This study analyzes the land-cover patterns, with spatial calculation of land-cover changes based on land-cover spatial information data and multi-temporal satellite imagery, for Cheonan-Asan City represented in pixel units. The change detection technique used in this study is empirical frequency distribution function. The study has obtained relative favorability score to determine and express how each factor affects land-cover change. And, this study used the relative favorability function to calculate the relative favorability score of spatial information data affecting land-cover change, in order to predict future land-cover changes at Cheonan and Asan City.
As a result, the four regular patterns in land-cover change were found. First, the most usual pattern in land-cover change is that forest becomes dry field and city, paddy becomes dry field and city, and dry field becomes city. The second pattern found is that there have been some cases wherein forest and paddy directly became city, but patterns with 10-year cycles were also noted, in which these areas become dry field first and then turn into city later on. The third pattern is that as the size of city increases, the area of forest and paddy decreases. However, dry field has not displayed much change in size but has shown more changes in locations. Finally, the patterns have been manifested more clearly in the 20-year-period between 1989 and 2009, rather than from 1982 to 1989. In addition, results show a relatively accurate prediction with an overall accuracy of 90.44% of the land-cover change prediction algorithm. The outcome obtained from predicting spatial land-cover change by combining relative favorability scores point to land-cover change from 2020 to 2040 indicates that the forests and paddies will be converted into dry fields and cities by 2020 according to the pattern of land-cover changes from 1989 to 2009. The surface area also displayed an increase in dry fields and cities. Forests, paddies and dry fields are predicted to change into cities by 2030 with a considerable increase in the surface area of cities. By 2040, almost no land-cover change is predicted. A constant decrease in the land-cover change rate is predicted during the years between 2020 and 2040. This is attributed to a discontinuation in urbanization from 2040, because the population increase rate has been predicted to be near 0 from 2040. It is expected that the prediction accuracy will be improved due to further studies on the algorithm combining relative favorability scores and studies on added value and selection of future land-cover change factors. Furthermore, such studies are expected to greatly contribute to establishing an efficient regional development plan if spatial prediction becomes possible in terms of land-cover change in other regions.