关键词: 城市交通, 郊域轨道交通, 土地利用演变, 贝叶斯网络, 开通前, 运营期

Abstract: Bayesian network (BN) is selected to study the land use change (LUC) caused by suburban-to-suburban rail transit. The influence area was set as the reference area, and the difference between LUC and suburban lines was analyzed. The case study of Tokyo Metropolitan shows that the suburban-to-suburban line has a great impact on LUC, especially on the forest, construction and sports land (operation period) and forest, sports, and water (before opening). The differences between the buffer zone and the reference zone are -32.08%, 12.59%, 8.24% and -24.23%, 4.62%, 4.40% respectively. However, the suburban line has no impact. The results show that, different from the multiple driving force to urban LUC, the suburban- to- suburban line as a strong driving force to suburban LUC causes significant changes in LUC. which can provide a basis for promoting suburban land development through the construction of suburban line. Due to the maturity and better inheritance of urban land use (LU), the current LU of suburban line is related to the previous LU at a high degree; As a strong driving force of suburban LUC, LUC is more active during operation of suburban-to- suburban line, so the current LU is less related to the previous LU at a high degree, which can provide a reference for the optimal time of land developing around the suburban-to-suburban line.The result of BN reasoning shows that the final state of the buffer zone during the operation period of suburban-tosuburban lines is construction land, which is less affected by the degree of construction land aggregation than the reference zone, indicating that the evolution of the buffer zone during the operation period of suburban line is relatively independent, which is also caused by the more activity of LUC caused by the suburban-to-suburban line; In order to support the construction land in the final state during the operation period of suburban-to-suburban line, 87.2% of the area should be buffer zones, and 98.5% of the initial neighborhood land should be non-forest land. The above can provide a reference for the layout and site selection of construction land in the process of suburban-to- suburban line land development. Taking the land use in 2006 of Duomo line as the initial state, the land use in 14 years is predicted, and the accuracy of the result is 89.9%, which indicates that BN has strong reasoning ability.

Key words: urban traffic, suburban-to- suburban rail transit line, land use change, Bayes network, before operation; operation period