Tuesday, July 26, 2005


Reevaluating the Grid

By Fanis Grammenos and Douglas Pollard

Current thinking about city and neighborhood planning is rooted in the Hippodamian method of laying out cities such as Miletus, Piraeus, Priene, and Rhodes, but also incorporates and reflects the ideas of more recent thinkers such as Raymond Unwin, Clarence Perry, Clarence Stein, and Constantinos Doxiadis.

For example, the 1902 Garden City ideals of Ebenezer Howard have since been overlaid with the functional requirements of contemporary transport modes, resulting in the expanded, modern city with its suburbs that are accessible mostly by automobile. This is particularly the case in newer cities and in countries where a high proportion of car ownership was achieved early in the century, such as in the US and Canada. The newness of the cities offered opportunities for innovative and unencumbered design directions, and the prevalence of the automobile influenced the directions that were chosen. The resulting city form found many critics, most influential among them the US-based theorists that advocated Traditional Neighborhood Design (TND) as the solution to what was seen as a failing urban system.

Since the emergence of the TND school of thought (1992), the Hippodamian grid system of laying out streets has gradually returned as an essential feature of good planning. The grid’s interconnectedness is seen as the foundation for fostering true community, and even as the necessary condition for nurturing it. Although other features of TND (such as mixed use, village centers, moderate density as opposed to low density, and housing diversity) are arguably more critical in achieving its objective of recreating small-town ambience, the grid has become its signature, as no other design feature is as clearly discernible or a more enduring mark of its approach. Although the grid system undeniably has positive attributes, such as interconnectedness and a clear mental image (that helps to find destinations), there are instances where its applicability may be rightly questioned. Questioning street patterns is fully justifiable, not only because roads are the most expensive infrastructural investment, but also because they are virtually laid for perpetuity and, as such, may leave a legacy of unsatisfactory conditions.

Rationale for the grid
The grid has been a fixture of urban environments since the Greek planner Hippodamus introduced it over two millennia ago. During this long period, many new cities were founded using the gridiron as the principal method for laying out city blocks, lots, and houses (particularly in newly colonized continents in the Americas, India, and Australia). Its prevalence and endurance would seem to be sufficient justification for its continued application. After 1928, however, the grid fell into disuse and prominent planners, particularly in North America, publicly pronounced its obsolescence: “The flood of motors had already made the gridiron street pattern, which had formed the framework for urban real estate for over a century, as obsolete as a fortified wall…” (Clarence Stein, cited in M. Southworth and Eran Ben-Joseph, Streets and the Shaping of Towns and Cities, 1997). Much of the explosive urban expansion that took place from 1945 to 1990 used street types and patterns other than the pure gridiron. Dead-ends and crescents displaced continuous, straight streets, while city blocks gradually abandoned regularity and orthogonal geometry. These features were originally recommended by Raymond Unwin and Frederick Law Olmstead and were followed systematically by Clarence Stein and, in a modified form, Constantinos Doxiadis. In addition, the new “system” differentiated between residential and non-residential roads by geometry and design specifications. Some were designed purely for conveying cars (e.g., highways and parkways), with no urban land uses framing them.

In the mid-1980s, the grid experienced a renaissance, propelled mainly by the TND movement. It was argued by Andres Duany and Elizabeth Plater-Zyberk in 1992 that, “Streets ought to be laid out largely in straight segments, as they were until the 1940s. After all, the vast majority of our successful towns and cities, from Cambridge to Portland, were laid out this way” (see “The Second Coming of the American Small Town,” winter 1992, Wilson Quarterly). Although many American cities were indeed laid out in pure grids of varying dimensions (Figure 1), the link between grid geometry and success is tenuous. There are as many grid-based towns and cities in decline (e.g., Winnipeg, New Orleans, Cincinnati, and Detroit) as in good health and, conversely, there are cities that are laid out in meandering routes and dead-ends, particularly Arab cities but also certain British towns, that are prospering. If success is to be measured by economic vitality and the general quality of life of a city’s inhabitants, it can be easily shown that city location, transportation, and economic base contribute far more than street layout. Cities of all forms, as evidenced, for example, by Athens, can fluctuate from peaks of wealth, power, and cultural attainment to valleys of insignificance and deprivation. Evidently, other forces or factors besides street patterns explain a city’s well-being.

Figure 1. Sacramento, Portland, and Houston grids (same scale). Block sizes:
Sacramento, 400’ x 420’; Portland, 200’ x 200’; and Houston, 250’ x 250’.

Since 1992, the grid’s main attributes, interconnectedness and clarity of layout, have regained focus and prominence as site-planning features. The grid’s crossweave inherently produces convenience, route options, and traffic distribution, and it is now presumed by many to alleviate congestion and to reduce car travel. Legibility, which is defined as creating a clear sense of a path to a destination and a comforting rhythm of intersections along this path, is the outcome of orthogonal geometry and increases a pedestrian’s comfort. Some of these advantages were recognized, with reservations, by those who experienced the grid at the time of its early application. New reservations have emerged in regard to the direct application of the grid in the current, much different, context of motorized transport. These include issues of land-use efficiency, which focuses on land consumption as an additional criterion for evaluating street patterns (Figure 1), and concerns about the proportion of impermeable surfaces (those, in other words, that do not filter water), including streets, as well as pedestrian safety and traffic flow.

Reasons for applying or modifying the grid
Since its earliest application, the grid was both praised and criticized as a layout system. In response to its emergence in newly founded Greek cities, Aristotle wrote (Politics, 7:10.4):

The arrangement of private dwellings is considered to be more pleasant and more convenient for other purposes if it is regularly planned, both according to the newer and according to the Hippodamian manner; but for security in war [the arrangement is more useful if it is planned in] the opposite [manner], as it used to be in ancient times. For that [arrangement] is difficult for foreign troops to enter and find their way about when attacking.

Thus the grid was lauded for its convenience and delight but disapproved of on the grounds of security (Figure 2). The old “organic” layout was praised for being confusing to invaders but, Aristotle implied, not to residents, who had memorized its paths and passages from childhood. When it came to the ease of finding one’s way (legibility), the grid’s orthogonal geometry, it seems, served visitors more than it did residents of small towns. Contemporary urban tourists would agree.

Figure 2. Miletus (Hippodamian), Athens (the ancient manner), Vitruvius (ideal radial plan)

The most notable variation on the original idea of the orthogonal grid was Vitruvius’ (first century BCE) departure. He introduced a version that more accurately reflected the centric nature of the city and its transportation needs. His radial plan retained the legibility of the pure grid while improving on its convenience by reducing the walk-length to the center.

Few cities followed Vitruvius’ model; the majority of founded towns and cities as well as many expansions to existing ones followed the Hippodamian prototype or slight variations of it, such as an elongated block (e.g., S. Kleanthis and E. Schaubert’s 1833 proposal for the expansion of Athens). One marked innovation was introduced by James Oglethorpe in 1743. In the plan of Savannah, Georgia, which Oglethorpe founded, he deviated from the pure grid by introducing a “ward” or “cell” structure with a central open space that intercepts parts of the grid, thus rendering some streets discontinuous, although still aligned (Figure 3). This arrangement enhances the perception of a physical neighborhood by emphasizing a common space framed by buildings.

Figure 3. Savannah, Georgia: ward or cell structure
as conceived by Oglethorpe, 1734

These assessments and modifications show that, within a walking distance, the grid is simply one option for legibility, but not the best option for convenience. As for neighborhood identity, the grid may require a slight modification to achieve distinctiveness in each case when a city has many neighborhoods.

Another modification that was proposed recently abandons the orthogonal geometry of the grid to achieve “closure” by changing the direction of streets at chosen intervals. This proposal argues that “…humans do not like endless vistas,” an opinion apparently derived from shopping center (marketing) experience (Canada Mortgage and Housing Corporation, 1954, and “The Second Coming…”). Urban experience tells a different story: some of the most celebrated shopping streets are straight and, in some cases, extend for a few kilometers (Yonge and Bloor Streets in Toronto; Broadway in New York; and the boulevards of Paris). This proposed departure, although it produces a picturesque effect, might, if used liberally, endanger clarity of layout (legibility) and, potentially, convenience, both essential functional attributes of a street pattern. It is generally agreed that curved streets produce both the beneficial and detrimental effects of “closure,” which is why they have been judged undesirable by many.

New reasons for applying the grid
After its relatively recent resurgence, new justifications for adopting the grid have focused on aspects of mobility and accessibility. These suggest that the grid:

• is more convenient for pedestrians
• provides more route options
• distributes traffic more evenly
• relieves congestion on arterials
• reduces car travel.

A closer look at these reasons yields doubts about the grid’s effectiveness in delivering expected results and suggests that beneficial modifications to it may be required.

Pedestrian convenience
Vitruvius bequeathed us a plan that betters the grid in convenience: the radial, spiderweb plan (Figure 2). If a destination is a “center” within walking distance, the most convenient route is “as the crow flies,” a radial path. (Some railway towns followed Vitruvius’ idea in part.) An orthogonal grid, by comparison, will increase the distance to the center variously by up to almost one and a half times the straight distance (Figure 4). Convenience is further compromised in the orthogonal grid as the blocks become elongated; the greater the elongation, the higher the likelihood of a path to a destination becoming circuitous.

Figure 4. Railway town: four radial streets start from the railway station

More route options
At each cross intersection, the grid offers pedestrians three choices of direction. Route options are valuable when at least one of the choices can shorten the trip to a destination. All options offered by the grid, however, require the exact same effort in reaching a destination and, consequently, are of little value.

For drivers, route options become valuable when one route has fewer intersections or turns, or is less congested. Each of these conditions would save time (effort is no longer a factor when driving). In a grid, all routes have an identical number of intersections; consequently, no one route offers a distinct advantage in that respect. Absence of congestion depends on the number of cars choosing the same route, the number of routes available, and the number of delays encountered; all routes are equally vulnerable to congestion depending on their collective flow capacity. An alternative route is not an automatic guarantee of a shorter trip. It would seem that, regarding convenience, unless the grid is modified, its route options are valueless to pedestrians and of doubtful value to motorists.

Traffic distribution
Pedestrian traffic hardly ever requires distribution, as the number of pedestrians rarely exceeds the capacity of a street. The grid has enormous reserve capacity for pedestrian circulation.

For vehicular circulation, however, the grid’s capacity has often been gradually used up due to the constant rise in car use. Traffic distribution with more and wider roads has increasingly become one of the few means to accommodate the growing number of cars. As the grid becomes finer (e.g., Portland), it theoretically provides greater capacity for traffic since more roads are available per unit of space. It was found, however, that this progression toward a finer grid reaches a threshold of diminishing returns: the finer the grid, the more intersections and, consequently, the slower the traffic and the greater the chance for gridlock. Similarly, the greater capacity of a finer grid means more traffic is eventually accommodated; consequently, more residents are negatively affected by traffic. To strike a balance between traffic flow that improves mobility and traffic distribution that limits the negative effects on residents, adjustments to the grid may be necessary. Neighborhoods in many cities have implemented modifications toward this balance.

Figure 5. Diagonal diverters and bollards (bollards inside red ellipse)

Congestion relief on arterials
A regular, uniform grid does not differentiate among its streets, although wider streets appear occasionally in plans (usually for ceremonial reasons). The designation, design, and treatment of certain streets in existing grids as “arterials” reflect the new need for motorized mobility: cars lose their commuting attraction when traveling far below their normal speeds. Another need, unimagined by the inventors of the grid, is for mass transit. To move both car traffic and mass transit together on designated “arterials,” access to them is either prevented or strictly controlled, while road width is increased where possible. These modifications initially enable, but also invite, more traffic, which eventually reaches congestion levels once again. Other streets can relieve the pressure, and the grid’s system of parallel streets offers convenient alternative routes to the same destination. This relief, however, comes with a negative side-effect: residents are burdened with traffic intrusion that lowers their expected quality of life (see James Daisa, Tom Closter, and Richard Ledbetter, Does Increased Street Connectivity Improve the Operation of Regional Streets? Case Studies from the Portland Metro Regional Street Design Study, 1997). To balance motorized mobility and quality of life, modifications to the grid are required.

Reduction of car travel
Car travel is currently considered a cardinal civic sin: it pollutes the air, contributes to the greenhouse effect, increases risk of accidents to pedestrians (particularly children) and cyclists, creates noise, and may affect the health of drivers by depriving them of potential exercise. Reducing car use per se and total travel cannot be overstated as site-planning goals. Of the two key factors that affect the choice of travel mode and distance, one, connectivity, is an inherent attribute of the grid. The other, the presence of transit routes and a variety of land uses nearby, is unrelated to street geometry: it depends on planning policy, zoning (including density), and development economics. In the absence of nearby useful destinations, the car becomes the most convenient means of reaching medium- and long-distance ones. Car travel cannot be reduced by the use of grid geometry alone: land-use distribution is an essential prerequisite.

Evolutionary and creative adaptations of the grid
The grid, a transportation system that originally balanced convenience for foot and horse-and-cart travel in ancient and preindustrial societies, showed signs of strain soon after motorized traffic appeared. Accidents and fatalities rose steadily, along with the worsening quality of urban space. Responses to these outcomes, some managerial and some physical, came in rapid succession and continue to emerge. On the managerial side, painted lane lines, horns and car signals, stop and other traffic signs, traffic police, automated signals, and one-way streets, became common in an ever-increasing effort to maintain order. On the physical side, selective road widening, strategic street closures, roundabouts (recently reinvented), elaborate at-grade or grade-separated intersections, and, recognizing the occasionally irreconcilable conflict between motorized and foot traffic, the creation of distinct circulation networks and districts for pedestrians only—foot and bicycle paths, inner-city pedestrian zones, plus-fifteen-level (elevated street) paths, and underground “street” networks connecting numerous downtown destinations—were all responses that alleviated but did not eliminate vehicle-generated conflicts.

Many of these conflicts that emerged in existing grid-based areas of cities were resolved in the design of new suburbs when their street patterns were adapted to the car: quiet and safety at the local street level, with speed and convenience at the district level. The grid geometry was abandoned as a method for laying out streets in favor of discontinuous streets such as cul-de-sacs, crescents for residential streets, and fully engineered continuous arterials for regional mobility and accessibility. These adaptations, however, rendered street patterns inhospitable and dysfunctional for pedestrians.

Creative mutations, cross-fertilization, and hybrids
The widespread use of new suburban street types and the appeal of the environments they created stimulated new approaches to adapting the grid in existing residential districts. In a number of cases, loops were created within the grid by constructing permanent diagonal islands across an intersection (Figure 5). Similarly, cul-de-sacs evolved out of straight streets after islands with bollards were erected at mid-block. These adaptations show that it is possible to have full pedestrian connectivity while maintaining vehicular imperviousness. They also show that the valuable legibility of the grid’s orthogonal geometry need not be sacrificed in achieving the discontinuity of motorized traffic. From this cross-fertilization of ideas, hybrid types of streets and street patterns are being applied and proposed: the open-end cul-de-sac, the connected cul-de-sac pattern, often complemented by a separate foot- and bike-path system (see Mike and Susan Corbett, Village Homes, 2004), and the crescent or cul-de-sac streets connected by paths to and through open spaces (see M. Southworth and Eran Ben-Joseph, Reconsidering the cul-de-sac, 2004). One proposal, the fused grid, uses the traditional grid as the basis for incorporating these adaptations and mutations into repeatable stencils (see Canada Mortgage and Housing Corporation, Residential Street Pattern Design, 2002). It structures a district on an underlying orthogonal grid using 40-acre modules as building blocks (Figure 6). These blocks are impermeable by vehicular traffic but fully accessible on foot and bicycle. Four modules (quadrants) in a square arrangement are framed by two parallel transit routes containing mixed uses (Figure 7). This configuration creates a continuous grid at the district and regional scales for motorized travel and a discontinuous grid at the neighborhood scale for foot travel. At the same time, it creates opportunities for green infrastructure.

Figure 6. The fused grid: four patterns using loops and cul-de-sacs
as the main street types within a neighborhood
(each block is 400m x 400m and a five-minute walking distance)

These new hybrids resolve the apparent incompatibility between the inherited grid system and the car-adapted suburban patterns of the last 50 years, and they show how the connectivity, structure, and legibility of the grid can be combined with the tranquility, safety, and economy of the more recent street types.

Figure 7. The fused grid structure: four quadrants framed by parallel roads,
between which mixed uses are located

Evidence suggests that this may be a workable model for new cities, for expanding existing ones, and for a careful, staged remodeling of built-up areas. Milton Keynes, a “new town” (1960s) in England, reports few traffic problems, a fact that is attributed to the one-kilometer grid of regional roads, a central feature of the fused grid. Stockholm transformed a central area of the city in 1970 by creating a quadrant that consisted of perimeter roads, loops, cul-de-sacs, and a diagonal open space weaving through the center of the block, a solution that is unmistakably analogous to the fused grid quadrant. (Figure 8)

Figure 8. The transformation of a Stockholm central district

Rapidly expanding cities in Greece as well as new satellite towns will inevitably experience the effect of rising car ownership on circulation and neighborhood quality. They could benefit from this contemporary interpretation of the Hippodamian grid, which is designed to accommodate pedestrian and motorized traffic equally well.

Fanis Grammenos and Douglas Pollard are senior researchers at Canada Mortgage and Housing Corporation, Canada’s national housing agency.
Page 1 of 1 pages