Thursday, May 28, 2015

On-site vs off-site parking and a Swedish model for suburbs

It is often said that one of the great advantages of cars, notably over transit, is that it offers door-to-door travel. You don't need to walk 5 minutes to a bus stop or walk 5 minutes from the bus stop at the end of the trip to your destination, you leave your front door, get in your car and drive to the front door of your destination, easy-peasy.

Except, this isn't actually true. Cars don't appear when people leave their home, nor disappear once at destination, they have to be left somewhere, in a parking spot. So, in fact, cars are parking-to-parking, not door-to-door. The only true door-to-door transport mode is walking, even bikes are parking-to-parking (and in some cases, bike parking can be even harder to find than car parking). Taxis I guess could be true door-to-door transport too.

In general in North America, planners and developers focus on providing plenty of parking for each individual building, with parking located near to the front doors, what can be called "on-site parking" (as in, the parking is on the same lot as the building it serves). This is one result of the off-street parking requirements, that has developers building the off-street parking on the same lot as the building, because that is the lot he can build on. Driveways, parking lots, even on-street parking (if not over-used) all serve that purpose of bringing cars very close to doors.
Driveways and abundant on-street parking help bring parking spots as close as possible to housing
Low-rise multi-family housing surrounded by parking
Sometimes, the focus on convenient parking location demonstrates little short of contempt for non-drivers:
Typical North American strip mall, 1-2 minute walking distance from bu stop located right in front of it, imposes detours to residents who live behind the mall, all to have a huge parking lot built right next to the doors of the stores, to minimize walking distance for drivers
Typical North American mall, surrounded by acres of parking
The reason why I bring this up is that the convenience of parking can have a significant impact on choosing to walk or drive for short trips. When I say that travel time is the dominant factor in mode selection, this is the time door-to-door, not just in movement inside a vehicle, time taken to get to and from the vehicle may affect the decision of people to use it or not. So the closer the parking, the more favored cars will be as a form of transport. There may also be some other issues, like manual garage doors which take some time to open and close before leaving, which is why it's so frequent for the garage to be rarely used for parking by people who make frequent use of their cars.

All of this means that one way to discourage using cars is simply to make it less convenient to use them by placing parking locations farther from housing and destinations, which goes against the current regulatory regime that insists on on-site parking.

A theoretical demonstration

So how much impact does parking location convenience has?

Well, first, let's take a typical case of on-site parking with every house having a parking on the lot, in the form of a driveway. Houses are on side streets, so drivers go down to the arterial road then drive to the destination, either an office, factory or a store, which also has parking on-site. Schematically, this is what it looks like:
In yellow, the trip using cars, in blue, the trip using transit, dotted sections are done on foot, bus stops are the blue dots
So, in this case, the time wasted on car trips is minimized, with parking very near doors of homes and destinations. All in all, car trips include only 1 or 2 minutes of walking. Meanwhile, walking to the bus stop may take 3 to 5 minutes (an optimistic figure), waiting at the stop can also take 3 to 5 minutes (supposing the transit user knows the schedule and goes there just on time), and the bus stop at destination is still far from the door of the destination, so that's another 2 to 3 minutes. All in all, the transit trip includes around 10 minutes of walking and waiting, and that's if there is no transfer.

So, if we do a mode race to see how much time it takes to reach the destination on foot, on a bike, on a car or on a bus, we'd get something like this:

Mode race of the case with on-site parking, cars have a massive advantage

In this case, a 5-km trip (roughly 3 miles) would take nearly 25 minutes on transit, but only 10 minutes in a car. Cars start with a 10-minute advantage over transit and only increase that advantage over longer trips thanks to their higher speed.

Now let's take the exact opposite case, where both residential and commercial/office parking are off-site. In this case, car drivers have to walk 2-3 minutes to their car from their home and 2-3 minutes from the parking to their destination. Bus riders still have to walk the same distance to their stop, but as there is no parking in front of the destination, they will save quite some time walking to their destination from the bus stop at which they alight.

Schematic representation of the case with off-site residential and destination parking, car drivers have to walk much more in this case, even more than transit users

Trip distance vs time for the off-site parking example, here, cars lose most of their advantage for shorter trips
In this case, the same 5-km trip as before would take about 16 minutes in a car and 21 minutes on transit. The difference is much reduced and allows transit to be much more competitive. Bikes also shine here, as bikes are easily portable and offer something much closer to a door-to-door trip.

Finally, there is the on-street parking scenario, which is in the middle. If there is plenty of parking spots on the street, the results are very close to the first case with on-site parking. If there is a shortage of parking spots, then people may need to walk a bit, approximating the off-site parking example. In both cases, if parking is on-street, then destinations tend to be built next to the sidewalk, reducing walking distances for transit users.

The Swedish model

OK, so what got me thinking about this is exploring Swedish cities through Google Maps, trying to get a sense of their organization. Many neighborhoods of Swedish cities have a particular form. These cities have very strong planning, and the Swedes often live in multi-family housing, 40% of them do, many of them inside apartments built by a government program, the Million programme, under which the government motivated the construction of one million dwelling units in 10 years, a massive undertaking in a country of 8 million people.

Anyway, many multi-family neighborhoods in Sweden have a peculiar form in modern developments, take a look at this area from Vaxjo:
Residential area in Vaxjo, Sweden
Let me identify the parts of that area:
Distinct subsections of the neighborhood
What we have here is a strictly segregated approach, with a vast residential area without streets for vehicles, where buildings are linked by footpaths and bike paths. The road connecting this development to the rest of the city only runs on the periphery, and between the road and the residential area: parking lots. That way, cars are strictly kept out of residential areas where people walk and bike, total separation between motorized and non-motorized transport. It also means that cars are left near the road, which can be nearly 200 meters away on foot (700 feet)... right next to the bus stops. This also leaves plenty of trees and greenery in the residential area. Make no mistake, this isn't really urban, this is clearly a suburban area, but one where taking transit is not penalized versus using a car.

Technically, I guess the parking is still on-site. as it seems clear this whole area is one big development project, but for all practical purposes, the parking is effectively off-site due to its distance from residential areas.

Here is a particularly striking example (at least, from the sky), from Norkoping, a concentric neighborhood with parking on the outside ring, then two rings of apartment blocs, all surrounding a park:
Concentric neighborhood in Norkoping

Here is another example in Karlstad:
Suburban neighborhood in Karlstad, Sweden, with the neighborhood's commercial area identified, as well as the main footpath/bike path
This particular instance is interesting because there are very few roads for cars through this area, in fact, using Google Maps to give me walking and driving trips from one apartment bloc to the commercial area yielded me the following:
Walking trip from an apartment bloc to the neighborhood commercial sector, time of trip: 4 or 5 minutes
And here is the directions for the shortest car trip, note the time: 5 minutes, about the same as walking, so you would save no time by driving instead of walking
Another advantage of this design is that since there is only one main road, it concentrates transit users on one bus route that travels along that road. As a result, the bus line is frequent, data available online shows that the bus that does the loop actually has an headway of 10 minutes off-peak, in a city with a population of 87 000 people.

This total separation of motorized vehicles from pedestrians and cyclists may also be partly responsible for Sweden's excellent traffic safety results. Indeed, Sweden is in the Top 5 safest countries on Earth both by traffic deaths per 100 000 people and by 100 000 cars.

Again, make no mistake, this is a suburban area, and one that has strict use separation as commercial uses are set apart from residential areas, but a suburban area that seems to focus more on walking, biking and transit than on cars. Not to say that Swedish cities are only made of such neighborhoods, not at all. They have plenty of traditional Euro-blocs in older areas and single-family areas with driveways in front of every house:
Single-family houses in Stockholm
In some instances, they even use elevated parking structures on the periphery of residential areas:
Elevated parking structure in the center of the image, to absorb the parking demand for the high-density housing without overwhelming the residential area with cars, in Stockholm
The use of elevated structures is particularly interesting, because it is one way to really build a lot of parking on a very small lot, without breaking the bank. It can serve the purpose of hiding cars from view and preserving more of the lots for pedestrian-friendly amenities or design.

This type of neighborhood design is more typical of multi-family housing, but some single-family housing also use parking lots and garages put at the periphery, leaving the inside of the neighborhood free from cars:
Single-family area with parking lots put at the periphery, in Lund
Image of one of the parking lots, in fact, parking garages, that would typically be built adjacent to houses but that are instead grouped in the same location

Images of the houses with footpaths separating the houses
Of course, that design of self-contained neighborhoods with limited roads is great for transit, but it offers relatively poor connections between neighborhoods for pedestrians, often with paths that are not in view of residential areas. It also lacks mixing of uses as commercial areas tend to be put in a different location from residential. Still, Swedish cities do a lot of effort to build bike networks, often offering faster, more direct paths to cyclists than cars. As a result, despite the quantity of greenery of Swedish cities, bike and transit mode shares are very high (bikes more for smaller cities, transit for bigger cities). All data here from TEMS.
Mode share in Stocckholm, the biggest city of Sweden (metro population of 2 millions), extremely high public transit usage, relatively low active mode shares

Mode share in Gothenburg, Sweden's second largest city (about 500 000 people)

Mode share for Linkoping, a city of 140 000, there we see bikes being much more used, with still a significant amount of public transit use

Mode share for Lund, population of about 100 000. To compare, transit mode share in Toronto and Montréal is about 20%, so 16% in a 100 000-people city is quite high

Mode share for Uppsala, pop. 140 000
One thing interesting here is that no matter whether these are big or small cities, car mode share is always between 45 and 55%. This contradicts the typical belief in North America that only large cities can be walkable or can have decent transit, and that it's "normal" that smaller cities be car-dependent. Smaller cities can also favor sustainable developments, but instead of subways, this takes the form of quality bus services (or even tramways/LRT in some instances), bike paths and walkable proximities.

I'm not so keen on the strongly planned cities of Sweden, but their aesthetics may be more in sync with North Americans used to lots of greenery. Enclosure may be favored in urbanism, but not everyone likes it. Still, they are pretty successful at offering modal choice to people and at favoring all modes of transport, save for cars. The biggest flaw of this approach is inflexibility, when roads and infrastructure are built only for a certain type of development, making the area evolve over time is quite difficult.

The advantages of off-site parking

I think this model showcases how useful off-site parking lots can be, especially for a transition from suburbs to more sustainable, walkable cities. One of the big problems of density in suburban areas is that it results in cars taking over the public sphere entirely, being everywhere. The Swedish approach of putting cars in their own little zone apart from residential areas can help make it more palatable to live in an higher density area in transition from car-dependence to multi-modality, where cars will still be very present. Like a shameful disease, cars are best kept out of view of the community, allowing for more human-friendly design to dominate the public realm. The inconvenience of the parking location also encourages other modes of transport by reducing the time advantage of cars.

Off-site parking could also have an additional advantage of being possibly adaptive to demand rather than simply being regulatory. When parking is required on-site, since unused spaces on one lot cannot be used by residents of another, every lot is obliged to provide enough parking for its residents' highest possible parking demand. Off-site parking could allow for parking spots to be attributed to those who need them, so less of them would be needed, because residents not using their "allotment" could simply sell the spaces or stop renting them, allowing someone else to use that space. If parking spaces get rare, a new parking lot could be built. So off-site parking lots could theoretically do away with the need of minimum parking requirements.

For example, let's take the typical example of houses which regulation says needs to have 2 parking spots, built on-site. The regulation being what it is, the parking spots are privately owned by whoever owns the housing, and the price for the parking is bundled with the house's price. But since the regulation is there to make sure that enough parking is provided, in fact, it may well be that a significant amount of parking is not actually regularly occupied by a car.
Schematic example of houses with on-site parking spots, in light blue, ones not regularly used by vehicles, in dark, spaces that are regularly used
Now, in this situation, with 14 units and 28 parking spots, fully 10 parking spots are unused, but they cannot be used by others because they are privately owned. So even if you want to add other units, you cannot tap into the unused parking spots to avoid having to fulfill the parking requirements, because these parking spots are off-limit.

On the other hand, if you have another approach, where the 28 parking spots have been put in one parking lot, with homeowners having the choice to separately buy parking spots, then you have the following:
Schematic example of houses with off-site parking spots
In this case, because the 10 unused spots are still available for purchase, you could theoretically allow 5 new houses to be built without adding additional parking, because there still is 2 free parking spots for each of these newly built houses. And since people have to pay for their parking separately, it makes it more worthwhile to opt out of having a car. If you're flexible enough, you could also allow for parking lots to be sold to be developed if they're not used enough, which would make cars have to compete with housing for the land.

To be fair, this can also be done with on-street parking, but I think I've made it clear that I dislike that solution, as cars parked on a street claim it for cars and clutter up urban areas. As the city is typically responsible for building and maintaining streets, reliance on on-street parking also makes public authorities responsible for providing parking for residents, opening the door to all sorts of conflicts as the city is then forced to manage parking and held accountable if there are parking problems. I find it much better to leave parking in the hands of private residents and businesses, so that the costs can be borne by those who use it, not by the community at large.

So anyway, the more I think about it, the more I think this is a way of tackling parking that could be promising. This is also valid for commercial parking lots. This approach is scalable, it could work in suburbs, where low land prices and density could at first lead to abundant, affordable parking, but also allow for a gradual reduction of the amount of parking as land prices increase due to development and better public transit options slowly make car ownership less important. In other words, I think it could be a complementary approach to incremental developments.


  1. Off-site parking was the originally intended method of providing parking in cities like Boston and NY, and you can still some of the old garages that were built for the purpose.

    But then drivers discovered that they could squat on public land and throw temper tantrums until finally government officials let them overrun the streets. And so it goes.

  2. Wouldn't developments like these where lots of homes have no vehicular access be dangerous, as no direct access would be possible for ambulances or fire engines in an emergency?

    1. You can drive next to the buildings if needed, when moving or in a case of an emmergancy. In other cases it's not allowed

  3. Many U.S. garden apartments have layouts like the Swedish ones. I've lived in one myself. But they usually are very auto-oriented because they are located in the suburbs where distances to destinations are extremely far, transit service is poor, and bicycle facilities are poor or nonexistent.

    To serve many garden apartments, fire companies have to be able to deploy multiple 300-foot hoses. Ambulance companies have to be able to wheel stretchers hundreds of feet. (By the way, this rather undermines the common argument that fire trucks need wide streets so they can drive two abreast in front of every house. If blocks are short enough, the fire trucks can park around the corner on adjacent streets.)

    The narrow streets of Philadelphia have the off-site garage arrangement -- the historic townhouses have no onsite parking and the streets are too narrow for on-street parking. Another great traditional model is Park Slope in Brooklyn. The housing is generally townhouses with no onsite parking (the back yards are delightful). Even though there is on-street parking, it can't supply all of the demand because the residential density is so high. Residents use garage parking elsewhere in the neighborhood.

  4. For emergency access they can always make sure that there are paved 10' wide walkways in the interior that ambulances can drive on if need be. But if a firetruck shows up to actually put out a fire, they would need to run a hose from the nearest hydrant anyway.

    1. No...they carry water. And there is sunken hydrants all over the place in Sweden residential areas. You don't see them as they are small square manholes on the streets and walkways. To enter these Ares there are special keys to the fates for emergency services and service companies.