JAPA has published my new article "Off the Grid… and Back Again? The Recent Evolution of American Street Network Planning and Design"
https://doi.org/10.1080/01944363.2020.1819382
Curious how US street network design has changed over time, especially in recent years? There's some good news! THREAD

First off, I'm also happy to share that it recently won the 2020 WRSA Stough-Johansson Springer Award for best paper!

Here's a blog post summary of the article: https://geoffboeing.com/2020/11/off-grid-back-again/

and here's an open-access pre-print: https://osf.io/t9um6/ 

What I found: urban planners abandoned the grid and embraced sprawl over the 20th century, but since 2000, across a variety of indicators, these trends have shifted back toward traditional street network design patterns that better support active travel and freedom of mode choice

US street network design changed drastically over the 20th century. In the early 1900s, dense interconnected imperfect grids were common. By WWII, urbanization shifted toward circuitous car-centric suburbs. By the 1990s, disconnected dendritic car-dependent sprawl dominated.

Given the semi-permanence of streets, we can examine snapshots of street patterns today to measure differences across places built in different eras.

To measure street network design over time, I estimate the vintage of each US census tract from a combo of census, tax assessor, and property transaction data. Then I group the tracts by decade and looked at their average values today across a basket of street network indicators.

The indicators' trends across decades tell the same story: street network griddedness, connectivity, density, and straightness declined steadily from pre-war highs through the 1990s. Most interestingly, however, all these variables' trends have reversed over the past two decades.

All else equal, each decade since 1940 is associated with lower griddedness than pre-1940. That is, post-war tracts may be larger, more spread-out, or hillier, but even when controlling for this, planners designed these street to be less grid-like than was typical before 1940.

But since 2000, griddedness and related indicators in newly built tracts have returned to levels not seen since the mid-20th century. This is important for making our cities less car-dependent...

All else equal, more grid-like places are associated with lower rates of car ownership today, which itself has a well-established relationship with vehicle miles traveled.

Where do street grids exist today? The Great Plains and Midwest, unsurprisingly, but there's also an archipelago of urban grids scattered across the country, as city centers tend to be more grid-like than suburbs or rural areas for several reasons.

For over a century, American spatial planning relied on the orthogonal grid for geometric ordering. But planners reorganized cities around the car after WWII, shifting away from dense interconnected street grids...

...in a bid to simultaneously attenuate the car's negative externalities (noise, pollution, streetscape blight, congestion, death) in residential communities while still empowering the populace to travel by car because it was fast and convenient.

Who cares? Well, street design matters for freedom of mode choice, whether we can walk or must drive, whether mass transit routes can be operated efficiently, how safe we are while biking, and how much pollution we emit.

My study finds evidence of promising recent trends toward more sustainable urban forms. But the initial layout of streets and land parcels determine urban spatial structure for centuries, locking in mobility needs and capabilities for generations to come. So what can we do today?

1) individual suburban retrofits can improve connectivity but are limited by the path dependence of infrastructure and land parcelization.

2) larger redevelopment projects offer strategic opportunities to incorporate (or restore) finer-grained, connected networks into their design for better human permeability.

3) greenfield development may offer the easiest opportunity to continue the aggregate trend back toward more-connected patterns, but such projects are often disconnected from the rest of the urban fabric and far from job centers.

4) street grids already exist in the cores and inner suburbs of most large US cities. Instead of building "connected streets" out on the urban fringe, planners can allow infill and densification to let more people live where infrastructure already supports freedom of mode choice.

A return to denser, interconnected grids? Looks like it. Millard-Ball and Barrington-Leigh recently identified a similar trend in network average node degree. @mnolangray also recently wrote in @CityLab about new street grids popping up in Texas https://twitter.com/mnolangray/status/1319000129995563008

I owe huge thanks to Adam Millard-Ball, Jana Cephas, @sarajcarr, @drschweitzer, @JakeWegmann and many others for their helpful comments on early versions of this paper.

For more on my methods and findings, check out the article itself: https://doi.org/10.1080/01944363.2020.1819382

Or if you don't have institutional access, here's a preprint: https://osf.io/preprints/socarxiv/t9um6/