Morphological measures

While accessibility and graph theory indices have been used to capture properties of the street network quantitatively, a number of scholars have also analyzed urban spatial configuration from a qualitative point of view. What we shall call morphological measures refers to a comparative analysis of urban form, where socially important features of built geometry are outlined through careful, qualitative analysis. 52

This sort of analysis of built configurations is best exemplified by the work of John Weeks, Team 10, Stanford Anderson, Michael Conzen, Anne Vernez Moudon, Allan Jacobs, Philip Panerai and others (Weeks 1963; Smithson and Team 10. 1966; Anderson 1978; Panerai 1980; Moudon 1986; Anderson 1993; Jacobs 1993; Conzen and Conzen 2004). Whereas graph theory analysis has focused on the spatial relationships between distributed elements of the built environment, morphological analysis has centered mainly on the attributes of individual elements themselves. We suggested at the outset that we shall not attempt to address the effects that layouts and aesthetics of individual structures might have on location choices. We therefore limit our discussion of morphological measures to only those attributes of built-form elements that contribute to their perception and access from the surrounding built environment.

Stanford Anderson’s analysis of the plan of Savannah offers an elegant example (Figure 17). Referring to the planar composition of wards laid out by General James Oglethorpe in 1735, Anderson argues that the spatial configuration of the plan has become a source of information over time: it guides location choices of land uses and engendering patterns of diversity and adjacency that are rare in most American grid plans (Anderson 1993). He outlines seven typologically different kinds of parcels (1-7) and six typologically different kinds of streets (A-F) generated by the plan (Figure 17).The seven parcel types are distinguished on the basis of their immediate exposure to and relationship with the surrounding streets and parcels35_. 1 4 5 6 2 3 7 A B C E F D

Figure 17 Plan of Savannah, GA, designed by General James Oglethorpe, showing six original wards around 1735.

Parcel type one is located on a corner of a continuous East-West street (street type A), and a continuous North-South street (street type B). Type A is the only street in the configuration with parcels opening onto the street on both sides36_{. Street type B is the only street that allows unobstructed North-}

South movement. The only parcels that open onto B are the back ends of the ‘trustee lots’ (parcels dedicated for public or communal uses in the Oglethorpe plan- type 7). Corner exposure to both of these streets gives parcel type one a topologically unique and distinctive setting, which could make type one and other analogous parcels in the grid more-or-less suited for certain activities. Anderson goes on to argue a

35 _{The same typologies repeat in all four quadrants of a ward through transformations of reflection and rotation.}

36 _{Though street type D is also double-loaded with parcels on both sides, it differs from A in its very narrow width, which makes}

D a service alley rather than a street.

similar case for each parcel type (one through seven) and street types (A through F), suggesting that the particular situation of each of these elements has created an unusually rich topological diversity in the plan of Savannah, which through time, has acted as a source of information for locating different activities in each ward. Because vehicular traffic is hindered in the center of each ward by the presence of a landscaped pedestrian square, then parcels facing towards the interior of each square appear to be better-suited for pedestrian-oriented activities than vehicle-oriented activities. The unique double sided quality of street type A, Anderson argues, along with its proximity to the original port on the river, might have historically incentivized Broughton Street to become the city’s primary commercial street.

“The decision to parcel the private development areas in a certain way (important features of which were the accessing of all such parcels on east-west streets; the establishment of no double-sided streets within a ward; the provision of as many private parcels at the edges as internal to the ward) established local use patterns that transformed the original arbitrary geometry into a structure filled with information. The small size of the wards and the importance of its periphery precluded the potential dominance of the central squares. Square, ward, and total structure, each had their character and strength in attracting appropriate uses. It was not as in most American grid cities composed of identical blocks, an arbitrary matter that one located a certain type of business or dwelling in one place or another. In Savannah, in growth, in decay, and in rebirth, energies knew where to flow first and from which to ebb last.” (Anderson 1993: 275).

Figure 18 Classification of parcel geometry based on levels of direct access to surrounding streets . A: Middle parcel B: Corner parcel C: Through parcel D: End parcel E: Island parcel.

Different degrees of variations in access and exposure can be observed in all urban block structures. A simple classification of parcel geometries in a rectangular grid distinguishes at least five different types of parcels, shown in Figure 18. The most common type is the middle parcel (A), located in the middle of a block, between three adjacent parcels, with only one side exposed to a street. The corner parcel (B) has access to two perpendicular streets at an intersection. A through parcel (C) also accesses two streets, but these streets are parallel, separated by a block. An end parcel (D) combines the advantages of corner parcels and through parcels by having access to two parallel streets and one perpendicular street. Finally, the island parcel (E) is surrounded by streets on all sides, having no immediately adjacent neighboring parcels on either side. Exposure to multiple streets allows tenants to benefit from several traffic streams on the door steps of an establishment. If all of the streets in Figure 18 accommodated an identical amount of foot-traffic, then the geometric advantages of the island parcel would allow (E) to benefit from exposure to four times the foot traffic of a middle parcel (A). More pronounced traffic exposure, combined with advantageous access radii that result from the parcel’s location, could well affect the location decision of a store owner. 54

The advantages of different parcel typologies are also apparent in their accessibility to surrounding parcels in a given walking radius, as we shall demonstrate in Chapter Three. We suspect that the widely encountered ‘corner store’ typology, for example, owes its popularity to an advantageous topological condition.

A substantively different, but methodologically similar analysis has been performed by Anne Vernez Moudon for a set of city blocks in San Francisco’s Mission district (Moudon 1986). Moudon’s detailed analysis of block sizes and subdivisions demonstrates how the geometric characteristics of these elements can also affect the types of buildings they end up accommodating. Moudon suggests that parcel dimensions can profoundly influence the typology and carrying capacity of buildings that come to occupy them. Geometric constraints on block and parcels can thus guide the process of growth and change that a city can accommodate through time. Mangin and Panerai have similarly shown how a careful dimensioning of parcels and city blocks can invite certain kinds of building types and lead to predictable patterns of densification and infill (Mangin and Panerai 1999). Allan B. Jacobs’ analysis of streets around the world has led the author to suggest that particular configurational qualities of streets, such as the density of cross streets, are related to the qualities of streets as hosts to a diverse set of pedestrian friendly activities. “Streets with one entry for every 300 feet (90 meters) are easy to find, and some of the best streets approach that figure, …., but there are more entries on the busiest streets” (Jacobs 1993: 302).

Jane Jacobs has argued that the availability of sidewalks, as well as their width, could act as important public spaces and facilitate access to businesses that cater to pedestrians (Jacobs 1961). Other analysts suggest that continuous and well-defined building fronts can render streets more pleasant for pedestrians, while wider setbacks from sidewalks could degrade their access (Figure 19). Double-loaded streets open twice as many doors to a road as single-loaded streets, animating the street with a wider visual diversity of buildings and their inhabitants and doubling the potential influx of foot-traffic in and out of buildings on a street segment. We aim to test whether some of these propositions affect retail location choices in our study area.

Probably, none of the morphological qualities mentioned here could alone define a good store location, but a cumulative collection of favorable spatial conditions might well inform the appropriateness of a place for selling goods of whatever type. The work of morphological analysis suggests that a location model that aims to capture configurational differences between various locations ought also to integrate a series of detailed place measures that go beyond graph theory type accessibility metrics.

Figure 19 Relative effects of a street’s building frontage on the suitability of a location for retailing. Left: single-sided streets tend to have half the doors on a street as double-sided streets (Middle). Right: Deep setbacks could be perceived less favorably than narrower setbacks by pedestrians who like animated and well-defined streets (Middle).