This is Part 1 of a series.
Just as we each have a carbon footprint, we each have a land footprint. A three person household in a home on 6,000 sf of land (including half the street in front of their home and half the alley) takes up 2,000 sf each person. As the average land footprint decreases, the square footage each person lives in might even remain the same as one story houses go to two or as two story apartments go to six, but the distance they have to walk to the store is lower. In addition to being more walkable, areas with smaller average land footprints can have better transit, pave over fewer forests and farmland, require less roads, sidewalks, sewer lines, electricity distribution, etc.
Using US Census data, we can see what kind of land footprint distribution an area has. This can tell us the most common living conditions in an area, and allows us to see how far apart people live in that area. This can also be helpful in comparing land footprints of different cities. Simply dividing the land area of a city by the number of people doesn’t work, since city boundaries are different and politically defined.
Below is what I’m naming a residential land footprint curve. The smallest geographical elements the US Census publishes data for is called a “block”. In cities these are often physical blocks bounded by streets, and outside of cities geographical boundaries are generally used. The Census Bureau has divided the US into 8.2 million blocks. All I’ve done is taken data from Seattle’s metropolitan area (the census definition is: King, Pierce, and Snohomish counties) and sorted the number of square foot per person in each block into “bins” of 100, then plotted the number of blocks in each bin. For example, there are 2 blocks in Metropolitan Seattle that have a density of between 0 and 100 square feet per person, 15 blocks between 100 and 200, 29 between 200 and 300, etc.
This curve allows us to easily identify the land footprint characteristics of a metropolitan area. For Seattle we can see that the most common living condition is around 3,800 square foot per person. Below that there’s a fairly linear distribution down to the densest blocks. This is mirrored above 3,800 until we reach around 7,000, where we start seeing a more gentle decrease. Few “blocks” (however they’re defined in rural areas) have 40,000 or more square foot per person inside our metropolitan area, but they do exist (I’ve truncated the chart to keep it readable).
In the chart below I’ve added pictures of actual blocks, to help visualize what these land footprints look like. The picture on the bottom left is a block on Capitol Hill, one of the 81 blocks with a density between 400 and 500 sf per person. At the top of the curve is a block in Magnolia, one of the 654 blocks with a density between 3,700 and 3,800 sf per person. On the right is an area in Renton, one of 47 blocks that have between 14,900 and 15,000 sf per person.
In my next post in this series I’ll compare Seattle’s land footprint curve to curves from other cities.
Time to empty those Craftsman bungalows in Ravenna out! Cabrini Green in the NW now!
I assume Roger Valdez will be directing via megaphone
It should be pointed out that the food a family requires takes up land. Related is the family farm, which has a large land footprint, but is used to provide for others. Density is important, but there are values to land beyond the efficiency of its habitation.
It should add that greater efficiency of habitation in the urban core preserves such farmland.
Exactly. Every farmer in America should be behind the concept of density and urban growth boundaries. Portland is a great example. You can drive 15 minutes from the center of downtown Portland and pick berries at a farm.
@Chris: Who was against Oregon’s urban growth boundary? Farmers living just outside of it. Why? Because without it they could make lots of money selling their land to developers, and with it they can’t.
Preserving American agricultural capacity generally can be a good argument in favor of urban growth boundaries, because even very fertile land is often valued higher for residential development than agriculture. But farmers looking out for their own financial interests would certainly prefer that their land take on a higher value due to development potential.
Aspiring and young farmers don’t want the highest value to be assessed on farmland near urban areas. It is only aging/retiring farmers that want to sell to pay for the high cost of healthcare etc after retirement. The aging farmer problem has therefore led to the huge need for young farmers in this country, espeically near urban areas, and yet we are all priced out of land ownership. Lease options are still not prevelent enough also.
Some of the best farmland in the country is, or was, in the Kent Valley, Santa Clara Valley, Orange County, Lemon Grove, probably Bellevue too, etc. Those are the farmers that “cashed out”. Although I’m not sure how well that term applies to the 60s-70s era when the entire society was pro-suburbia and farmers received only a fraction of what they would have gotten two decades later. But nevertheless, we’d have a lot more high-quality farmland near the cities if growth management had been in effect then.
What ever became of King County’s program to buy development rights to preserve farmland? I know there are large parts of the valley between Redmond and Woodinville where this was done. My recollection is that the program simple ran out of money. But I seem to recall that at one time it was common for upzones to be granted only when developers were willing to buy out offsetting development rights to preserve agricultural land. My guess is they found buying elections was cheaper.
…. meanwhile, Moscow is halfway through demolishing very dense ‘Khruschevkas’ soviet style housing units.
http://themoscownews.com/realestate/20111208/189270546.html
I wonder what the density of those are. Plop in a subway and voila, TOD in spades.
Now what’s not to like about that?
Perhaps the people that ended up living there, didn’t like it, didn’t take care of it and it had to be demolished. Perhaps TOD will eventually follow the same path.
Replace a few place names and that could be an article about the demolition of housing projects in any major American city (Chicago and St. Louis come to mind). From the history (broadly) of a now failing attempt to tackle the problem of homelessness and poor city living conditions, to the general problem of ideology overriding truth on the ground, to specific issues of shoddy construction and poor maintenance, to today’s low occupancy rates and depressing conditions, to mixed feelings and new housing uncertainty in people being forced to move out.
Hopefully something better can rise in their place. Privately developed housing fundamentally lacks some of the problems of public housing… though of course it has many different ones.
Soviet apartment blocks are like American urban renewal projects of the mid 20th century, only on steroids. Concrete ugliness, a short but boring walk to the metro, and an alienating atmosphere that literally overcrowded families and led to alchoholism. In Stalin’s era they didn’t even have kitchens because families were expected to dine in the company cafeteria for solidarity. Kruschev’s designs were less austere, but even then there was a single swinging faucet for both the bathroom sink and bathtub. A family of four commonly lived in a 2-room or even 1-room apartment (plus kitchen/breakfast nook). It’s less “tight” now but many families have more relatives or friends living with them than in an American apartment. Some of the huge buildings have multiple apartment complexes inside, walled off from each other, each with its own outside entrance.
I’m not sure how much of this overcrowding was inevitable given Russia’s severe housing shortage, but the huge buildings definitely reflect centralized “one-size-fits-all” planning and a “scientific” assessment of how many square meters and which amenities all families of each size needed.
TOD, ah, well, people are voluntarily living in TOD and paying premium prices for it. Eventually people will probably discover that we can build better TOD than we’re doing, by making TOD districts larger but each building smaller and more human-scaled, like in Vancouver’s second-ring districts (Kitsilano and Broadway, akin to Ballard). And in the first-ring highrise districts (West End and Yaletown, akin to Belltown and SLU), adding more amenities like full-sized supermarkets and drugstores and schools on ground floors.
Theoretically that could turn around and people may shun TOD someday, but it’s unlikely. The peak of car ownership and quasi-country houses was in the past. Practical concerns like getting around in an era of uncertain prosperity and energy supplies, and more elderly people who can’t drive, will become more prominent. That will inevitably lead to increasing walkability in both cities and suburbs. However, future TOD may be different than it is today, and suburbia may generate its own type of semi-TOD.
It’s not even the size of the buildings that’s the issue (see Battery Park City or the Barbican Estate). It’s the size of the apartments, the design of the apartments, and the design of the buildings. The Soviet apartment blocks were High Brutalism of the most severe sort.
@Nathaniel: It’s as much the specifics as the generalities. Isn’t half the point of brutalism supposed to be expressing the form and purpose of the interior of the building in its exterior? Such a philosophy shouldn’t preclude buildings with good interiors. It’s not just apartment size or amenities, either. Cities that built housing projects were replacing slums whose conditions were more crowded and, at least at first, in worse condition.
So there have to be some other explanations. First, that cities (here I mean incorporated central cities as opposed to their suburbs) were trying to fix exceptionally difficult problems with housing projects. It’s possible nothing they could have built would have worked. Take Chicago. Brutalist high-rise housing projects had their problems. But much sparser detached housing projects in other parts of the city did, too, and continue to. And crime and gang activity abound outside the projects, of course (the variety of results in neighborhoods without major housing projects is interesting). Second, that the projects didn’t address the whole problem of housing. They addressed many material needs and the need to modernize living conditions but not people’s need for dignity and respectability, and certainly not their need for continuity in their communities. Perhaps even more than the slums they replaced housing projects quickly gained a stigma. Designers came up with buildings with unsafe interior spaces and dead exterior spaces. Instead of identifying and strengthening what worked in existing communities they tore everything down and built from scratch — they uprooted people and let chaos form in the social vacuum.
Brutalist buildings on US college campuses, where conditions were much different, projects were of a different scale and scope, and buildings’ users had real power to influence building decisions and improve their surroundings, turned out much differently.
I live in a mixed-use building which, with its adjacent buildings, looks like one huge rectangle covering the whole block. When I said make each building smaller, I meant a tiny sliver of space between them, or a sharp differentiation in color and style, so that they look like several different tall-narrow buildings rather than one block-sized monolith.
Where are those two blocks with 100 square feet or less?
Downtown Seattle’s King County prison was #1. I don’t think we’re aiming at density quite that high ;-)
A suburban apartment complex was #2, though I don’t recall where. I have a feeling either there’s an error in the census data with that one, or a whole lot of people living in each unit.
Is it a very inexpensive area, rent-wise? When I first moved out of my parents’ house, I lived in a cheap apartment complex in a safe but unattractive suburban strip-mall district in Ypsilanti, MI, and nearly everyone in my building was either a college student or a Mexican immigrant. These were 700 square foot 1-bedroom units, and while I doubt there were any with 7 people in them, families of four were not that uncommon.
(My experience living there was actually pretty good, although I kind of wished I knew Spanish so I could ask to be part of their parking lot barbecues. The food always smelled ridiculously delicious.)
If I recall correctly, it looked like cheap apartments among other cheap appartments. Maybe three story. Let’s see, if each was 500sf but counts as 1,000sf (because of stairs, roadway, etc.), and assuming 3-story construction, that would be an average of a bit over three people per unit. That’s certainly possible. If it were only 1-story you’d need 10+(!) people per unit for this density. If it were 6-story you could have over twice the size of units and still have the same density.
It might be interesting to look at — some suburban apartment complexes are pretty dense buildings next to pretty big parking lots; if the block was cut in such a way that the parking wasn’t counted you might get a pretty high density figure. Such a cut would probably be an error of sorts… or if multiple families that didn’t live in a unit concurrently both used it as their mailing address for census purposes. Now I’m really curious what this was.
[Al] I bet that’s it. I don’t remember seeing parking. If I have time tonight, I’ll look up the location.
Seattle is one of the most expensive cities in the country. Only a few are higher: San Francisco, New York, DC, and a couple others. (For the same cost of living and better transit, go to Chicago or parts of L.A.) Everywhere else is cheaper, often much cheaper like Dallas or Atlanta. But you pay for it in non-walkability and skeletal transit.
Seattle was cheap in the 80s when its population was lowest. You could get a 2 BR apartment in the northern University District (55th) for $450, and sometimes even better bargains. But those are long gone.
According to Forbes (2009) Seattle doesn’t even make the top ten list; NY, LA, White Plains NY, San Fran, Honolulu, Miami, Chicago, Boston, Houston and D.C. (aka, the other Washington). Huffington Post reports on a 2012 study, “Cost Burdens of Moderate-Inocme Households by Metro Area.” Miami is #1 followed by Riverside, Tampa, LA, San Diego, Atlanta, Sacramento, Phoenix, Houston, Portland Dallas, Detroit, Chicago, Cincinnati and then Seattle just ahead of NY. That measure takes in to account income levels and cost of transportation as well as housing costs. NY is obviously at the top in housing but you can get around without owning a car. San Francisco comes in #20, Boston #21 and DC
Bernie,
Here is the full report:
http://www.nhc.org/media/files/LosingGround_10_2012.pdf
But it is a great read on how transportation costs affect people of moderate incomes in different metro areas.
One interesting thing: we likely wouldn’t think anything was amiss if an old parking-free apartment block on Cap Hill or in Belltown or the U District had a very high level of density, though some of its residents owned cars that they mostly parked on other blocks (whether by hunting for open spaces or by buying monthly passes at surface lots).
I was talking only about housing costs, not other expenses.
Re: The second densest block. There’s definately something strange with that data point. Here it is, and it’s listed as having 320 people in 26,600sf. It not only has parking, but a big grassy area.
So… 26,600 square feet is basically 160 feet on a side, which is much smaller than the complex. From the overhead view I’d guess there are about 140 units in the complex (spanning both of its blocks), maybe fewer. Do 320 people really live there? It sounds possible but maybe a little high.
This looks like a pretty new complex, so maybe the block data reflects what was there before, and everyone’s address is geocoded into a single small block. As to its real size, according to Google Maps it’s about 800 feet e-w and almost 500 n-s, so roughly 400,000 square feet. If we take a population of 320 at face value, that gives about 1,250 square feet per person. Still quite dense for Puget Sound.
In terms of location, the complex is a walk of between a half-mile and a mile to the commercial center of Enumclaw (as well as the peak-only bus to Auburn and the Sounder). Unfortunately there’s no similar density closer to those things and the zoning probably precludes businesses much closer than that.
What about plotting “number of people” on the vertical instead of “number of blocks”? Like, in each bin, plot the total number of people who live in all the blocks in that bin. Then you get an idea of the typical density conditions under which people live… there are a lot more people in the dense blocks than in the non-dense blocks, so this will raise up the right side of the curve some, and arguably give a better representation of where most people live.
Make any sense??
I like that. I’ll consider it for my next post.
This helps create the kind of density Jack mentions (quoted with permission by the author in a passworded forum) works pretty well, too, and doesn’t create the sterile 26 story towers:
”
If you read the stuff that I have written, I rarely use the word “density” and I’ve probably never directly advocated for it. There is a good reason for this: the planning profession has destroyed the concept of density.
Another fine example of this had me shouting at my computer screen this past weekend when the Minneapolis Star Tribune reported that “Minneapolis sees high-density future”. According to the report, high density is — as planners and their codes are so eager to bring about — means apartments and condo towers.
Minneapolis, the city of the single-family home on a tree-lined boulevard, sees its future in the apartment towers rising 20 stories above busy downtown streets.
Several major projects underway illustrate a trend toward tightly packed, urban living that is playing out in cities across the United States, giving Minneapolis planners hope of recapturing population the city lost starting in the 1950s. More apartments and condominiums got the green light in Minneapolis this year than any in recent history — about 2,800 in 22 new buildings so far.
“If we’re going to compete in the 21st century as a competitive global city, we have to attract people who want to live in cities. And cities are dense, urban environments,” said the city’s director of community planning and economic development, Jeremy Hanson Willis.
Driving the empty streets of Detroit this year, I was amazed at how this approach presents itself after the inevitable decline. There would be all of these (largely abandoned) houses and then, every now and then, a 20-story (largely abandoned) tower, with 1980’s era siding and a huge parking lot full of weeds. It was an eerie site.
When I was in graduate school, we had an “urban laboratory” course where we were tasked with leading the redevelopment of a 10-acre site. It was an abandoned dump of some type so, for all intents and purposes, bare ground. We were told that we must accommodate some athletic fields as well as something like 80 units of housing. Ever the deviant, I was shocked at how easy it was to convince my classmates (except my friend, Eduardo, who was cynically in on the joke) that a Le Corbusier tower in the park was the optimum design. Planners are obsessed with density.
Not to tough a painful nerve, but saying that density is the solution to the problems of our cities is like saying _______(choose your simplistic solution)_________ is the solution to gun crime. Is it more gun control? Better enforcement? Registration? Addressing mental illness? Video games? Violent television shows? We would like to think there is a simple solution, but the reality is that our society incubates violence in a very complex cauldron of cultural, social and environmental variables.
As it is with the health of our cities — it is a complex set of circumstances that defies an easy fix like density.
But if density is your cure, I’m with you at the end of the day, only not in the way that the code books or developers would have us proceed. If we simply allowed every single family home in Minneapolis to rent out the room above their garage, a flat in the backyard or a basement apartment, we could accommodate all of the population growth of the city without any new infrastructure, any bizarrely-placed towers completely out of context with the neighborhood and any need for substantial public investment.
Ask yourself why we don’t just do that and you’ll go a long ways towards understanding how we got into this mess and why getting out is going to be difficult and painful.”
From Chuck Marohn at Strong Towns–send him a note if you like. He would be up for the dialog. He is a professional engineer and planner, so by his own admission he can get geeky.
If you allow every person in Minneapolis to rent out their basements and garages, only some will do it, and a lot of those won’t find tenants. Draw a couple maps of your city. Where is there rental demand for small units? And where is there extra space for garage and basement apartments? If you have a serious lack of affordable housing in the suburbs there might be some overlap there. But even within the suburbs the most desirable locations for small apartments (near town centers with jobs and retail) aren’t where the extra space is (in cul-de-sac-ville).
Detroit didn’t fail because it built tall buildings. Detroit failed because it became economically dominated by three companies in the same industry (making the same products for the same consumers with the same labor), which failed.
Sorry, I mean the left side of the curve. Read the axis wrong, oops.
This is a great graph.
One thought: It seems like “number of blocks” is an arbitrary value, especially since low-density blocks are probably different in size from high-density blocks. Perhaps “number of people” or “amount of land” would give a better measurement for the y-axis.
seconded. this is a great graph, and I would also like to see a similar view with number of people.
very similar to what I’ve seen done eleswhere around tracking latency per type of web page.
With these sorts of graphs, you can meaningfully report on median, 10% percentile, and 90% percentile across geographies.
Very surprised the curve is so smooth. I would have expected to see some knees in the data around specific lot sizes.
We should expect the distribution to be log-normal. Some kind of reverse Gibrat’s Law evidently holds for blocks, too.
I’d like to see this data plugged into a mapping tool to give a block-by-block density map of the metro area. I think that would be very interesting for any transit nerd.
I’d love to see the US Census build a tool like that. You could see not only density, but everything they collect – income levels, race, number of cars, etc. I suppose the data’s out there if anyone with the skills and time wants to build it.
I live in a 850 ft sq apartment, but there are three floors, therefore I maximally used 850/3 = 240 sq ft of ground foot print and possibly less as there is a family above and group below.