토지이용정보를 활용한 미시적 토지수요예측 연구

Land use forecasting takes significant meanings in urban planning, now that urban planning is basically based on the forecasts on the futures. Since 1990, many studies and researches concerning land use forecasting have steadily been published in Korea. Most of them, however, have the following limits in common; microscopic land use forecasting was not available because they take a broad view (macroscopic) on land uses: their hypotheses are not satisfyingly reliable as most of them depended on the rule-based algorithms. In this background, this study aims to suggest a microscopic land use forecasting model with considerations to socioeconomic factors. This study also aims to provide methods to evaluate probability models by sampling land use measurements (by 3%) from forecasting results, so that the probability models can contribute to the forecasting on a broader land units. In addition, this study evaluates the probabilities of land use changes on a given land unit, and verifies whether the resulted probabilities can fundamentally contribute to the decision making processes of urban planning and/or urban development. This study evaluates the land use change probabilities in 2015 of Dae-gu metropolitan area, using the area's past land use pattern data sets (of 1995 and 2005), through multinomial logit model. The urban land use patterns are categorized into two types; “residential” and “non-residential”, and the standard size of land unit is set to 30 meters ×30 meters grids. According to the results, 1.17% of the metropolitan area in 1995 has converted to residential area in 2005, while 2.39% to non-residential. The results also indicate that neighborhood characteristics have strong effects on land use changes: clustering is significant over space, but high residential densities can impede future development of non-residential land use. The area's land use change probabilities in 2015 are forecasted on each grid, with an assumption that the land use change patterns within the last 10 years will extant.