ABSTRACT

The process of urbanization differs in different parts of the world. Whereas, cities in the developing world are growing at a rapid pace primarily due to the rural urban migration, cities in the developed world are facing different patterns of growth namely suburban sprawl. The size and organization of the cities itself lead to wastage of physical resources needed to support it and damage of its own and its hinterland's environment. Besides, environmental concerns such as air and water pollution, etc. are featuring seriously in the minds of the urban policy makers for the first time. Thus, traditional urban land use and transportation planning process is not enough to plan for our future cities where sustainability would be the key, requiring the intervention of multi-disciplinary approaches.

Travel Demand management (TDM) and Transit-oriented development (TOD) has been the focus of sustainable urban transportation research for the last few decades. Numerous studies have recognized the importance of enhancing transit connectivity at the city scale and at the same time increasing development density and mixed-use development at the neighborhood scale to enhance transit usage. Enhancing development density at the neighborhood scale not only promotes modal shift towards transit use but also increases the modal share of walking and bicycle use for different trip purposes whereas, mixed-use development at the neighborhood scale has a profound influence on the urban structure and form thus effecting location choice by households which in turn influences trip generation and distribution. Increasing development density within the urban area also reduces urban sprawl and suburbanization thus increasing land use efficiency.

Development density of a particular neighborhood can be enhanced either through increasing the built-up area percentage or through increasing the height of existing buildings or a combination of both. Neighborhood environmental concerns such as amount of open area, parks and greens becomes the decisive factor for increasing the built-up area percentage whereas, people's preference for living or working in high rise buildings and construction cost becomes the decisive factors in increasing building height. This is where the TDM design problem is converted to a multi-criteria planning and design problem at the neighborhood level. This is referred as the land use design problem (LDP) in literature and the resulting mathematical formulation is a multi-objective optimization problem. Finally, the land use design problem could be further extended to take into considerations city level design targets for natural support systems such as parks and greens, agriculture, water bodies etc. thus incorporating concerns for urban food security and livability into the urban land use transportation planning process.

The basic objective of this research is to model the interaction between urban land use, building floor area density, population density, transportation and the natural support systems and provide an integrated analytical planning model for urban scale application. The specific objectives are as follows:

・Comprehensive analysis of how built environment characteristics influence modal choice at the neighborhood level for different trip purposes.

・Incorporating a proactive component in the urban land use transportation planning process in form of an optimization model at the neighborhood scale to evaluate neighborhood planning targets.

・Integration of the different model components such as residential location choice, trip generation, trip distribution, modal choice and neighborhood optimization for city scale simulation.

・Development and application of the models for Nagoya.

・Estimating, evaluating and proposing long terms plans for urban structural modifications towards our vision an efficient urban form for Nagoya using the modeling framework developed in this research thus attaining reduction in total travel cost, travel time and CO2 emission levels in the process.

The present model utilizes an integrated modeling approach using multi-objective optimization to reconcile the trade-offs among various urban planning and environmental parameters at the neighborhood scale and simultaneously simulating the land use, building use and transport dynamics, both at the neighborhood and city scale using aggregate models such as the multiple regression models for trip generation and attraction, discrete choice models such as the logit model for modal split and a multi-objective optimization model for land use and building use plan generation. The sub-models developed could also be applied independently or in combination to evaluate other planning targets at the city or neighborhood scale.

At first, the case study area, Nagoya city is introduced. The city is then divided into zones or neighborhood for the development of our model. Then, detailed databases were developed for each neighborhood on demographic, built environment, trip and network characteristics. In the next step, linear multiple regression models were developed for different trip purposes using neighborhood parameter estimates like total floor area for building categories, area for land use categories, population etc for predicting trip production and attraction from the neighborhoods. For trip distribution, the standard gravity model formulation was adopted and calibrated for Nagoya.

In the next step, modal split models were developed for 6 broad trip purposes in Nagoya. The modal split model adopted in our study is a post trip distribution modal split model based on the discrete choice framework. We adopted a two stage model with a binary choice at each stage. The first stage (Level 1) explained the mode choices between slow and fast mode and the second stage (Level 2) explained the mode choices between car and transit mode. A comprehensive set of built environment characteristics at both the trip origin and destination were evaluated to determine the influence of built form on modal choice.

The next model developed is the neighborhood regeneration model. Based on TDM planning objectives, neighborhood environmental concerns and economic efficiency four objectives were formulated: maximizing transit use; maximizing walking and bicycle use; minimizing built-up area coverage; and minimizing total transport cost. The decision variables used in the model were average building height, neighborhood building floor area density and % of existing buildings to be demolished for reconstruction. Five types of constraints were used in the development of the model. The first constraint is the lower bound of building floor area density. This constraint is set by considering the existing density of the building categories. The second constraint is the upper and lower bounds on percent of buildings to be demolished. This constraint is set by considering the percent of buildings which have reached the end of their service life (lower bound) and feasible percent of buildings that could be demolished (upper bound). The third constraint is the upper and lower bounds on average building height. This constraint is set by considering the existing average building height of the building categories (lower bound) and people's preference for living and working in high-rise buildings (upper bound). The fourth constraint is the upper and lower bounds on total cost of building construction. This constraint is set as 0 (lower bound) and cost benefit achieved by minimization of transportation cost (upper bound). The fifth constraint is the lower bound on land use savings. (Built-up area decrease due to higher density reconstruction) This constraint is set by considering the land use design target. Seven building categories are investigated in our model. The model planning period was fixed as 5 years which is usually also the period between subsequent PT surveys undertaken in Nagoya. Another general assumption in our model is that service life of buildings in Japan is 30 years. All economic and environmental benefits are also evaluated over the same period. Considering Nagoya's case, we assumed additional floor area demand for a building category to be absorbed within the existing land use for that building category which would actually make the reconstruction activity profitable for developers. One limitation of our optimization approach is that, neighborhood density at trip origin is only considered for modification whereas; modal split is affected by both neighborhood densities at trip origin and trip destination. To address this issue we have used an iterative approach for neighborhood optimization. The resulting optimization problem is a nonlinear, non-convex programming model. Premium Solver Version 7 software was used for optimization and the method employed was 'Standard Interval Global'. Even though, minimizing total transportation cost has been used as the primary objective, minimizing transport CO2 emission or total CO2 emission from both transportation and the building sector could also be also used as an objective function in this model.

In the next step, a residential location choice model for rented multi family housing location choice and a multiple regression model for estimating new demand for commercial floor area in the different neighborhoods were estimated.

Finally, the model framework was applied to Nagoya. The simulation framework is developed at the beginning. Then, baseline estimates were obtained for trip characteristics for all the neighborhoods in Nagoya. The next step involved selection of neighborhoods for application of the neighborhood regeneration model based on our planning targets. The neighborhood regeneration model was then applied to 66 neighborhoods. Three objective functions within the neighborhood regeneration model, minimizing total transportation cost, minimizing transport CO2 emission and minimizing total CO2 emission from both transportation and building sector was evaluated separately using three sets of calculations for the 66 neighborhoods selected for regeneration. Finally, estimates for trip characteristics and built environment characteristics were obtained along with related costs and benefits for all the selected neighborhoods for regeneration for the three different evaluated objectives.

The following general conclusions could be drawn from the present research.

Neighborhood built environment characteristics play an effective role in determining travel behavior of the population. However, their effect on mode choices at the neighborhood level not only varies according to the mode but also according to the trip purpose. Thus, adopting ad-hoc decisions to modify built environment characteristics at the neighborhood level may result in unexpected effects. For example, increasing residential density at a zone to increase modal share of transit use may actually result in decrease of transit rider ship due to modal shift to slow modes such as bicycle or walking. Thus, the effect though beneficial may result in unexpected complications like underutilized transit infrastructure.

Neighborhood built environment characteristics at trip origin was found to be more effective in influencing probability of choosing slow mode than neighborhood density at trip destination. However, for transit mode choice, neighborhood density at trip destination is more effective than neighborhood density at trip origin. Thus, it would be more effective to increase business and commercial density at transit destinations to increase transit mode choice.

Built environment characteristics and particularly, development density of a particular neighborhood can be enhanced to improve the modal share of slow and transit modes. However, actual physical barriers, like limited plot size for high rise development or due to the existing age of buildings which may be relatively new and people's willingness to change, the effective improvement in density may be insufficient to influence significant changes in modal choice. Thus, long term planning targets for improving the development density at the neighborhood level should be adopted.

Environmental concerns such as amount of open area, parks and greens at the neighborhood level and CO2 emission both at the neighborhood and city scale from the transportation and building sector, amount of land area required for natural support systems for agriculture, forestry etc, could be evaluated as fixed values or as decision variables using the present modeling approach.

Finally, the present modeling approach could be utilized to generate a more efficient urban form in terms of travel time, travel cost as well as CO2 emissions as was demonstrated in the case of the model application in Nagoya city.

An analytical model for integrated planning at neighborhood scale and transportation planning at the city scale has been left unexplored due to the complexity of the interactions between these two design problems. The present modeling approach can be one of the useful tools for planning for TDM and TOD, allowing planners and decision makers to make more informed choices about F.A.R. (Floor area ratio) and B.A.R. (Building area ratio) regulations, construction of new transit lines and design policies for changes needed within the neighborhoods in terms of land use, building use, building floor area and density.

環境負荷の小さい持続可能な都市を形成するために都市工学の分野ではさまざまな研究が行われてきた。その中で、二酸化炭素の発生源になっている自動車交通由来の環境負荷を削減するための都市計画や、二酸化炭素排出の小さい居住形態の実現などが検討されているものの、それらは個別に考えられている。交通と住宅・業務施設の立地は本来深い関係を持つにもかかわらず、それらの相互関係を含めた形での都市の解析は十分になされていない状況である。そのような解析は、コンパクトシティの形成など、持続可能な都市を実現する上で重要である。

本論文はこのような背景の元に行われた研究の成果をまとめたもので、Neighborhood scale optimization for density, transportation, land use and the natural support systems towards an efficient urban form(効率的な都市形成に向けての密度、交通、土地利用、自然システムの地区単位での最適化)と題し、10章からなる。

第1章は序論である。従来から提案されてきた交通需要管理(TDM)、公共交通指向型都市開発(TOD)などの対策は有効ではあるが、土地および建物利用と交通を組み合わせたプロアクティブな解析は十分に行われていない状況である。そのため、将来の望ましい都市の構造が具体的には示されていない。このような認識を示した上で、土地利用、建物利用、床面積密度、人口密度と交通、緑地の関係をモデル化し、実際の都市に適用するという本研究の目的を示している。

第2章は既往の研究の整理である。

第3章は、本研究で用いる基本的なモデル群の構造とそれらを用いた解析の流れについて整理したものである。本研究では、(1) 発生・集中交通量モデル、(2)分布交通量モデル、(3)交通手段選択モデル、という交通モデルに加えて、(4)近隣地区の単位で建物の建て替えを想定し、多目的の最適化を行う地区再生モデル、および(5)近郊部を含む同一都市内での住宅と商業の立地場所の移転を予測する立地選択モデルを構築して、これらを有機的に組み合わせ、あるいはフィードバックさせて用いることを示している。

本研究は名古屋市を対象にしており、同市の人口、建物情報とパーソントリップ調査結果を原データとして用いている。現状に対して、これらのデータを用いて上記(1)、(2)、(3)のモデルを作成し、その結果を用いて効率の良い都市の姿を提案するために、地区規模では(4)を、都市全体としては(5)を適用して将来の都市シナリオを定量的に示し、(1)、(2)、(3)のモデルでそれらの評価を行う、という基本的な方法をとっている。

このように、交通モデルと地区再生モデルや立地選択モデルを組み合わせたインタラクティブな将来予測モデルシステムを提案しているところが本研究の特徴である。

第4章は、対象とした名古屋市の基礎データについての整理である。

第5章は、発生・集中交通量モデルおよび分布交通量モデルについての解析である。

名古屋市の実際のデータに基づき、対象とする地区の人口密度や建物用途、密度などの地区特性との相関解析によって発生・集中交通量を求めるモデルを構築するとともに、地区間の移動容易性を考慮したグラビティモデルによって交通量を配分するモデルを構築している。

第6章は、交通手段選択モデルについての解析である。交通手段選択を、徒歩・自転車のような低速交通手段と、高速交通手段に分け、後者は更に公共交通機関と自動車に分けられる二段二項ロジットモデルによる交通手段選択モデルを採用している。ここでは、人口密度や地区間の移動容易性、乗用車の保有率などこれまで使われてきた地区特性ばかりでなく、交通発生地区や目的地の用途別床面積等もロジットモデルの変数として考慮している。交通発生地区の人口密度が低速交通手段を促進させるほか、目的地の床面積や居住密度が高くなるほど公共交通機関選択率が高くなることを示している。これらの相関関係が、地区の再生に伴う交通のエネルギー消費削減の評価の基本となっている。

第7章は地区再生モデルであり、本研究の独自のものである。このモデルは多目的の最適化を達成するために地区内の建築物の改築を想定するものであり、床面積密度の上限と下限、改築比率の上限と下限、建物高さの上限と下限、改築コストの上限、改築による空地創出量の下限、をそれぞれ制約条件として与え、最適化計算を行っている。このモデルは、都市の中で対象とする地区のそれぞれに対して計算を行うものである。

第8章は立地選択モデルについてであり、居住地の変更が容易な、借家集合住宅の居住者が利便性の高い地域に移転することを予測し、それに伴う商業立地の増加を予測するモデルである。

第9章は、前章までに述べたモデルを名古屋市に対して統合的に適用した解析結果について述べたものである。現状の土地利用、業務床面積と密度、居住床面積と密度、公共交通機関へのアクセス容易性の観点から、209の地区の中から再生の対象となる66の地区を選び出し、それらに対して地区再生モデルを適用した。また立地選択モデルによる郊外部からの人口の流入を予測した。これらの解析結果から、理想的な立地を行った際の費用削減と二酸化炭素排出削減量、緑地の創出量を求めた。

第10章は結論であり、結果を総括すると共に、今後の課題を述べている。

本研究は、都市構造の改変によって都市が与える環境負荷の低減を、現実の居住者の行動を反映したモデルによって評価したものである。日本の都市のように人口が一定もしくは減少に向かう都市であっても都市内部の構造の変換によって環境負荷を低減できることを示したものであり、持続可能な都市のあり方を定量的に評価する研究として評価される。

以上、本研究において得られた成果には大きなものがある。本論文は環境工学の発展に大きく寄与するものであり、博士(工学)の学位請求論文として合格と認められる。