Good street networks do not emerge; they are designed

The last time I blogged — a long, long time ago  — I wrote about the extreme disconnectedness of the South Lamar neighborhood. The maps below, for example, show two points near the geographic center of the neighborhood. Although they lie within a hundred yards of each other as the crow flies, they are not within a mile of each other by street. The South Lamar neighborhood, just two miles from downtown, is supposedly urban  — it certainly is redeveloping at an urban density — but it doesn’t have an urban street network.

This lack of connectivity imposes substantial costs on both neighborhood residents and nonresidents. It forces neighborhood residents to take roundabout routes through congested South Lamar Boulevard, wasting their time and intensifying the congestion endured by everyone. It makes the neighborhood impenetrable to nonresidents, depriving them of convenient, alternate routes. This is a serious flaw in a city with such a fragile transportation network, where a one-car accident on I-35 can paralyze the region’s roads for hours.

So it’s worth asking, “Why didn’t the South Lamar neighborhood develop a sensible street network?”

South Lamar developed much later than the surrounding neighborhoods, so one possibility is that the neighborhood’s streets were like that when the city got them. Historical aerial photographs, however, show that the herky-jerky pattern emerged after the area was annexed by the city.

Below is an aerial photograph showing the neighborhood in 1954, shortly after its annexation by the city in 1951.  

The neighborhood’s “main street(s)” — the Clawson-Lightsey-Del Curto concatenation — was there in 1954, but the neighborhood was mostly a blank canvas. There were a few streets running parallel or perpendicular to Clawson. There even were a couple of bona fide street blocks. These could have formed the backbone of a decent grid.  

Forward to 1973. The streets added between 1954 and 1973 are shown in blue:

Rather than grow a connected network, the neighborhood developed the tree-and-branch street pattern that plagues it today. Nothing emerged to knit together the neighborhood. There was still no alternative to Clawson-Lightsey-Del Curto for getting from north to south. There was still no street connecting the eastern and western halves.

Little has improved in the last 40 years. The photo below depicts the result of the city’s continuous stewardship since 1954. The neighborhood’s streets are a helter-skelter, disconnected mess of cul-de-sacs, streets that jump vacant lots for mysterious reasons, and giant tracts of land that lack any streets at all.

The problem here is pretty simple, and it lies with the process the city uses to create its street network.

There are essentially two methods a city can use to plan its streets. One method — which might seem the more obvious — is for the city to plan out its street network itself. This requires city planners to hunch over a map to plot out future streets. The original 1839 plan for Austin by Edwin Waller was such a plan. It laid out the grid pattern in downtown Austin that has endured virtually unchanged for nearly 175 years. (Technically, it was not a plan by “city planners” because it was made before there was even a city, but close enough).

Austin does very little planning of this sort. There are a few exceptions. Austin plans out arterial roads like William Cannon, working in conjunction with the County and CAMPO. CAMPO itself plans out highways and major thoroughfares. Austin also has a very few local plans for areas like transit oriented development districts or the East Riverside corridor. But, these exceptions aside, Austin does not plan its local streets.

Austin (and, to be fair, most American cities) uses a different method to generate its street network. Rather than plan the local streets itself, the city regulates street planning. That is, the city has adopted criteria that private property owners must satisfy when they develop their land. These criteria are surprisingly simple, and have changed little since Austin adopted its first subdivision ordinance in 1946:

And that’s pretty much it, aside from some miscellaneous rules that don’t matter here (e.g., subdivisions must generally have two entrances).The city otherwise leaves the choice of where to build a local street — or even whether to build a street all — to developers.

The assumption implicit in these rules is that an acceptably connected street network will emerge from the iteration of subdivision development. And these rules do guarantee some minimal connectivity when a developer starts with a large tract of undeveloped land. The block length rule restricts the distance between intersections, the alignment rule requires the streets to line up with and connect to adjacent streets, and the cul-de-sac rule prevents excessively long dead-end streets.

But these rules do a very poor job of ensuring connectivity in an area that was subdivided into smaller tracts before anyone thought of putting in streets. The tracts may be too large not to redevelop, but too small for the connectivity rules to generate an “organic” street grid. 

The close up of Del Curto below shows long, evenly spaced driveways and houses set far from the street. This pattern indirectly reveals the underlying subdivision of relatively narrow, very deep, semi-rural homestead lots. Significantly, although these lots are very deep (the lots abutting Del Curto on the west are approximately 600 feet deep), they do not reach the streets to the rear. They simply abut other very deep lots.

 .

These lots are now being redeveloped, one or two at a time. This is the point at which the city’s subdivision rules should begin to generate some connectivity. But the three simple planning rules do not do the job. Because the lots are being developed only one or two at a time, the maximum block length rule does not apply. The lots are long, but not long enough to violate the cul de sac rule. Most importantly, because the lots abut other lots to the rear, there is nothing for a new street to connect to.

This development on Thornton Road illustrates the dynamic specifically.

This shows a large undeveloped lot (in yellow) on Thornton Drive in 1997. Abutting that lot at the rear is a large lot (blue) fronting on Kinney Road. The yellow lots lies along a half-mile stretch of Thornton that lacks an intersection, so it would have made a lot of sense to cut a new street here to connect Thornton and Kinney Road.

The Thornton tract has since been redeveloped intensively with a mix of single-family homes and duplexes.

The developer did not provide a street connection to the Kinney Road lot because there was no street to connect to – the subdivision rules do not require a developer to assume that its neighbor will eventually provide a street. The mini-subdivision was also too small to trigger the maximum block length rule or cul de sac rule.

The Kinney lot has since been subdivided, but it remains lightly developed. If the owners choose to redevelop someday, the subdivision regulations won’t require them to build a street through to Thornton and there would be nowhere for the street to go in any event. The opportunity to connect Kinney to Thornton through that lot has been permanently foreclosed. 

This is the way the lack of connectivity within the neighborhood is propagated and reinforced. Each lot gets developed as intensively as possible, with no street connection, foreclosing future street connections in the process. The city has been running a controlled experiment for 60 years to prove that infill development alone cannot produce a decent street network. If we are ever to have a connected street network, the city will actually have to draw lines on a map. There is no street fairy who will slip in one night to fix the streets for us.

This does not imply that the infill subdivision process can never supply connectivity, but these events are so epochal that they are sure to draw objection. In 2015, the stars aligned just right for the South Lamar neighborhood: a developer proposed to redevelop a lot bridging Lightsey Road and Aldwyche Drive, indicated by the red line. The city’s connectivity standards not only applied, they required a street connection, the first ever to knit the two halves of the neighborhood together:

But the neighbors on the Aldwyche side objected, and Council ordered the new street to be gated.

In Austin, a stunted street network like this imposes increasingly high costs on everyone. This is one problem that won’t be solved by a laissez faire reliance on the land development market.