Tall Buildings as Urban Habitats: A Quantitative Approach for Measuring Positive Social Impacts of Tall Buildings' Lower Public Space

Abstract

After decades of high-speed development, designing tall buildings as critical components of urban habitat, rather than simply standing aloof from their environments, has become an important concern in many Asian cities. Nevertheless, the lack of quantitative understanding cannot support efficient architectural design or urban renewal that targets better place-making. This study attempts to fill the gap by providing a typological approach for measuring the social impact of tall buildings' ground conditions: that is, public space, podiums, and interfaces. The central business districts (CBD) of three Asian cities, Shanghai, Hong Kong, and Singapore, were selected as cases. Typical patterns and categories of lower-level public spaces among the three CBDs were abstracted via typological analyses and field study. The following evaluation is achieved through the analytic hierarchy process (AHP). This quantified approach helps to provide a visualization of high or low positive social impacts of tall buildings' lower-level public spaces among the three cases. This study also helps to suggest a design code for tall buildings aimed at a more human-oriented urban habitat.

1. Introduction

The rapid urbanization of Asia over the past few dec-ades has brought to bear high pressure to create lively urban habitats. In 1950, only 17% of the population in Asia lived in urban areas. The data show a rise to 40% in2005, and it is estimated that over 55% of the Asian pop-ulation will be urban by 2030 (World Urbanization Pros-pects, 2018). The large demands of economic and demog-raphic growth have led to an efficient urbanization parad-igm, focusing on a high-density built environment. In this context, new-build central business districts (CBDs), fully occupied by tall buildings, have become common. Alth-ough most of these CBDs achieved success from an econ-omic perspective, their quality from a human-oriented perspective was rarely considered. Positive social impact, including vibrancy, pedestrian quality, and perceptual feelings encouraging social activities on a human scale, is usually an afterthought (Safarik, 2016). Hence, numerous urban problems have been addressed, such as traffic con-gestion, the devastation of the urban fabric, and the dec-line of environmental quality and living conditions. Ashigh-density and high-rises constitute a present reality and future condition, redressal of those districts’ urban envir-onments becomes essential in further development (Cho et al., 2017).

In recent times, the rethinking of spatial quality and urban design in those CBDs has provided possibilities for solving these problems. However, we should be aware that those works should be undertaken not only on an urban scale, but also on a human scale. This requires fur-ther study of tall buildings’ lower-level public spaces, the most important and “urban” part of tall buildings, and their social impact. Public spaces, podiums, and the interfaces of tall buildings under five floors, or so-called “lower public spaces,” can encourage positive social activities by providing visual pleasure to pedestrians (Gehl, 2006; Me-hta, 2014).

Tall buildings have been criticized historically because of their isolation from the surrounding built environment.In this context, efforts have been made to integrate them into the urban context, particularly with regard to their ground conditions (von Klemperer, 2015; Safarik, 2016). Nevertheless, most of these attempts are proposed within an experienced-based and qualitative approach. This is inadequate for promoting more efficient design and maxi-mizing positive social impact for the public. To address this requirement appropriately, a typological analysis of typical patterns of ground conditions and a quantitative measurement of these patterns’ social impacts are needed.

This paper attempts to fill the gap by providing a typo-logical analysis of typical patterns of tall buildings’ ground conditions and a quantitative measurement of these pat-terns’ social impacts. Taking three CBDs in Shanghai, HongKong, and Singapore as cases, typical spatial patterns and categories in their lower public spaces have been identi-fied typologically via field studies. Related studies have been reviewed in the next section for selecting appro-priate analytical methods. After the summary of data and methods, the results section presents the analysis between those morphological categories of lower public spaces and their positive social impact. The paper concludes with suggestions for urban design paradigms that could assist in the transformation of tall buildings in urban habitats.The limitations of the study and future research direction are also discussed.

2. Related Studies

The relationship between the quality of public spaces and their social impact — that is, visual pleasure and accompanying social activities — has been discussed by many experts and academics. Gehl (1971) notes that the built environment tends to affect social relations by prom-oting or inhibiting users’ behaviors. In turn, people’s perf-ormances and behavioral preferences may be taken as an index for evaluating environmental attributes. In this con-text, the Revealed Preference (RP) survey has been gen-erated to assist in the evaluation of preferences through data from user surveys and investigations of on-site beh-avior (Bradley et al., 1991). A representative example of RTP survey is made by Nasar (1990), who interviews local residents about their preferences and the physical features of the surroundings. Objective indicators, such as body mass index (BMI) and air quality, are included as well as questionnaires by Frank et al. (2006). Mehta (2014) eval-uates activities in different scales of public spaces in sev-eral cities. In short, the RP survey is widely used in less-complicated and small-scale environments where inter-ference and variables can be easily controlled.

Increasingly, the Stated Preference (SP) survey is more suitable for complicated scenarios, where observed beha-vior is inadequate. By providing various combinations of different features, a quantitative model can be built to weigh the effects. For instance, Ulrich et al. (1991) meas-ure stress recovery in different participants after they were exposed to videotapes with nature and urban settings. Craig (2002) and Ewing and Handy (2009) suggest dif-ferent approaches for evaluating public spaces. Experts are asked to state their preferences on various kinds of scenes containing combined spatial features in photos or videos. Statistical tools are then used to assess the weight-ing of the different spatial features composing the scenes. Using this approach, the SP survey can enable researchers to reveal rare or formerly non-existent spatial patterns(Broach et al., 2012). This would therefore be more suit-able for complex built environments, such as CBD areas.

However, when faced with a high-density urban envir-onment, the SP method has difficulty with regard to theweightings among numerous categories. Hence, a reliable evaluation method is needed. The introduction of the Ana-lytic Hierarchy Process (AHP) fills this gap. The AHP has been widely employed to deal with models that com-bine both subjective and objective criteria, and has beenused in many fields such as real estate, forest manage-ment, and urban public transportation (Bender et al., 2000; Ananda et al., 2008; Nosal et al., 2014). Lo et al. (2003)uses AHP to identify the importance of design criteria inHong Kong’s open spaces. Similar work is done by Lee etal. (2008) using the AHP to assess urban renewal in HongKong. They suggest a list of major objectives and design criteria for providing sustainable renewal proposals. While notable, their work does not fully take into account the impact from surrounding buildings in place-making. In this study, the combination of SP and AHP enables the achievement of a systematic and objective evaluation of the social impact of tall buildings’ lower public spaces.

3. Research Methodology

3.1. Site Selection

This study chose three typical CBDs in Asia as study areas: Lujiazui in Shanghai, Central in Hong Kong, and Martina Bay in Singapore (Fig. 1). They were selected for the following reasons, in addition to their international competitiveness and significance in the world city network.

1) All of these CBDs started their high-density devel-opment during the latter half of the twentieth cen-tury. Their construction was completed within a short period of time.

2) These districts have been developed to meet the demands of international business. High land prices and similar land use make them the most concentra-ted districts of tall buildings in their cities.

3) Capital intensification has brought negative effects in those districts. Many spatial features and functions that are identified as being important for urban envir-onments, such as retail stores, open spaces, and parks, are more or less lacking in these districts (Wong,2004; Taylor et al., 1989)

Figure 1. The definition of lower public space in this study.

The definitions of tall buildings and their lower public spaces need to be elaborated for a clear discussion. Acc-ording to the definition given by CTBUH height criteria, the term “tall buildings” can be used to describe those buildings of 14 or more stories –– or those more than 50 meters in height (CTBUH, 2018). “Public spaces” as used in this paper are defined according to their accessibility to the general public. This includes not only their spatial typologies, objects, and artifacts therein, but also the interfaces that help to define the physical boundaries of the spaces are also connoted (Mehta, 2014). Taking 10meters as the maximum width of streets, the lower public spaces refer to those public spaces located below the sixth floor of tall buildings, including open spaces, podiums, and interfaces (Fig. 2). Public spaces over this range are reg-arded as inaccessible or difficult to discern from the street, according to Gehl’s (1971) study.

Figure 2. The three selected study areas; (a) Shanghai, (b) Hong Kong, (c) Singapore.

3.2. Analytical Framework

The perception and evaluation of built environments is a process that obtains awareness of sensory information(Ewing et al., 2009). Thus, complete perceptual behavior contains two main variables: physical features from the environment and appraisal by perceivers. This indicates that an in-depth study of lower public spaces’ social imp-acts should be conducted using typological analyses and subjective assessments.

Fig. 3 shows the analytical framework of this paper. First, we selected study areas around 1 km2 in three CBDs, each of which included 50-150 tall buildings. Typical pat-terns and categories of their lower public spaces were sorted after field studies and typological analysis. AHP was applied to calculate the weight of each category for rating. The technical details are enumerated in the follow-ing section, and the statistic and rating results are discussed in the conclusions.

Figure 3. Analytical Framework.

3.3. Survey Schemes and Data Analysis

Although discussions around tall buildings and their ground conditions need to be enriched, related studies about streets’ and open spaces’ qualities provide referen-ces for this study. Ewing (2009) suggests evaluating urban design qualities through physical features, including side-walk width, tree canopy, street furniture, transparency, etc .Elsheshtawy (1997) provides a morphological method for measuring the complexity of streetscapes according to their visual dimensions. Earlier studies of public spaces focus primarily on morphological patterns (Carmona et al.,2012). To establish a comprehensive model, we suggest that categories of public spaces and those of surrounding buildings should be considered simultaneously. Based on field studies and assistance from Google Street View, we classified typical patterns of tall buildings’ lower public spaces into five categories on three dimensions (Fig. 4).

Figure 4. Typical typological patterns of lower public spaces of tall buildings.

1) Spatial feature

a) Spatial typology: how public spaces are located and assembled in the lower part;

b) Spatial connection: the typology of connection between public spaces located at the same or different height);

2) Building feature

a) Building element: elements on façades that face the street. Given their distribution, this category is divided into features at street level and those at higher levels

b) GL (Ground Level) function: commercial office, shop, residential, parking, etc.

3) Street element: public or semi-public facilities esta-blished along the streets that are confirmed to have strong social effects.

As mentioned previously, AHP is introduced to eval-uate the social impact of five categories and their sub-categories. The AHP model is composed of three parts. The first, the goal level, explains problems that need to be decided. The goal level in this study is to identify the social impact of different patterns of tall buildings’ lower public spaces. The second, the criteria level, incudes five main categories: spatial typology, street element, building element (both street level and higher level), spatial con-nection, and GL function. To determine the priorities in these categories, this level is further broken down into different sub-criteria levels. Pair-wise comparisons are made at both levels.

According to Saaty (2008), a group decision can reduce bias against/toward particular group criteria. To obtain accurate results, this study invited 12 experts specialized in urban design and public building fields from TongjiUniversity to consult on the approach. All experts were required to arrange every category separately in order and consider their impact on perceptual feelings and encour-aging social activities (Table 1) Every item in this ques-tionnaire is shown by abstracted diagrams rather than by real images taken from the study area to avoid potential bias.

Table 1. Questionnaire for ranking typological categories and sub-categories

The final ranks are computed on the basis of the order derived from the questionnaires. Categories in these ranks received scores (S) rated from 1 to 9 by their position:

(수식 1)

where:

F_i: the item’s frequency in every position;

ω_i: weight of the item, determined by their positions.

The item in the first position has the largest weight;

a: number of items in each rank

Thereafter, the weights of these categories were deter-mined with the help of AHP and pair-wise comparisons. A=(a^ij)_n×n of every category. Values of a_ij were taken from rank results (S), which stand for intensities of importance between every two items. a_ij= 1 means that item i and j share the same level of importance. aij= 9 means i is far more important than j. YAAHP software was used to cal-culate the relative weights of criteria; thus, further mathe-matical details are not shown here.

The weights of categories and sub-categories were used to evaluate lower public spaces of every building located in the study areas:

(수식 2)

Where:

F_n: anticipates whether one sub-category is obtained in this building: 0 if none, 1 if true;

ω’_n: weight of sub-category, determined by AHP hierar-chy model;

ω_i: weight of category, determined by AHP hierarchy model.

4. Results and Analyses

4.1. Typical Patterns in three CBDs’ lower public spaces

The identification of typological patterns reveals their overall distribution and modes of combination in single building. Most tall buildings in the study area provide sim-ilar typologies of lower public spaces, usually “streets”(53.62% in Lujiazui, 55.26% in Central, and 31.65% in Maryna Bay) or “covered streets” (32.89% in Central and44.95% in Marina Bay). The “covered podium” exists in only one case in Marina Bay, and there are only four cases of a “sunken plaza” in Lujiazui’s lower stratum. In fact, the areas of urban open spaces per capita in all three CBDsare much lower than the international standard of 10 m2/inhabitant (Lo et al., 2003; Tan et al., 2017). The statistics of “spatial connection” also match this conclusion, whereover 70% of tall buildings provide only one public space in their lower sections. The “building elements” in the higher levels also show commonalities in these buildings. Especially with regards to newly-built tall buildings, 46.4%in Lujiazui, 30.7% in Central, and 41.04% in Marina Baychoose glass curtain walls as facades. Earlier office build-ings and most residential buildings use window walls instead (28.07% in Central and 29.48% in Marina Bay).

Compared with those in Central, public spaces in Lujia-zui and Marina Bay have more plants in the streets. Leav-ing aside implications from the climate, both Shanghai and Singapore introduced local policies to promote green-ing according to the Regulation of Shanghai Municipality on Afforestation and Greening in 1987, and the garden city Program in 1963. Although the Hong Kong govern-ment added clauses about open spaces and greening into the Hong Kong Planning Standards and Guidelines in2002, the main construction in Central had already been completed. Therefore, these policies were limited in their capacity to bring about more positive effects in this area.

The most typical patterns from the sites are represented in Fig. 6, which describes a common mode of categories in one building. Although all three patterns have similar ground conditions, their details are different. Owing to different setbacks guided by local construction laws, streets in Lujiazui are the widest. However, those setbacks are usually separated from the streets with roadside greenery, such as bushes or lawns. Only a few entrances are established for pedestrians at the end or in the middle of blocks. Meanwhile, in Central and Marina Bay, limited site areas and high land prices prompt buildings to fully use the sites they occupy. Streets usually contain seating or other resting places along the sites’ edges. Shops exist on almost every ground plane in Central, while less than half of the studied area in Marina Bay had shops facing the street. Shops can provide positive impacts for inhabit-ants’ social communications in these areas (Metha, 2014). The percentage of street-facing “shops” in Lujiazui is the lowest (20.41%) of the three areas, and inadequate to support daily living needs in the area.

The most typical patterns from the sites are represented in Fig. 6, which describes a common mode of categories in one building. Although all three patterns have similar ground conditions, their details are different. Owing to different setbacks guided by local construction laws, streets in Lujiazui are the widest. However, those setbacks are usually separated from the streets with roadside greenery, such as bushes or lawns. Only a few entrances are established for pedestrians at the end or in the middle of blocks. Meanwhile, in Central and Marina Bay, limited site areas and high land prices prompt buildings to fully use the sites they occupy. Streets usually contain seating or other resting places along the sites’ edges. Shops exist on almost every ground plane in Central, while less than half of the studied area in Marina Bay had shops facing the street. Shops can provide positive impacts for inhabit-ants’ social communications in these areas (Metha, 2014). The percentage of street-facing “shops” in Lujiazui is the lowest (20.41%) of the three areas, and inadequate to support daily living needs in the area.

Figure 5. Statistical results in three CBDs.

Figure 6. Typical patterns; (a) Shanghai, (b) Hong Kong, (c) Singapore.

4.2. Evaluating Results in Urban Scale

The relative weights represent the importance of cate-gories (Table 2). At the criterial level, “spatial typology” is identified as the most important of the qualities of tall buildings’ lower public spaces. Following that are “street elements” and “GL function.” “Building elements” in the higher levels have the lowest weighting. Most experts regard them as hard for pedestrians to discern from the street. At the sub-criterial level, “shop,” “covered street,” and “covered plaza” have the highest weight. The result suggests that shops on the ground floor with a particular spatial typology have the most positive impact on lower public spaces. Following that are “passing through,” “sun-ken plazas,” “podiums,” and “plants.” The “wall” on the higher levels has the lowest impact, despite its transpa-rency. “Handrail” and “no street element” also rank as less important in the place-making of lower public spaces.

Table 2. Weights of various categories via AHP analysis

The weights of categories were then used to evaluate impacts from particular buildings. Ranging from 0 to 100, a higher rate indicates a better social impact in lower public spaces or, in other words, a stronger capacity to encourage positive social behavior and use of the envir-onment. The results show that the average rates for threeCBDs are close to each other (Fig. 7(a)-(c)). The reasons should be considered from both morphological and static viewpoints. “Shops” contribute 83.8% of the “GL function” in Central, and thus make a great contribution to support-ing public activities and commercial atmosphere. Never-theless, pedestrians’ pleasure is decreased on narrow side-walks without any setbacks, and a lack of public street furniture and greenery significantly worsens their mental stress (Ulrich et al., 1991). This becomes more conspicu-ous in Central’s hinterland. Lujiazui and Marina Bay have more types of “street elements,” but the low proportion of“shops” to “GL function” (usually banks instead of retail enterprises, which are less attractive) hampers public acti-vities.

Figure 7. Rating results (a. Shanghai, b. Hong Kong, c. Singapore).

4.3. Evaluation Results of Buildings’ Scale

Targeted analysis of buildings explains how those vis-ual categories affect the qualities of lower public spaces.It is clear that there are some similarities among the low-rated buildings in the three CBDs. Most of those build-ings have the “street” in “spatial typology,” and provide narrow public spaces for pedestrians (Fig. 8(b)). The“building elements” of these buildings are usually “walls(low-transparency)” without shops to support street life. Itis worth noting that the abundance of “street elements,” although having higher weight in the hierarchy model, shows a weak influence on a particular building’s rates. The largest distinction appears in “GL function.” Although both No. 03b and No. 03c in Central perform well in other categories, the “parking lots” on the ground floor down pull their final rates (Fig. 8(a)).

Figure 8a. Lower public spaces in No. 03c in Central.

Figure 8b. Lower public spaces in No. 46 (Right) and No. 48 (Left) in Marina Bay.

However, providing various typologies for lower public spaces with better accessibility makes a notable contribu-tion to high-rises’ lower public spaces. In Shanghai, No.32a and No. 32b are considered to have obviously better spatial qualities, because both provide a “sunken plaza” with multiple levels, which help to attract pedestrians to stay or rest (Fig. 9(a)). Moreover, both are connected by a “podium” with a roof garden above, containing green-ery hanging over the rails that is easily noticed from street level. Examples in Marina Bay and Central prove the importance of a multi-level walking system to the success of lower public spaces in tall buildings. No. 94 in MarinaBay is linked with another building on the opposite side via a sky bridge. The whole walkway on the second floor is also fully exhibited along the street façade. In addition, an access point in the middle of No. 94 provides a short-cut for pedestrians to cross the block, which improves the quality of its public space as well (Fig. 9(b)). The gather-ing effect of a multi-level walking system appears strongly in Central. Nos. 41, 43, 103a, 103b, and 107 connected by a central elevated walkway extend their public spaces among each other, thereby overcoming the disadvantages brought by narrow streets and heavy traffic (Fig. 9(c)). These results prove the importance of a multi-level walk-way in a high-density metropolis like Hong Kong.

Figure 9a. Lower public spaces in No. 32a and No. 32b in Lujiazui.

Figure 9b. Lower public spaces in No. 94 in Marina Bay.

Figure 9c. Central elevated walkway among Nos. 43, 41, 107, 103a and 103b (from left to right) in Central.

5. Discussion

The analysis of around 300 tall buildings in three CBDsreveals some interesting understandings of the social imp-act of their lower public spaces. These spaces show a highdegree of similarity, partly because most of the tall build-ings counted share a similar “spatial typology” (usually“streets” or “covered streets,” see No. 46 and No. 48 in Marinna Bay). High-density development in those areas has crowded out room for public space, and concentratedthose joint spaces into large and separated plazas that are historically foreign to Asian cities (Miao, 2013). Outside of those large-scale open spaces, the combination of tall buildings and their lower spaces are limited to a few typologies from cases in Lujiazui (No. 32a and No. 32b)and Marina Bay (No. 94) where we can see that the diver-sity of lower public spaces has a large and positive effect on the urban environment.

Based on the AHP analyses and cases from Central (Nos.43, 41, 107, 103a and 103b), another possible strategy is to promote multi-level walking systems in CBD areas. The improvement from connecting public spaces among dif-ferent blocks shows an obvious effect in the results above. Vertical pedestrian networks with high accessibility could mitigate the limits of cramped public spaces in particular buildings. Besides, creating public spaces in the vertical dimension could also release contradiction between FARand civic activity demands.

Moreover, results from the morphological analysis andAHP could help to suggest appropriate design paradigms for the three CBDs. First, Lujiazui has been criticized for its large-scale and imperfect city functions (Liu et al., 2012). Given that reconstruction of the urban context and reform of public spaces has become a primary goal, it would be better to expand building volume into additional podiums using existing setback spaces, and provide room for retail enterprises and other urban facilities (Fig. 10(a)). A multi-level public space system could be built with a small-scaleplaza and roof garden instead of large setbacks, bringing public spaces on a human scale back into this area. Second, the tall buildings in Central ought to engage with their sur-roundings more positively through plazas or pocket gardens around the sites’ corners or edges (Fig. 10(b)). These pub-lic spaces could create open spaces and even passages for pedestrians. Given the local climate, the plaza could be partly covered. However, the case of a fully covered plaza in at the HSBC building in Hong Kong has led to unsatis-factory use of the environment due to its “hostile” design(Cuthbert et al., 1997). Hence, these spaces should be filled with plants and urban furniture for validating social acti-vities. In addition, Marina Bay is faced with a new periodof transformation because of the uneven and highly selec-tive nature of multinational corporations (Wong, 2004).Creating an urban public system that connects this district as a whole matches the increasing demands for flexibility.A paradigm focusing on accessibility between the urban environment and several lower public spaces would be helpful. We suggest creating more passable public spaces on the ground level, with higher levels of gardens and plat-forms linked via clear connections, such as escalators (Fig.10(c)). Covered streets should also be continuous between neighboring buildings given climate concerns. In the end, we evaluated the three design paradigms with the assess-ment method noted above. All of these design paradigms obtained ratings higher than 90 in our evaluation system.

Figure 10. Design paradigm (a. Shanghai, b. Hong Kong, c. Singapore)

6. Conclusion

Asian cities experienced a rapid transformation of their built environments in the late twentieth century. The high-density development model brought numerous challenges to the local urban environment, especially in newly-built CBD areas. Taking three CBDs in Shanghai, Hong Kong, and Singapore as cases, this study provides a quantitative approach to assessing the social impact of lower publicspaces in these areas. The integration of typological ana-lyses and AHP reveals how different patterns of tall build-ings and their surrounding environments affect perceptual feelings. In this study we find that “spatial typology” and the presence of a “shop” have the strongest impact on posi-tive social communications. The connection and accessi-bility of multiple public spaces also show great import-ance in place-making, which indicates a possible solution for high-density urban areas. The study also suggests app-ropriate design paradigms for the three CBDs based on the analysis. Both the assessment methods and design paradigms could be used for further urban designs that aim to create a human-oriented habitat in these areas.

However, this study contains several limitations as well. First, the current analysis takes into consideration only different impacts among several categories. The potential effects caused by high or low values in sub-categories are not fully taken into account. Further study is required to make in-depth assessments with detailed measurements of inner values among sub-categories. Second, the pool of experts invited for assessment needs to be expanded in the next step. Experts from other fields and users of lower public spaces in CBD areas ought to be included as well. Third, methods such as virtual reality (VR)/wearable dev-ices will be included in our future studies. These new tech-niques can introduce an objective approach to measuring people's potential preferences for different patterns. They will also help to provide an immersive environment to improve the validation of analyses, which may help to control possible errors caused by the existing methods.