Tuesday, December 3, 2013

Complexifying the Street Grid

There is a post up over on blog-friend Strong Towns that is getting quite a bit of linkage play about differentiating street grids based on geometric size of the block pattern and the ratio of the quadrilateral "private" developable block and the size, scale, and amount of public infrastructure that inter-meshes with that private space.  Here is a graphic the Atlantic Cities used to demonstrate the difference between Portland  (left) and Salt Lake City's (right) grids

Beyond just the visual geometries, they went a step further to show the relationship between public space and private space.  The implication being that less public space is better.  After all, if we're going by some definitions of urban design, the entire practice (of design as well as governance) is about maximizing private utilization of land through the layout and design of public space and the conduits through which the economy happens (ie roads).  In general, that would be the least amount of public for the most amount of private.  However, that is not always the case and there is no performance measure based on this except that Manhattan/San Fran are low.  Those cities are in high demand and thus high value, therefore...good.

However, the connection is pretty loose.  Both are on the water where land supply is restricted.  Furthermore, can we definitively state that the inherent efficiency of the street grid had much to do with the various local economic sectors are in hyper-drive due to that singular efficiency?

This is an interesting development in the quantitative understanding of cities, but it is only one sub-set and there are far many more factors to also be considered that do play a role in the qualitative function and behavior of cities and those inhabiting them.

First and foremost is shape.  This measurement only works if the grid is highly regular, uniform, and repetitive.  While it was used in a number of American cities simply as a quick and easy way to survey and convey land from public to private hands for development, many more are highly irregular, change shape and size, and/or vary angles.  See: Dallas.

And Dallas (at least historic Dallas) is fairly regular in that it still has gridiron forms.  There are also the irregular medieval-ish patterns such as in Boston, the Baroque-ish patterns of DC, and the exceedingly highly irregular, oft-curvilinear patterns of cul-de-sacs and sprawl.

I don't want to get into a debate of the merits to the grid vs more irregular shapes.  Because in all likelihood there is no objective right answer nor definitive metric for concluding such a debate.  Either way that decision is generally best determined locally, if not hyperlocally at the actual site and determined by geography.

Which brings me to another way to measure cities that is not dependent on grids being highly regular.

Civil Engineering Professor from UConn, Norman GARRICK (ed. I mixed up Mailer and Garrick into Garrick.  I'm getting old.) has pioneered this work which is more translatable across the various geometric patterns of all of our cities, which is intersection density per area, typically square mile.


Typically, the grid has higher intersection density than the loopy doopy model.  So what, amirite?

Well, it turns out that Gailer's work actually has performance based results.  Once intersection density gets above 225 intersections per square mile, safety improves (there may not necessarily be less traffic accidents, but they get less severe and therefore, less deaths and severe injury or damage) and walkability increases, including higher mode share for all forms of transportation.

 I explored this locally in Dallas and showed that some of the most depressed areas of the city are that way because BIG infrastructure was introduced, thus reducing intersection density, thus reducing connectivity, which thus reduces opportunity, which then leads to disinvestment and decay.  That sounds complex, but it's actually quite simple.  It's the chain reaction (plus many of the biases that we overlay upon these issues) that makes it seem more complex than it is.

There is plenty to be said about the benefits of high intersection density vs low in how it alters behavior as well as real estate values and patterns.  This is something I'll be exploring in detail in a future post called the Funnel vs. the Filter.

However, this too is still a fairly simple arithmetic that doesn't say much about one-ways vs. two-ways, the scale of the streets in question, etc.  Intersection density does however play a role in another issue that is not taken into account...

All land within a block is not created equal.  In a post entitled the value and efficiency of small street and block structure (which, yes, implies high intersection density), I explore the hypothetical notion that 'storefrontage' is worth more than 'storehouse'.  In other words, the land at the perimeter of the block is worth more per square foot than the land interior to the block.  And this goes for just about any high intensity land use, whether that be office (the boss gets the window), retail (the 'lures' go in the shopfront), or residential (where you're often paying a premium for views while the unleaseable square footage associated with storage units, corridors, elevator shafts, etc are relegated internally).

I say hypothetical because the value increment between storefront and storehouse is in all likelihood also highly localized, a variable that changes not only from street to street but from city to suburb and city to city.  Who knows, there may be a relative constant, but oh my the research that would entail.

Ancient Rome vs. Modern Rome, "eroded" to be more useful, adaptable, and maximize value and performance.
Ancient Big Blocks: $46.53 / sq.ft.Modern Small Blocks: $31.5 / sq.ftVerdict: In other words, about a 50% premium for the small street and block structure.
In this scenario, where you have a tighter grid and more frontage you will likely have greater utilization of land and greater overall value.  According to the simple geometric model, Portland comes out looking poorly because its blocks are so small, but that also means a greater proportion of private land is higher value 'frontage' in relation to the sum total of private land because the depth of that gradient premium is a relative constant. It's the tightness of the grid which enforces a discipline upon private development.

Lastly (but certainly not last since there is no end to the potential ways to measure cities), there is the concept of centrality or the centeredness of something with regards to its surroundings.  This has some relationship to intersection density.  However, it could be argued that the intensity of the network is a response to centrality.  Greater demand to be near a 'centered' point of gravity creates demand for tighter grid and more streets.  Space Syntax is measuring this concept with their spatial integration models:

You may recognize the above map as London.  These spatial integration maps were used in the planning of the London Olympics, and were striking enough to director Danny Boyle, that he recreated them as part of the opening ceremony.  THAT was London.  Real London, evident of the beating heart of social and economic activity occurring.

What spatial integration maps do is measure the degree to which things are connected, allowing the definition of centers of activity to emerge.  In turn, understanding the direct relationship between movement, accessibility, and value.  The researchers that advanced this mathematical understanding of cities have shown the direct relationship between high degrees of connectivity (real connectivity) with land value, walkability, and crime reduction.  What they're showing is the direct impact infrastructure has upon real estate patterns, ie performance.

Utilizing this model on this side of the pond, I took the spatial integration software and applied it to resilience scientist Eric Klinenberg's work in Chicago to show why similar demographic profiles reacted so differently to catastrophic events, ie performance.  What happened is changes to local infrastructure networks severed physical networks which in turn severed social and economic bonds, the glue of resilience.  A neighborhood center was decentralized and in turn it ended up killing its function as a neighborhood center, which had the unfortunate butterfly effect of eventually killing many people during the Chicago heat wave because of the severing of connectivity and local centralization.

So what does the grid actually say about our cities?  Less than we think, but also probably more.  We just haven't put all the right metrics together yet.  But saying low public space to highly private space is a dangerous oversimplification that needs to take performance measures into account.