Skip to navigation – Site map

A Spatially-Explicit Scenario for Achieving “Wise Shrink” Toward Eco-Urbanism

Yoshiki Yamagata, Daisuke Murakami and Hajime Seya

Abstract

In order to achieve a sustainable and resilient eco-urbanism, we propose a “Wise Shrink” approach grounded on a consideration of trade-offs and synergies between city compaction and disaster risks. The objective of this study is to show the possibility of such scenarios in the Tokyo metropolitan area using a newly developed spatially-explicit urban land-use model (SULM). Here, a scenario that considers both city compaction and disaster risk reduction (Wise Shrink) is compared with a city compaction scenario without disaster risk reduction (Compact City) and a business-as-usual scenario (BAU). The results show that increasing green areas under the Compact City scenario would facilitate adaptation to climate change risks. However, there exist trade-offs for flooding and other risks depending on the location of the compaction. By harmonizing the risks, this study suggests a possible Wise-Shrink approach for achieving sustainable and resilient eco-urbanism.

Top of page

Full text

Introduction

1Urban shrinkage is becoming a usual path in developed countries (Großmann et al. 2013). Actually, Turok and Mykhnenok (2007) reveal that almost 42% of cities in Europe are shrinking. More and more studies are discussing concepts, agendas, and tools toward sustainable urban shrinkage (e.g. Haase et al. 2014). Since there is a wide variety of patterns in urban shrinkage, it is not possible to devise a ‘one-size-fits-all’ explanatory approach (Haase et al. 2016). In other words, a case-by-case assessment is necessary. Fortunately, numerous approaches to assessing sustainability in urban shrinkage, including top-down approaches (e.g. material flow analysis; see Ayres and Kneese (1969), bottom-up approaches (e.g. agent-based approach, Benenson 2004), and their hybrid (e.g. Chrysoulakis et al. 2013), have been proposed (Chen et al. 2014). The discussion on urban shrinkage typically focuses on how to design sustainable compact cities under depopulation, deindustrialization, suburbanization, and post-socialist transformation (Pallagst et al. 2013).

2One important aspect that should be considered when designing urban shrinkage/compaction is disaster risk. That is, if we try to increase the population of districts with a high hazard of natural disasters, then the disaster risk may increase. Consideration of disaster risk is increasingly important with the progress of climate change. However, disaster risk has not been considered in the context of urban shrinkage/compaction in previous studies, partly because disaster risk is not directly related to these urban processes. Unfortunately, unusual event risks including the risk of natural disasters are often subjectively underestimated (normalcy bias, see Omer and Alon 1994). A quantitative approach is needed to evaluate the risk trade-offs, that is, the trade-offs between disaster risks and other factors (amenity, natural environment, convenience, etc.).

3The above discussion suggests the importance of a risk-adaptive urban shrinkage/compaction scenario, which we call a “Wise Shrink” scenario. We define wise shrink as a scenario with the following properties: (a) it promotes city compaction; (b) it explicitly adapts to disaster risks; (c) it considers the risk trade-offs. Note that the Compact City scenario considers (a) only. Our study area is the Tokyo metropolitan area. Voss (2006) estimated that the insurance risk is highest for the Tokyo metropolitan area among megacities, globally.

4The subsequent sections are organized as follows. Section 2 reviews positive and negative aspects of compact city development. Section 3 briefly introduces methods for future scenario development, and section 4 introduces the spatially explicit urban land-use model (SULM) that we used in our scenario analysis. Then, section 5 demonstrates how to apply this model towards wise shrinking. Finally, section 6 concludes our discussion.

Positive and negative aspects of urban compaction

5Numerous studies have demonstrated the environmental, social, and economic benefits arising from compact city policies. Regarding environmental benefits, Taniguchi and Ikeda (2005), Kennedy et al. (2009), Baur et al. (2013) and Mishalani et al. (2014) empirically verified that compact population distribution decreases per capita greenhouse gas emissions from transportation sectors. The compact city also enables us to implement efficient smart grids and community heating at low cost. Besides, it can contribute to revegetation (e.g. Beatley 2012). Concerning social aspects, Kaido and Kwon (2008) clarified that city compaction increases the accessibility of residents to multiple urban facilities. Also, it lowers transportation expenses, and increases the mobility of low and middle income families (Howley 2009). With regard to economic benefits, Burchell et al. (2002) demonstrated that compact city policy saves a considerable amount of infrastructure costs. Morikawa (2011) showed that the service industry significantly increases its productivity by concentrating people in an area.

6By contrast, some studies have suggested negative aspects of compact city development. For example, Mindali et al. (2004) demonstrated that there is no statistically significant difference between CO2 emissions in compact cities and those in dispersed cities. This is partly because urban compaction increases access to shops, restaurants, and so on, and increases short-distance trips (Bouwman and Voogd 2005). Longden (2015) showed that the relationship between population density and CO2 emissions becomes positive after regional differences are controlled using a fixed effects model. Furthermore, while compact city policy encourages people to live in high density areas, Neuman (2005) demonstrated that residing in areas with high population density and limited natural resources lowers livability. Numerous hedonic studies (e.g. Rosen 1974) have also demonstrated that people prefer living in areas with an abundance of natural resources (e.g. Jim and Chen 2007, Cho et al. 2008). Moreover, the compact city does not specifically consider climate change adaptations (Williams et al. 2010). A central area with a high density of people and economic activities can be vulnerable to disaster risks, especially in developing countries (Dodman 2009).

7The above discussion suggests the importance of quantifying the influences of urban compaction on disaster risks. A number of methods that allow us to project future urban development have been developed. Some of them are useful in evaluating the effectiveness of compact city policies based on multiple criteria including disaster resilience. The next section briefly reviews approaches showing future urban scenarios.

Methods for creating urban scenarios

8Land-use models are useful to analyze the implications of future urban forms, such as compaction and dispersion scenarios (Yamagata et al. 2013). According to van Schrojenstein et al. (2011), land-use models can be classified based on the land-use change principle: (a) continuation of historical development, (b) suitability of land, (c) results of neighborhood interaction, and (d) results of actor interaction. Each type of model can also be constructed based on the method principle : (a) simply extrapolating the trend of past land-use changes into the future (e.g. White et al. 2009); (b) determining land use in a zone based on the characteristics of the zone, such as accessibility and soil (e.g. zones nearby railway stations would need a greater chance to become urban land); (c) using the cellular automata and Markov-Chain models to determine land use based on neighboring relationships (e.g. a zone surrounded by wasted lands tends to be a wasted land zone as well). Statistical land-use models considering both (b) and (c) in the former classification have been extensively studied in the last decade (e.g. Chakir and Parent 2009, Brady and Irwin 2011, Li et al. 2013). Recently, urban modeling approaches emphasizing agents’ (actors’) behavior (d) have become increasingly popular as computer performance advances.

9To develop persuasive urban scenarios, some studies introduced economic principles in their own land-use models (e.g. Brady and Irwin 2011, Koomen et al. 2015). The SULM (e.g. Ueda et al. 2013, Yamagata and Seya 2013) is one of such models that describes the utility maximization behavior of households and profit maximization behavior of developers and landlords in building and land markets. Since the late 1980s, the effectiveness of this model has been demonstrated in benefit evaluations of transportation policy (e.g. Sato and Hino 2005, Chen et al. 2013), land-use policy (e.g. Nakamichi et al. 2013; Yamagata and Seya 2013, Yamagata et al. 2013), and so on.

10Our objective is to model urban activities, and to reveal how to shrink urban areas wisely. The SULM is suitable for this purpose because it explicitly describes the activities of aforementioned agents and their behavioral change due to urban shrinkage/compaction (e.g. choice of residential location), and it captures behavioral change due to adaptation to disaster risks. Thus, we apply SULM to project a Wise Shrink scenario and other scenarios.

Spatially explicit urban land-use model (SULM)

Model

11Our SULM, which was applied in Yamagata et al. (2013), estimates the behaviors of households, developers, and landlords, aggregated into zone index by i{1, 2, ... N}. In our analysis, the zones are defined by 22,603 micro districts called cho-cho-moku in the Tokyo metropolitan area. In this model, households and developers transact in a building market whereas the developers and landlords transact in a land market. A partial equilibrium among these agents is described by modeling their profit maximization behaviors. In Japan, SULM, which is inspired by Anas (1982), is slightly modified to reflect the situation of the real estate market, i.e., that the land price market and the real estate market are separated. The subsequent subsections introduce the behavior models for households, landlords, and developers.

Households’ behavior

12The households’ indirect utility function in the i-th zone can be formulated as follows:

13where Vi denotes the maximized indirect utility, yi is the average income per person in zone i, ri is the residential floor rent per area, ci is the generalized cost of a private trip, and αa and αx are unknown coefficients. Eq.(1) suggests that the utility of households increases if the households have a large income and low floor rent and travel costs. Due to the limitation of data availability at the micro district level, we assume that yi and ci are exogenous variables (i.e., fixed variables). The floor rent, ri, is estimated by maximizing Vi. αa and αx are specified based on Roy’s identity equation as follows (see Yamagata and Seya 2013):

14where ai is the per person consumption of residential floor area in zone i, zi is the per person composite good in that zone, and αz is a coefficient, which is estimated by applying the ordinary least squares (OLS) for the third equation in Eq.(2).

15Once Vi is estimated by using the OLS estimates of αa and αx, and ri, which is estimated by the developer model (see below), the residential location choice behavior of household type h (e.g., one-person, married couple, and so on; see Table 1) is modeled as follows:

16where fi is a vector of location-specific attributes, δ[h] and are coefficient (scalar) and coefficient vector, respectively, for h-th household type, and Pi[h] denotes the ratio of households with type h that select i-th zone as their residential location. δ[h] and can be estimated by aggregated logistic regression. Pi[h] is updated using the estimates of δ[h] and .

17Once Pi[h] is updated, the floor area demand in the i-th zone is estimated as follows:

18where is the total number of households with h-th type, and si[h] is the number of h-th households in the i-th zone.

19One problem is the collinearity between rent ri and population Pi. Since ri influences Vi negatively (see Eq.1), and Vi influences Pi[h] positively (see Eq.2), ri and Pi[h] are implicitly assumed to be negatively related. However, they are likely to be related positively in the Tokyo metropolitan area, where people tend to be concentrated in high-rent areas. To cope with this problem, we need to control the quality of floors (Yamagata and Seya 2013, Yamagata et al. 2013), e.g. using the pure-repacking model of Walsh (2007) and the quality-adjusted price (e.g. Mehta 2007). Following Yamagata and Seya (2013), we applied the former.

Developers’ behavior

20Developers are assumed to maximize their own profits following Eq.(5):

21where [DH] denotes developers. Ai[DH] is the floor area supplied by developers, Li[DH] is the land area demanded by developers, and pi is the land rent per unit. Ki is the material inputted for the production of floor service, m is the price for material construction, and μ1, μ2, v are unknown parameters. These parameters can be estimated by applying OLS for equations derived by solving Eq.(5) (Yamagata and Seya 2013).

Absentee landlords’ behavior

22It is assumed that landlords’ behavior is to maximize their profit based on Eqs.(6), (7).

23where [LH] indicates landlords’ behavior, and Li[LH] denotes landlords’ residential land supply. C(Li[LH]) denotes the land maintaining cost, in which is the area of available land and σi denotes a parameter that is obtained during the calibration to reproduce the observations.

Calibration

24Figure 1 summarizes the functional relationships among factors in SULM. This model estimates population distribution, rent, demand/supply of building (floor area), and demand/supply of land (land area), which are shown in red in Figure 1, under equilibrium conditions in building and land markets. These equilibriums are formulated by Eqs.(8) and (9), respectively:

25Eq.(8) associates the households’ behavior and the developers’ behavior, whereas Eq.(9) links the developers’ behavior and the landlords’ behavior. Once exogenous variables, which are shown in black in Figure 1, are provided, the endogenous variables shown in red are estimated by iterating the maximizations of the objective functions of each agent sequentially until the iterations converge.

Figure 1: Image of SULM. Red: endogenous variables to be calibrated; black: exogenous variables that are assumed to be fixed.

Figure 1: Image of SULM. Red: endogenous variables to be calibrated; black: exogenous variables that are assumed to be fixed.

Source: authors.

Figure 2: Location of the Tokyo metropolitan area (left) and its population distribution (2005; right)

Figure 2: Location of the Tokyo metropolitan area (left) and its population distribution (2005; right)

Source: authors.

Application of SULM towards a wise shrink

Outline

26In this section, SULM is applied to estimate flood and earthquake risks and revegetation in 2050 under several future scenarios including a Wise Shrink scenario. The target area is the Tokyo metropolitan area that is the most populated metropolitan area in the world, with a population of 3.6 million in 2005 (Figure 2). We implemented the SULM in micro districts (N = 22,603) that compose this area.

27Table 2 summarizes the data used in SULM. Among them, land rent and floor rent were predicted from point-referenced land-rent and floor-rent data (Table 2), respectively, using ordinary kriging (Cressie 1993), which is a standard spatial interpolation method. Land areas in each micro district were allocated by distributing the municipal land areas (source: Fixed property tax cadaster) based on the standardized product of population number and area. The available residential land area was prepared using the area zoned for residential use in the City Planning Law, Japan (Ministry of Land, Infrastructure and Transport 2003).

Table 1: Variable description

Notation

Variables

Data source (Agency)

Pi[h]

Number of people in the following household types:

National census (MIC 1)

[1] One-person (65 years of age or over)

[2] One-person (under 65 years of age)

[3] Married couple only

(either of them 65 years of age or over)

[4] Married couple only

(under 65 years of age)

[5] Married couple with child(ren)

[6] Single parent with child(ren)

[7] Other types

ci

Total available time/Wage rate

Surveys on time use and leisure activities (MIC)

yi

Income

Statistical survey of actual status for salary in the private sector (National Tax Agency)

pi

Land rent per unit

Officially assessed land price (MLIT 2)

ri

Floor rent per unit

Provided by a private company (At home corp.)

Li,

Land area

Fixed property tax cadaster (MIC)

Available land area

Ai

Floor area

Ki

Material price for construction

Statistics on building material and labor demand (MLIT)

Building construction navigation

(Construction research institute)

Material inputted for production of floor

Statistics on building construction (MLIT)

f

Average elevation

National land numerical information (MLIT)

Average slope

National land numerical information (MLIT)

Liquefaction risk

Wakamatsu et al. (2006)

Earthquake risk

Japan seismic hazard information station (NIED3)

Building-to-land ratio

UDS Co., Ltd.

Floor area ratio

Proportion of residential zone

Proportion of urbanization control zone

Proportion of commercial zone

Office density (number of offices/area)

Establishment and enterprise census (MIC)

Spatially weighted average number of offices4

National land numerical information (MLIT)

1 MIC: Ministry of International Affairs and Communications, Japan.

2 MLIT: Ministry of Land, Infrastructure, Transport and Tourism, Japan.

3 NIED: National Research Institute for Earth Science and Disaster Prevention

4 We imposed inverse distance weighting. The weight for the zone in itself is set to zero (LeSage and Pace, 2009).

28Voss (2006) estimated that Tokyo is the city with the highest insurance risk among megacities in the world. Indeed, the Tokyo metropolitan area is projected to suffer from a great earthquake called Nankai-Trough earthquake within decades. Besides, the central area includes low elevation areas whose inundation risk is high. Considering the further increase of the risk of extreme events due to climate change projected in Asian countries (e.g. Pachauri et al. 2014), increasing resiliency and adaptability to disaster risks is an urgent task in the Tokyo metropolitan area. On the other hand, depopulation is another problem in Japan. Although the population around Tokyo is still growing, it is projected to decrease from around 2020. Accordingly, urban compaction is also an important issue in this area.

29Based on the above understanding, we developed the following scenarios for 2050 emphasizing urban compaction and disaster risk adaptation: the Business-As-Usual scenario (BAU); the Compact City scenario (Compact City); and the Compact City and Adaptation scenario (Wise Shrink). In BAU, the number of households in each type (Table 1) in 2050 was projected by extrapolating the population estimate for 2030 (source: National Institute of Population and Society Research, Japan) using the log-linear extrapolation method (Smith et al. 2006).

30In the Compact City and Wise Shrink scenarios, the city centers to be intensified were determined by applying the Moran scatterplot (Anselin 1996) for the office density (see Table 1). Suppose that zi is the office density in zone i, and wi,j ( and ) is the spatial connectivity between zones i and j. Then, this method plots against zi. The plot revealed two types of office agglomeration zones: HH: high density zones with high density neighborhoods (i.e., both zi and are high); HL: high density zones with low density neighborhoods (i.e., zi is high and is low). Figure 3 plots the estimated HH and HL zones. This study assumes that HH or HL zones are central areas. In the Compact City scenario, people living within 500 m of central areas are subsidized by 1,200 USD/year, which is the same amount as the subsidy in the compact city policy in Toyama, Japan (see Section 1).

31In the Wise Shrink scenario, disaster risk adaptation is additionally introduced. Figure 4 displays the anticipated flooding depth (source: National Land Numerical Information [MLIT]) and the probabilities of suffering from earthquakes with seismic intensities exceeding 6.5 within 30 years (source: Japan Seismic Hazard Information Station [NIED]). We assumed a land-use regulation that reduces available residential land by 50% in areas whose flooding depth is more than 0.5 m. However, we did not explicitly impose any regulation for the earthquake risk because the risk is quite uncertain. Actually, high-risk areas of floods and earthquakes have similar distributional properties (Figure 4). So, the land-use regulation for flood risk adaptation might also reduce the damages in an earthquake.

Figure 3: Estimated office agglomeration clusters (black) and subsidized areas (incentive zones).

Figure 3: Estimated office agglomeration clusters (black) and subsidized areas (incentive zones).

Source: authors.

Here, High-High or High-Low means high-density zones with high- or low-density neighborhoods.

Figure 4: Flood hazard (left) and earthquake hazard (right).

Figure 4: Flood hazard (left) and earthquake hazard (right).

Source: authors.

Estimated populations and building land areas

32The SULM estimates the populations, building areas, floor areas, and floor rents in each zone in 2050 (see Figure 1) under the BAU, Compact City, and Wise Shrink scenarios. The population distributions estimated in each scenario are plotted in Figure 5 (top). As a result of depopulation, the population in 2050 (BAU) is smaller, particularity in suburban areas (see Figure 2). Compared with the BAU scenario, the Compact City scenario has a larger population in the center of Tokyo and in the centers of suburban areas.

33The Wise Shrink scenario also shows a larger population in the suburban sub-centers, but not in the center of Tokyo because of the flood risk. Wise shrink decreases the population by 23,996 people inside the area with a flooding depth of more than 0.5 m (see Figure 4), while the Compact City scenario increases the population by 1,617 people inside the anticipated flooding area. Besides, as analyzed in Yamagata et al. (2016), the Wise Shrink scenario reduces the population who are expected to suffer from earthquakes with seismic intensities of more than 6.5 by 11,978 people by 2050, whereas the Compact City scenario increases the population expected to suffer from the earthquakes by 147 people. Thus, the Wise Shrink scenario is more risk adaptive.

34Figure 5 (bottom) displays a significant decrease in building land in suburban areas in the Compact City and Wise Shrink scenarios. Yamagata and Seya (2013) showed that the decrease significantly reduced CO2 emissions and energy use, and mitigation of urban heat island effects. In summary, both the Compact City and Wise Shrink scenarios increase sustainability, although the latter is preferable in terms of climate resilience.

Figure 5: Estimated populations in 2050. This figure displays the population in BAU (left), differences of populations between Compact City and BAU (middle), and the same between Wise Shrink and BAU (right).

Figure 5: Estimated populations in 2050. This figure displays the population in BAU (left), differences of populations between Compact City and BAU (middle), and the same between Wise Shrink and BAU (right).

Source: authors.

Economic assessment of future scenarios

35As discussed in Section 1, to shrink urban areas wisely, we need to consider (a) sustainability (b) risk adaptability, and (c) trade-offs between risks and other factors. While we have verified that the Wise Shrink scenario fulfills (a) and (b), it is less clear how to address (c). Thus, this section focuses on (c). More specifically, we first quantified the economic values of factors characterizing urban areas (e.g., accessibility and greens). Subsequently, (c), the trade-offs under each scenario, is quantitatively analyzed.

Hedonic analysis

36In this section, the hedonic approach is applied to evaluate the economic value of green spaces, accessibility and so on. The hedonic approach is a representative way to quantify the economic value of non-market goods (e.g., green spaces, disaster risks). This approach hypothesizes that the benefits from non-market goods are capitalized into the value of land. Under this hypothesis and certain additional conditions, the following regression model can be derived from the hedonic theory:

37where yi is the land price at i-th location, xi,p is the p-th explanatory variable, βp is a coefficient, wi is a spatially dependent disturbance, ui is a spatially independent disturbance. Consideration of wi is important in quantifying the economic impacts of each explanatory variable in the presence of spatial dependence. wi is estimated using the eigenvector spatial filtering technique (e.g. Griffith 2003).

38Table 2 describes the variables used in our hedonic analysis. The explanatory variables include land-use types, accessibility, and flood risk. Note that because of the collinearity between flood risk and earthquake risk, we considered only the former.

Table 2: Variables description: Hedonic analysis

Category

Variables

Description

Accessi

bility

Tokyo_dist

Logarithm of the distance from Tokyo Station to the nearest railway station [km]

Station_dist

Logarithm of the distance from the nearest railway station [km]

Disaster Risk

Flood_depth

Anticipated flood depth [m]

Urban

Land

Industry

Area of

Industrial land

in 1 km grids

[10 km2]

Road

Road

Public

Land for public facilities

Vacant

Vacant land

Green

Land

Paddy

Paddy fields

Agriculture

Other agricultural land

Forest

Forest

Wild

Wild land

Park/green

Park and recreation areas

Water

Land

River/lake

River and lake

Beach

Beach

1) All of these variables are collected from the National Land Numerical Information (NLNI) download service (http://nlftp.mlit.go.jp/​ksj-e/​index.html).

Table 3: Results of the Hedonic analysis

Category

Variables

Est.

t-value

Category

Variables

Est.

t-value

Constant

Constant

15.58

438.08

***

Green

Paddy

-16.03

-35.49

***

Access.

Tokyo_dist

-1.02

-100.54

***

Agriculture

-11.32

-22.71

***

Station_dist

-0.20

-36.92

***

Forest

-2.36

-5.60

***

Disaster Risk

Flood_depth

-0.06

-11.28

***

Wild

-17.36

-6.13

***

Urban

Industry

-4.93

-5.07

***

Park/Rec

6.73

2.49

**

Road

3.49

2.83

***

Water

River/lake

-4.58

-8.18

***

Public

1.38

1.52

Beach

101.99

1.82

*

Vacant

-3.67

-2.59

**

Log likelihood: -2210.21 (Non-spatial: -4121.81)

1) *, **, and *** denote statistical significance levels at 10%, 5%, and 1% error, respectively.

39Table 3 summarizes the results of our hedonic analysis. The log-likelihood of the spatial model in Eq.(10) is -2201, which is larger than that of the non-spatial model (i.e., Eq.2 without wi). The effectiveness of the spatial hedonic model is verified. The estimates for Station_dist and Tokyo_dist suggest the attractiveness of nearby railway station areas with good access to Tokyo Station. Areas with a dense road network are also found to be attractive, whereas areas nearby industrial land and vacant land are not. It is also revealed that flood risk significantly lowers land price. These results are intuitively consistent.

40An interesting finding is the negative economic values of natural land uses, including paddy fields, other agricultural land, forest, and wild land. These negative signs suggest that non-urban land use is not preferred compared to urban land use, probably because urban services are relatively more attractive than services from natural resources. Park and recreation areas are the only green areas with positive economic value. Parks/recreation areas should be increased for revegetation as they raise the attractiveness of their neighborhoods.

Comparison of scenarios: Revegetation

41In this section, the hedonic analysis is applied to compare Wise Shrink and Compact City scenarios from multiple viewpoints. Here, compact scenarios are subdivided into the following three scenarios: Compact_0, which converts reduced building land areas to areas with any land-use type; Compact_g1, which converts building land areas into green areas (i.e., paddy fields, agricultural areas, forest, wildland, or park/recreation areas); and, Compact_g2, which basically converts building land areas into any type of green areas, just like Compact_g1, but, only park/recreation areas are allowed in grids with population increase. Wise_g2 might be the best because it increases green areas while utilizing the positive economic value of parks and recreation areas.

42To estimate which land-use type is the most likely under each scenario, we used the following spatial compositional data model (e.g. Pawlowsky-Glahn and Buccianti 2011, Yamagata et al. 2016):

43where li,d is the composition of d-th land-use (d {1, 2, ..., D}) in zone i, zi,p is the p-th explanatory variable, bp is a coefficient, wi is a spatially dependent disturbance, which was estimated by the eigenvector spatial filtering technique (e.g. Griffith 2003), ui is a spatially independent disturbance. This model describes the influence of explanatory variables and spatial dependent components on the land-use composition, li,d.

44The explanatory variables include population (source: National census), Station_dist, Road_dens, elevation, distance to the nearest primary river (source: NLNI), and dummy variables indicating urbanization control area, lake, alluvial fan, natural levee, back marsh, delta, and sandbar, respectively (source: Japan Seismic Hazard Information Station, National Research Institute for Earth Science and Disaster Prevention). Eq.(11) was estimated using the data of land-use composition in 2006 in 1 km2 grids (source: NLNI; see Table 2). See Yamagata et al. (2016), for the parameter estimation result.

45Figures 6 and 7 show an increase in forest and park/recreation areas relative to BAU, which were estimated using Eq.(11). The forest areas under the Wise Shrink scenarios are much greater than those in the Compact City scenarios. It is verified that the Wise Shrink scenarios are preferable in terms of revegetation. Figure 8 shows that the increase in parks and recreation areas is small under the Compact_0 and Wise_0 scenarios. However, these green areas considerably increase in Compact_g2 and Wise_g2 as expected. Thus, it is verified that the degree of revegetation can change significantly depending on greening policies.

Figure 6: Estimated revegetation in 2050: Forest

Figure 6: Estimated revegetation in 2050: Forest

Source: authors.

Figure 7: Estimated revegetation in 2050: Park and recreation areas

Figure 7: Estimated revegetation in 2050: Park and recreation areas

Source: authors.

Comparison of scenarios: Economic value

46This section analyzes the trade-offs between risks and other factors using Eq.(12):

47where PiS and PiBAU are the populations in i-th unit estimated under the S-th scenario and BAU scenario, respectively, and is the corresponding regression coefficient estimated in section 5.3.1. xi,pS is the value of p-th explanatory variable (see Tables 2 and 3) in i-th unit under the S-th scenario. Only variables categorized into Urban land, Green land, and Water land in Table 2 are assumed to change depending on scenarios, based on the spatial compositional data model in Eq.(10). VS,p quantifies the total benefit from p-th explanatory variable that was added by using the scenario S instead of BAU.

48In each scenario, VS,ps were calculated for each explanatory variable, and aggregated into the following four categories (Table 2): Green land, Other land (Urban land + Water land), Accessibility, and Disaster risk adaptation. Figure 8 summarizes the economic values of these four factors under the Compact City and Wise Shrink scenarios. Values are positive if they are greater than those in the BAU scenario, and negative if they are smaller than those in the BAU scenario. The Compact City scenarios greatly increase accessibility because they concentrate people around nearby railway stations. Benefits from land-use variables are also increased because more people live in urban areas, which is highly valued, whereas fewer people live in non-urban areas, which is lowly valued. As a result, the total benefits from urban compaction are positive (top right in Figure 9). Still, compact scenarios tend to concentrate people in high risk areas whose economic values are low (bottom middle of Figure 8).

49By contrast, wise shrink significantly increases adaptability to flood risk compared with the BAU scenarios. The benefits from all of the other factors are also higher in the Wise Shrink scenario than those in the BAU. The effectiveness of the Wise Shrink scenarios is verified. There are significant differences among the three Wise Shrink scenarios. The total benefit received in Wise_g2 is 2.15 times greater than that in Wise_0 and 2.13 times greater than that in Wise_g1. The result suggests that urban compaction must be accompanied by an effective eco-urbanism as well as disaster risk adaptation.

Figure 8: Economic value of the Compact City and Wise Shrink scenarios

Figure 8: Economic value of the Compact City and Wise Shrink scenarios

Source: authors.

Discussion

50There are many approaches to eco-urbanism in the world. However, it is not easy to find out how we can create eco-urban districts in mega-cities. Our approach might serve as a tool to explore the possibilities by quantifying the implications of Wise Shrink scenarios as an alternative to compact city policies. Because of the trade-offs and synergies among multiple factors, compact city policies could have a variety of implications for different aspects. To assess Wise Shrink scenarios which consider the complexities, we have applied SULM to look at the trade-offs and synergies among factors determining sustainability. Our analysis has revealed some drawbacks of Compact City scenario, namely that it could increase flooding and other disaster risks and its impact on revegetation could be relatively minor. By contrast, it is shown that the Wise Shrink scenario could effectively reduce the disaster risks and, at the same time, increase green areas while enhancing the economic value of the metropolitan area from multiple viewpoints.

51Considering the recent increase of climate risks (e.g. Hirabayashi et al. 2013), the application of our modelling approach to support urban resilience-related decision making on a local scale could be an urgent task. However, our model still has some major limitations. 1) It needs to be integrated with a transport model. The transport sector is a very important factor for both economic and environmental constraints. We need to analyze whether urban compaction can actually reduce the use of cars (Mindali et al. 2004, Bouwman and Voogd 2005). 2) Livability needs to be modeled as that could be lowered by city compaction (Neuman 2005). Livability depends on a wide variety of micro environments, such as openness, sunlight, abundance of green areas, and landscape. In the future, a high-resolution 3-dimentional urban model that can quantify local properties related to livability and human well-being needs to be developed (Yamagata et al. 2015). 3) Advantages and disadvantages of SULM must be clarified by comparing it with other models. Viguié and Hallegatte (2012) created several future urban scenarios (e.g. scenarios with greenbelt, subsidy for public transportation, and so on) using the NEDUM-2D model, which is an economic model they had developed. Also, agent-based models have often been used for scenario development (e.g. Koomen et al. 2015). A comparative analysis of the methods of scenario development would be an interesting future topic. A weakness of SULM is that it ignores sociological phenomena, including collective actions and free riding. Barnes and Sheppard (1992) suggest that activities of rational actors in a geographic space do not necessarily yield rational consequences due to these phenomena. Also, the historical relationship among agents might have shaped urban form (Barnes 2012). An explicit consideration of these factors would be an important next step to maximize the wideness in urban shrinkage.

References

Anas A. 1982. Residential Location Markets and Urban Transportation. New York, Academic Press.

Anselin L. 1996. The Moran scatterplot as an ESDA tool to assess local instability in spatial association. Spatial analytical perspectives on GIS 111: 111-125.

Ayres RU, Kneese AV. 1969. Production, consumption, and externalities. American Economic Review 59(3): 282-297.

Baur AH, Thess M, Kleinschmit B, Creutzig F. 2013. Urban climate change mitigation in Europe: looking at and beyond the role of population density. Journal of Urban Planning and Development 140(1), March 2014 [online], http://ascelibrary.org/​doi/​pdf/​10.1061/​(ASCE)UP.1943-5444.0000165.

Beatley T. 2012. Green Urbanism: Learning from European Cities. Washington, Island Press.

Benenson I. 2004. Agent-based modeling: from individual residential choice to urban residential dynamics. Spatially Integrated Social Science: Examples in Best Practice 42(6-7): 67-95.

Barnes TJ. 2012. Roepke lecture in economic geography: Notes from the underground: why the history of economic geography matters: the case of central place theory. Economic Geography 88(1): 1-26.

Barnes TJ, Sheppard E. 1992. Is there a place for the rational actor? A geographical critique of the rational choice paradigm. Economic Geography 68(1): 1-21.

Bouwman ME, Voogd H. 2005. Mobility and the urban-rural continuum. Global Built Environment Review 4(3): 60-69.

Brady M, Irwin E. 2011. Accounting for spatial effects in economic models of land use: recent developments and challenges ahead. Environmental and Resource Economics 48(3): 487-509.

Burchell RW, Lowenstein G, Dolphin WR, Galley CC, Downs A, Seskin S, Still KG, Moore T. 2002. Costs of sprawl-2000. Report on Transit Cooperative Research Program, Transportation Research Board, 74, http://onlinepubs.trb.org/​onlinepubs/​tcrp/​tcrp_rpt_74-a.pdf.

Chakir R, Parent O. 2009. Determinants of land use changes: A spatial multinomial probit approach. Papers in Regional Science 88(2): 327-344.

Chen S, Chen B, Fath BD. 2014. Urban ecosystem modeling and global change: potential for rational urban management and emissions mitigation. Environmental Pollution 190: 139-149.

Chen H-T, Tsutsumi M, Yamasaki K, Iwakami K. 2013. An impact analysis of the Taiwan Taoyuan International Airport Access MRT System – Considering the interaction between land use and transportation behavior. Journal of the Eastern Asia Society for Transportation Studies 10: 315-334.

Cho SH, Poudyal NC, Roberts RK. 2008. Spatial analysis of the amenity value of green open space. Ecological Economics 66(2): 403-416.

Chrysoulakis N, Lopees M, Joe RS, Grimmond CSB, Jones BJ, Magliulo V, Klostemann JEM, Synneta A, Mitraka Z, Castro EA, Gonzaiez A, Vogt R, Vesala T, Spano D, Pigeon G, Freer-Smith P, Staszewski T, Hodges N, Mills G, Cartalis C. 2013. Sustainable urban metabolism as a link between biophysical sciences and urban planning: the BRIDGE project. Landscape and Urban Planning 112: 100-117.

Cressie N. 1993. Statistics for Spatial Data. New York, Wiley.

Dodman D. 2009. Urban density and climate change. Analytical Review of the Interaction between Urban Growth Trends and Environmental Changes Paper, 1.

Griffith DA. 2003. Spatial Autocorrelation and Spatial Filtering: Gaining Understanding through Theory and Scientific Visualization. Berlin, Springer.

Großmann K, Bontje M, Haase A, Mykhnenko V. 2013. Shrinking cities: Notes for the further research agenda. Cities 35: 221-225.

Haase A, Rink D, Grossmann K, Bernt M, Mykhnenko V. 2014. Conceptualizing urban shrinkage. Environment and Planning A 46(7): 1519-1534.

Haase A, Athanasopoulou A, Rink D. 2016. Urban shrinkage as an emerging concern for European policymaking. European Urban and Regional Studies 23(1): 103-107.

Hirabayashi Y, Mahendran R, Koirala S, Konoshima L, Yamazaki D, Watanabe S. Kim H, Kanae S. 2013. Global flood risk under climate change. Nature Climate Change 3(9): 816-821.

Howley P. 2009. Attitudes towards compact city living: Towards a greater understanding of residential behavior. Land Use Policy 26(3): 792-798.

Jim CY, Chen WY. 2007. Consumption preferences and environmental externalities: A hedonic analysis of the housing market in Guangzhou. Geoforum 38(2): 414-431.

Kaido K, Kwon J. 2008. Quality of life and spatial urban forms of mega city regions in Japan, in Jenks M, Kozak D, Takkanon P (eds.) World Cities and Urban Form: Fragmented, Polycentric, Sustainable? New York, Routledge: 161-175.

Kennedy C, Steinberger J, Gasson B, Hansen Y, Hillman T, Havránek M. Pataki D, Phdungslip A, Ramaswami A, Mendez GV. 2009. Greenhouse gas emissions from global cities. Environmental Science and Technology 43(19): 7297-7302.

Koomen E, Diogo V, Dekkers J, Rietveld P. 2015. A utility-based suitability framework for integrated local-scale land-use modelling. Computers, Environment and Urban Systems 50: 1-14.

LeSage JP, Pace RK. 2009. Introduction to Spatial Econometrics. Boca Raton, Chapman and Hall/CRC.

Li M, Wu J, Deng X. 2013. Identifying drivers of land use change in China: A spatial multinomial logit model analysis. Land Economics 89(4): 632-654.

Longden T. 2015. CO2 intensity and the importance of country level differences: an analysis of the relationship between per capita emissions and population density. FEEM Working Paper No. 047.2015, https://ssrn.com/​abstract=2609298.

Mehta N. 2007. Investigating consumers’ purchase incidence and brand choice decisions across multiple product categories: A theoretical and empirical analysis. Marketing Science 26(2): 196-217.

Mindali O, Raveh A, Salomon I. 2004. Urban density and energy consumption: a new look at old statistics. Transportation Research Part A: Policy and Practice 38(2): 143-162.

Ministry of Land, Infrastructure and Transport. 2003. Introduction of Urban Land Use Planning System in Japan, http://www.mlit.go.jp/​ common/000234477.pdf (Retrieved May 7, 2013).

Mishalani RG, Goel PK, Westra AM, Landgraf AJ. 2014. Modeling the relationships among urban passenger travel carbon dioxide emissions, transportation demand and supply, population density, and proxy policy variables. Transportation Research Part D: Transport and Environment 33: 146-154.

Morikawa M. 2011. Economies of density and productivity in service industries: An analysis of personal service industries based on establishment-level data. The Review of Economics and Statistics 93(1): 179-192.

Nakamichi K, Yamagata Y, Seya H. 2013. CO2 emissions evaluation considering introduction of EVs and PVs under land-use scenarios for climate change mitigation and adaptation – focusing on the change of emission factor after the Tohoku earthquake. Journal of the Eastern Asia Society for Transportation Studies 10: 1025-1044.

Neuman M. 2005. The compact city fallacy. Journal of Planning Education and Research 25(1): 11-26.

Omer H, Alon N. 1994. The continuity principle: A unified approach to disaster and trauma. American Journal of Community Psychology 22(2): 273-287.

Pachauri RK, Allen MR, Barros VR, Broome J, Cramer W, Christ R, Church JA, Dahe Q, Dasgupta P, Dubash NK, Edenhofer O, Elgizouli L, Field CB, Forster P, Friedlingstein P, Fuglestvedt J, Gomez-Echeverri L, Hallegatte S, Hegerl G, Howden M, Jiang K, Jimenez Cisneroz B, Kattsov V, Lee H, Mach KJ, Marotzke J, M, Mastrandrea MD, Meyer L, Minx J, Mulugetta Y, O’Brien K, Oppenheimer M, Pereira JJ, Pichs-Madruga R, Portner HO, Power SB, Preston B, Ravindranath NH, Reisinger A, Riachi K, Rusticucci M, Scholes R, Seyboth K, Sokona Y, Stavins R, Scholes R, Seyboth K, Sokona Y, Stavins R, Stocker TF, Tshakert P, van Vuuren D, Ypserle JP. 2014. Climate Change 2014: Synthesis Report. Contribution of Working Groups I, II and III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Geneva, Switzerland.

Pallagst K, Wiechmann T, Martinez-Fernandez C. 2013. Shrinking Cities: International Perspectives and Policy Implications. London, Routledge.

Pawlowsky-Glahn V, Buccianti A. 2011. Compositional Data Analysis: Theory and Applications. Chichester, Wiley.

Rosen S. 1974. Hedonic prices and implicit markets: product differentiation in pure competition. The journal of political economy 82 (1): 34-55.

Sato T, Hino S. 2005. A spatial CGE analysis of road pricing in the Tokyo metropolitan area. Journal of the Eastern Asia Society for Transportation Studies 6: 608-623.

Smith SK, Tayman J, Swanson DA. 2006. State and Local Population Projections: Methodology and Analysis. New York, Kluwer Academic Publishers.

Taniguchi M, Ikeda T. 2005. The compact city as a means of reducing reliance on the car: A model-based analysis for sustainable urban layout, in Wikkiams K (ed.) Spatial Planning, Urban Form and Sustainable Transport. Aldershot, Ashgate: 139-150.

Turok I, Mykhnenko V. 2007. The trajectories of European cities, 1960–2005. Cities: The International Journal of Urban Policy and Planning 24: 165-182.

Ueda T, Tsutsumi M, Muto S, Yamasaki K. 2013. Unified computable urban economic model. The Annals of Regional Science 50(1): 341-362.

Van Schrojenstein Lantman J, Verburg PH, Bregt A, Geertman S. 2011. Core principles and concepts in land-use modelling: a literature review, in Koomen E (ed.) Land-Use Modelling in Planning Practice. Dordrecht, Springer: 35-57.

Viguié V, Hallegatte S. 2012. Trade-offs and synergies in urban climate policies. Nature Climate Change 2(5): 334-337.

Voss S. 2006. A Risk Index for Megacities. München, Munich Re Group.

Wakamatsu K, Matsuoka M. 2006. Development of the 7.5-arc-second engineering geomorphologic classification database and its application to seismic microzoning. Bulletin of the Earthquake Research Institute, University of Tokyo, 81: 317-324.

Walsh R. 2007. Endogenous open space amenities in a locational equilibrium. Journal of Urban Ecoonmics 61(2): 319-344.

Wegener M. 2005. Overview of land use transport models, in Hensher DA, Button KJ, Haynes KE, Stopher PR (eds.) Handbook of Transport Geography and Spatial Systems. Amsterdam, Elsevier: 127-146.

Wegener M, Fürst F. 2004. Land-use transport interaction: state of the art. SSRN eLibrary, No. 1434678, http://dx.doi.org/​10.2139/​ssrn.1434678.

White EM, Morzillo AT, Alig RJ. 2009. Past and projected rural land conversion in the US at state, regional, and national levels. Landscape and Urban Planning 89(1): 37-48.

Williams K, Joynt JL, Hopkins D. 2010. Adapting to climate change in the compact city: the suburban challenge. Built Environment 36(1): 105-115.

Yamagata Y, Murakami D, Seya H. 2015. Latest high-resolution remote sensing and visibility analysis for smart environment design, in Thenkabail PS (ed.) Remote Sensing Handbook, Volume III: Water Resources, Disasters, and Urban. Monitoring, Modeling, and Mapping: Advances over Last 50 Years and a Vision for the Future. Boca Raton, Taylor and Francis: 599-621.

Yamagata Y, Seya H. 2013. Simulating a future smart city: An integrated land use-energy model. Applied Energy 112: 1466-1474.

Yamagata Y, Seya H, Nakamichi K. 2013. Creation of future urban environmental scenarios using a geographically explicit land-use model: a case study of Tokyo. Annals of GIS 19(3): 153-168.

Yamagata Y, Seya H, Murakami D. 2016. Urban economics model for land-use planning, in Yamagata Y, Maruyama H (eds.) Urban Resilience – A Transformative Approach. Basel, Springer International: 25-43.

Top of page

List of illustrations

URL http://articulo.revues.org/docannexe/image/3164/img-1.jpg
File image/jpeg, 12k
URL http://articulo.revues.org/docannexe/image/3164/img-2.jpg
File image/jpeg, 16k
URL http://articulo.revues.org/docannexe/image/3164/img-3.jpg
File image/jpeg, 16k
URL http://articulo.revues.org/docannexe/image/3164/img-4.jpg
File image/jpeg, 12k
URL http://articulo.revues.org/docannexe/image/3164/img-5.jpg
File image/jpeg, 16k
URL http://articulo.revues.org/docannexe/image/3164/img-6.jpg
File image/jpeg, 16k
URL http://articulo.revues.org/docannexe/image/3164/img-7.jpg
File image/jpeg, 12k
Title Figure 1: Image of SULM. Red: endogenous variables to be calibrated; black: exogenous variables that are assumed to be fixed.
Credits Source: authors.
URL http://articulo.revues.org/docannexe/image/3164/img-8.png
File image/png, 144k
Title Figure 2: Location of the Tokyo metropolitan area (left) and its population distribution (2005; right)
Credits Source: authors.
URL http://articulo.revues.org/docannexe/image/3164/img-9.png
File image/png, 345k
Title Figure 3: Estimated office agglomeration clusters (black) and subsidized areas (incentive zones).
Credits Source: authors.
URL http://articulo.revues.org/docannexe/image/3164/img-10.png
File image/png, 371k
Title Figure 4: Flood hazard (left) and earthquake hazard (right).
Credits Source: authors.
URL http://articulo.revues.org/docannexe/image/3164/img-11.png
File image/png, 467k
Title Figure 5: Estimated populations in 2050. This figure displays the population in BAU (left), differences of populations between Compact City and BAU (middle), and the same between Wise Shrink and BAU (right).
URL http://articulo.revues.org/docannexe/image/3164/img-12.png
File image/png, 713k
Credits Source: authors.
URL http://articulo.revues.org/docannexe/image/3164/img-13.png
File image/png, 739k
URL http://articulo.revues.org/docannexe/image/3164/img-14.jpg
File image/jpeg, 12k
URL http://articulo.revues.org/docannexe/image/3164/img-15.jpg
File image/jpeg, 16k
Title Figure 6: Estimated revegetation in 2050: Forest
Credits Source: authors.
URL http://articulo.revues.org/docannexe/image/3164/img-16.png
File image/png, 2.0M
Title Figure 7: Estimated revegetation in 2050: Park and recreation areas
Credits Source: authors.
URL http://articulo.revues.org/docannexe/image/3164/img-17.png
File image/png, 1.2M
URL http://articulo.revues.org/docannexe/image/3164/img-18.jpg
File image/jpeg, 12k
Title Figure 8: Economic value of the Compact City and Wise Shrink scenarios
Credits Source: authors.
URL http://articulo.revues.org/docannexe/image/3164/img-19.png
File image/png, 62k
Top of page

References

Electronic reference

Yoshiki Yamagata, Daisuke Murakami and Hajime Seya, « A Spatially-Explicit Scenario for Achieving “Wise Shrink” Toward Eco-Urbanism », Articulo - Journal of Urban Research [Online], 14 | 2016, Online since 13 December 2016, connection on 28 June 2017. URL : http://articulo.revues.org/3164 ; DOI : 10.4000/articulo.3164

Top of page

About the authors

Yoshiki Yamagata

Yoshiki Yamagata graduated from the University of Tokyo (PhD in System Science). Since 1991, he works at the National Institute for Environmental Studies (NIES, Tsukuba). Currently, he is Principal Researcher at the Center for Global Environmental Research (CGER)where he studies climate risk management. He is also affiliated with the IIASA in Vienna (Austria) and the Institute of Statistical Mathematics (ISM, Tokyo). His recent research topics include: Land use scenarios, resilient urban planning and International regime networks. He has lectured at the Universities of Tokyo, Tsukuba and Hokkaido. Internationally, he served as a lead author of the Intergovernmental Panel on Climate Change (IPCC), at the steering committee of “Global Carbon Project” and at the editorial board of “Applied Energy”, among other. Email: yamagata@nies.go.jp

Daisuke Murakami

Daisuke Murakami is a Research Associate at the Center of Global Environmental Research, National Institute for Environmental Studies. He received his PhD in engineering from the University of Tsukuba in 2014. His research interests include spatial statistics, regional science, downscale of socioeconomic scenarios, and geographical information science. Email: murakami.daisuke@nies.go.jp

Hajime Seya

Hajime Seya is an Associate Professor at the Departments of Civil Engineering, Graduate School of Engineering Faculty of Engineering, Kobe University and holds a PhD in engineering from the University of Tsukuba. His research interests include urban transportation planning, regional science, geographical information science, integrated land-use-transport modeling, and spatial statistics/econometrics. Email: hseya@people.kobe-u.ac.jp

Top of page

Copyright

Creative Commons 3.0 – by-nc-nd, except for those images whose rights are reserved.

Top of page
  • Logo DOAJ - Directory of Open Access Journals
  • Revues.org