증거기반 국토계획 수립을 위한 토지이용변화 예측 및 모니터링 : Markov Chain과 MLP 기법을 중심으로

토지이용 변화 예측·모니터링을 국토계획 수립에 활용하는 방안을 연구한다.

South Korea faces limitations in land availability. Therefore, preventing urban sprawl and managing land resources efficiently is essential to preserve them for future generations. Territorial monitoring is essential in this effort, as it identifies land-use changes based on empirical evidence and integrates this information into policy development. However, despite its importance, territorial monitoring has not effectively incorporated predicted land-use changes into policy-making, and territorial planning remains largely normative. This study aimed to predict land-use changes using time-series data and to incorporate these projections into national land policies.

It forecasts the 2030 land-use for 30 cities in Gyeonggi Province, which are expected to experience development pressure, based on land cover data. The prediction process employed the Markov Chain and Multi-layer Perceptron (MLP) models. The analysis revealed that built-up areas would increase across all regions of Gyeonggi Province by 2030, although the magnitude and spatial patterns of growth differed by city. Additionally, indiscriminate urban expansion was observed at both metropolitan and municipal levels. These results suggest that national land policies should be tailored to each region, reflecting the diverse patterns of land-use change.

Furthermore, the land-use changes resulting from these policies should be continually re-evaluated and incorporated into future forecasts, thereby ensuring ongoing monitoring.