that changes under the influence of urban development projects according to the set peripheral distances (25, 50 m, and 100 m). Normalized Difference Vegetation Index (NDVI) and Vegetation Health Index (VHI) were selected as indices for evaluating the main effects, as well as Landsat and Sentinel-based satellite imagery data were calculated through the Google Earth Engine platform (GEE). Land cover maps provided by the Environmental Spatial Information Service as well as average temperature and precipitation data of the Korea Meteorological Administration were constructed through ArcGIS 10.5. National inventory data of the Environmental Impact Assessment Information Support System (EIASS) were processed and applied as variables. The data analysis method evaluated the vegetation distribution patterns of the research sites using the Artificial Neural Network (ANN) and Random Forest (RF) machine learning algorithms, and it was set to predict the range of influence on the vegetation index according to the ecotone multiple buffer size. As a result of the analysis, it was confirmed that NDVI was mainly concentrated on high values after the urban development project, while VHI tended to have a high pre-project value, which turned to the opposite trend. This can be interpreted as a significant result of the establish of new urban green spaces in accordance with the provisions of the Act on the Expansion, Management, and Creation of Urban Green Areas for Urban Landscape Planning. As a result of the performance of the machine learning models, the RF model showed the optimal predictive performance in both vegetation indices and along the ecotone distances. The modeled probability heatmap shows significant results at 90% confidence level (p<10%). Moreover, significant results were obtained when comparing the observed and predicted values visualized using the assessment tool. Both NDVI and VHI showed the tendency of the impact of the target site due to urban development to reach a maximum distance of 50 m. This proposal of quantitative evaluation tools is meaningful in that it may emphasize the decisive role of environmental impact assessment in terms of vegetation management by providing information on regional ecological restoration. It is expected that the extent of impact on the vegetation environment by urbanization can be identified to support the project plan while minimizing the loss of vegetation cover.