17:15 〜 19:15
[HDS08-P03] Improved agent-based intra-regional relocation model considering population and urban dynamics and its application to flood risk assessment in the Kyoto basin
キーワード：エージェントベースモデル、洪水暴露評価、人口動態、都市計画
INTRODUCTION
Integrated flood risk management (IFM) is a combination of various risk reduction measures, i.e. structural and non-structural measures. It has gained public attention due to increasing severe flooding damage in recent years. One important aspect of IFM is to reduce flood exposure such as the number of exposed households and residential buildings along rivers. Although land-use regulations in high-risk areas are a powerful policy to reduce the exposed population, it is practically difficult to conduct land use regulations in small countries such as Japan because most cities have been developed in flood-prone areas. Instead, urban planning is expected to guide people to live in areas that are as safe as possible, for example, by developing low-risk areas. In such guided urban planning, the spatial distribution of households changes dynamically. Since climate change and river management also change over time, it is expected that all these influences should be considered dynamically to discuss future flood risk.
STUDY OBJECTIVES
Our previous study developed an agent-based model (ABM) that represents dynamic changes in housing distribution over time. The model can simulate the influence of time-varying factors by reproducing the behaviors of agents such as individual households and developers and their interactions. Furthermore, it successfully improved the reproduction accuracy by incorporating transport accessibility as an index of the comfort of living. However, the model has a limitation that the boundary conditions are unable to treat temporal changes in urban development, transport accessibility and total population, which makes it difficult to estimate long-term household location choices.This study extends the model so that it considers temporal changes in population and urban structure development, which have traditionally been treated as fixed. The extended model is tested with the temporal population changes, the construction of new railway stations and new town developments in the Kyoto Basin between 2000 and 2015.The Kyoto Basin is located at the confluence of four rivers and features topography with high flood risk. The number of households increased from approximately 490,000 to 600,000 between 2000 and 2015, and significant changes have occurred in the spatial distribution of households and housing market, reflecting developments of new towns and railway stations.
SUMMARY OF RESULTS
1. Reproducibility of the extended model
We checked the validity of the model by comparing simulation results with the observed data on household distribution and housing prices
- The model successfully reproduced the distribution of households associated with new town developments and new railway stations. The simulated households showed a high correlation with observed data.
- The simulated housing prices were better than the previous ones, but it was difficult to adequately reflect local effects such as price surges around new railway stations and new towns.
2. Calculation of flood risk
We simulated the number of flood-exposed households and the amount of flood damage and assessed the impacts of urban development.
- The number of flood-exposed households was simulated to increase with urban development in high-risk areas, while a certain reduction in the number of households due to the inducement effect on low-risk areas was also simulated.
- The model under/overestimated potential flood damage amount due to difficulties in accurately simulating house price fluctuations observed nearby new railway stations and in new towns.
Integrated flood risk management (IFM) is a combination of various risk reduction measures, i.e. structural and non-structural measures. It has gained public attention due to increasing severe flooding damage in recent years. One important aspect of IFM is to reduce flood exposure such as the number of exposed households and residential buildings along rivers. Although land-use regulations in high-risk areas are a powerful policy to reduce the exposed population, it is practically difficult to conduct land use regulations in small countries such as Japan because most cities have been developed in flood-prone areas. Instead, urban planning is expected to guide people to live in areas that are as safe as possible, for example, by developing low-risk areas. In such guided urban planning, the spatial distribution of households changes dynamically. Since climate change and river management also change over time, it is expected that all these influences should be considered dynamically to discuss future flood risk.
STUDY OBJECTIVES
Our previous study developed an agent-based model (ABM) that represents dynamic changes in housing distribution over time. The model can simulate the influence of time-varying factors by reproducing the behaviors of agents such as individual households and developers and their interactions. Furthermore, it successfully improved the reproduction accuracy by incorporating transport accessibility as an index of the comfort of living. However, the model has a limitation that the boundary conditions are unable to treat temporal changes in urban development, transport accessibility and total population, which makes it difficult to estimate long-term household location choices.This study extends the model so that it considers temporal changes in population and urban structure development, which have traditionally been treated as fixed. The extended model is tested with the temporal population changes, the construction of new railway stations and new town developments in the Kyoto Basin between 2000 and 2015.The Kyoto Basin is located at the confluence of four rivers and features topography with high flood risk. The number of households increased from approximately 490,000 to 600,000 between 2000 and 2015, and significant changes have occurred in the spatial distribution of households and housing market, reflecting developments of new towns and railway stations.
SUMMARY OF RESULTS
1. Reproducibility of the extended model
We checked the validity of the model by comparing simulation results with the observed data on household distribution and housing prices
- The model successfully reproduced the distribution of households associated with new town developments and new railway stations. The simulated households showed a high correlation with observed data.
- The simulated housing prices were better than the previous ones, but it was difficult to adequately reflect local effects such as price surges around new railway stations and new towns.
2. Calculation of flood risk
We simulated the number of flood-exposed households and the amount of flood damage and assessed the impacts of urban development.
- The number of flood-exposed households was simulated to increase with urban development in high-risk areas, while a certain reduction in the number of households due to the inducement effect on low-risk areas was also simulated.
- The model under/overestimated potential flood damage amount due to difficulties in accurately simulating house price fluctuations observed nearby new railway stations and in new towns.