 |
| BART survey data (2008). Top: rider home locations; bottom: rider employment locations. Courtesy of BART. |
BART has released its 2008 Station Profile Study, updating its last study from 1998. The data, which is collected from rider surveys, is BART’s version of the census. It reveals the demographic profile of BART riders, and it provides valuable information on how riders use the BART system: where they are coming from, where they are going, how they travel from their home a nearby station, and how they travel to their destination after riding BART. The data, which is available both system-wide and for each individual station, confirms what we know anecdotally about the role of urban vs. suburban stations: 81% of riders at 16th/Mission walked to BART, while merely 3% walked to Orinda; 72% of riders drive to North Concord/Martinez, but a miniscule 1% drive to Powell. I plan to do some number-crunching on the data in the future; but for now, I wanted to share some interesting results and initial impressions. In addition to clarifying how BART riders currently make use of the system, the survey data reveals how the Bay Area could better take advantage of this critical regional asset than we do today. The lesson we learn from the data is the lesson that we already knew: we need to do a better job of linking transit and land use, particularly along BART’s heavy rail metro lines. This is something that we are always talking about, and the BART surveys do suggest that the region is moving in the right direction in terms of promoting transit-oriented development. Bicycle trips from home to station bumped up from 3% to 4%, while transit trips declined from 23% to 15%. Nearly half (49%) of riders access stations by car (34% solo, 10% dropoff, 5% carpool), the same as in 1998. However, more people are now walking to BART stations from their home than they were a decade ago: 31% in 2008, compared to 26% in 1998. More people are also walking from BART to work or other destination: 74% in 2008, compared to 67% in 1998. Furthermore, at 6 major CBD stations (12th St, 19th St, Lake Merritt, Berkeley, Montgomery, Powell) and 5 other mostly urban stations (Ashby, North Berkeley, El Cerrito Plaza, Colma, and Balboa Park), home origin points increased by 10% or more, while car and transit origins decreased. More home-based pedestrian trips at downtown stations reflect a trend toward urban/downtown infill housing, epitomized by Jerry Brown’s 10K housing initiative in Downtown Oakland and San Francisco’s Rincon Hill plan.
Better Stations, Better Neighborhoods
But the survey data also shows that we can still do much better on both ends of the stick — both by augmenting service where it is needed, and by increasing density on valuable parcels adjacent to stations, in places where land use is currently not as intense as it should be. Contrast home origin locations for riders using 16th Street Mission with home origins for riders using North Concord/Martinez Station. The map on the left for 16th Street shows a dense residential population walking, biking, or taking transit to the station (median of 0.46-mile trip); the map on the right for North Concord/Martinez shows a dispersed ridership that mostly drives to the station (median of 6.04-mile trip):
Home origin locations, by mode. Left: 16th/Mission (circle = 1/2 mile radius).
Right: North Concord/Martinez (outer circle = 1 mile radius). Purple = walking,
green = bicycle; blue = transit, orange = drop off; red = car. Courtesy of BART.
The data also conveys another idea: while commuters are willing to travel a longer distance from their home to a station, they prefer their place of work to be more immediately located to a station. This principle should influence how we design transit-oriented development, and it should be taken into account in all station area specific plans that cities might prepare. To the extent that a “transit village” houses both commercial space and significant residential stock, the parcels immediately adjacent to the station are best reserved for significant office space with ground floor retail (perhaps destinational in nature, or of regional appeal), while outer parcels are well-suited for residential or mixed use, with ground floor commercial space that serves the neighborhood. All or most of this development should be located in a one-half mile radius of the station. This suggests that BART’s current practice of reserving station-adjacent land for large above-ground parking garages should be corrected at existing stations and avoided for any new stations that are built.
Filling In the Urban Core
We can also infer from the data prime locations to augment service and increase the value of the system by constructing infill stations. An infill station is under construction at West Dublin/Pleasanton, filling the long gap between Castro Valley and Dublin/Pleasanton, and that station will anchor a transit village. But we can also build infill stations to bring under-served urban neighborhoods (where there is already a transit-dependent population) into the fold of the system. BART stations are closely-spaced in the San Francisco and Oakland CBDs, but they are usually at least a few miles apart in the suburbs. This has prompted the observation that the BART system is not quite a proper urban subway, nor quite commuter rail, but rather is a hybrid of the two. Within the urban core, but outside of the CBDs, the stations are more closely-spaced than in the suburbs — but not as closely-spaced as they really ought to be to properly serve the corridor. BART’s data confirms the rule of thumb that most people are willing to walk to a station within one-half mile but become disinterested in walking distances greater than one-half mile. Ideally, then, BART’s urban stations would be spaced so that mostly everyone living on the corridor is within one-half mile of a station. That is often not the case, because the system was originally designed to facilitate quick trips from suburbs to urban CBDs that would be competitive timewise with freeway driving. The result is that many urban areas outside the CBDs, even those neighborhoods on the right-of-way, lack convenient (i.e. easy non-vehicular) access to BART.
 |
| Courtesy of BART. |
Nowhere is this more true than in Oakland, southeast of the lake, where a roughly 7-mile segment of the BART corridor parallel to Interstate 880 and East 14th Street is served by just two BART stations, Fruitvale and Coliseum. The fabric of East Oakland grows increasingly suburban as you move toward the high-numbered avenues; but the area is still denser than most other places in BART’s domain, and it deserves better service. There is no station in the 3-mile stretch between Lake Merritt and Fruitvale, which means that BART trains whisk right past the Eastlake and San Antonio neighborhoods. This is one the densest sections of the East Bay, but the local residents are in large part not riding BART. This can be inferred from the survey data. Check out the map excerpted above; orange dots represent residential points of origin. There is a gray hole with no orange dots, roughly midway between the Lake Merritt and Fruitvale stations, indicating a lack of rider surveys from that area. A more fine-grained analysis shows that the highest density of riders using Lake Merritt Station live north of 8th Avenue; similarly, the highest density of riders using Fruitvale Station live south of 25th Avenue. The transit-dependent population that lives in the middle zone currently relies on AC Transit to get around rather than BART; but high population density, coupled with the presence of nearby commercial districts, suggest that this would be a successful infill station.
 |
| Courtesy of BART. |
Another natural location for an infill station is 30th Street/Mission, which would close in an almost 2-mile gap between 24th Street and Glen Park, the longest stretch of BART track in San Francisco without a station. As shown at the map at right, the survey data demonstrates a hole in ridership near 30th/Mission. The hole is less pronounced than that near San Antonio — in part because of the shorter distance to nearby stations, and in part because this intersection is already a nexus of several Muni routes, many of which connect directly to 24th Street or Glen Park. Indeed, the fine-grained data shows that the densest ridership at 24th Street lives north of Cesar Chavez; similarly, the densest ridership at Glen Park lives south of Holly Park. Residents near 30th/Mission that use BART generally choose 24th Street station, but they mostly access the station via Muni or drop off, rather than by walking. The station at 30th Street would have the benefit of drawing riders from neighborhoods like Bernal Heights and Baja Noe, further removed from Mission Street, thus opening up a new cross-section of residential neighborhoods (and an additional commercial segment of Mission Street) to BART service. The station site also presents nice opportunities for transit-oriented development, both at the large Safeway parking lot and at smaller vacant lots scattered throughout the area.
Transbay Blog 
As always, awesome analysis. Did you read the Smart Growth Around America entry on Arlington’s growth? If not see here: http://blog.smartgrowthamerica.org/2009/05/08/arlington-virginias-story-of-smart-growth-the-movie/
Given BART’s and WMATA’s similar history and general operational & design philosophy, I’m saddened we didn’t choose a similar zoning practice back when BART was being built like the Arlington/Rallston corridor did.
Also, IIRC back when SFCityscape was blogging, I think the track grade at 30th & Mission causes some problems with trying to build an infill station there. I would move to that hood in a heartbeat if there was a BART station there.
Posted by eddo | 10 May 2009, 5:00 pm* not Arlington/Rallston, Rosslyn/Ballston ;)
Posted by eddo | 10 May 2009, 5:02 pmeddo: Thanks for writing in. Yeah, it’s a 3% grade at the current track alignment, which won’t work. But I think that that issue could be resolved by adding more level platform track just outside the mainline. At any rate, the engineering problem wasn’t insurmountable, just not maximally straightforward. If Steve sees this post, he’ll probably remember more on that.
I did read that Smart Growth post, thanks for adding the link. The BART/WMATA topic is definitely a “tale of two cities” sort of comparison. It’s a great encapsulation both of how much we have left to do.
Posted by Eric | 10 May 2009, 5:23 pm@eddo
There is an issue with track grade at 30th and Mission, however BART has already studied several ways to deal with that and put the price tag at $500M (http://www.bart.gov/docs/planning/30th.b.pdf)
Not chump change, but a fraction of what’s going into the San Jo extension.
Posted by Josh | 10 May 2009, 5:40 pmThe big get from all this for me was the fact that we need to connect BART to our major destinations if we want people to ride. It also points to the fact that we really need to stop extending further out and focus more on the core. A la Geary (I’m still writing!) and Broadway. Those are the places that are most likely to have an Arlington type return.
Posted by The Overhead Wire | 10 May 2009, 11:49 pm^^ I’ll drink to that. The infill stations are part of it too. 30th/Mission and Eastlake are awesome commercial districts, and having stations within walking distance would be a boon to the neighborhood. But I also think we need to leverage BART where we have it now, to transform current non-destinations into destinations (which is how we got Arlington, Bethesda, etc. in the first place). We should extract as much worth as we can from even the more distant stations. I fear our “transit village” efforts to date have been too timid. Politically, though, it is easiest to put growth downtown rather than in neighborhoods, which is probably part of why we saw some of the the biggest increase in home origins at CBD stations.
Posted by Eric | 10 May 2009, 11:58 pm“This suggests that BART’s current practice of reserving station-adjacent land for large above-ground parking garages should be corrected at existing stations and avoided for any new stations that are built.”
Completely backwards. This means that the only people who are willing to take BART are people withing walking distance of the station, UNLESS a parking facility near BART is built. If you get rid of the parking, you will change the pattern of BART users around the station, but you will draw on a smaller pool of customers.
Posted by DaveO | 11 May 2009, 7:51 amDaveO: read that sentence you quoted again very carefully. “Getting rid of parking” is not its only logical conclusion. That is the conclusion you’ve imputed to it. Suburban stations will require some parking, but BART’s placement of that parking — either with surface lots, or with unnecessarily gigantic parking garages immediately as you step out of the station — needs to be rethought and redesigned. One reason why I said that is because of the BART-SJ extension, which anticipates some truly mammoth parking structures.
As for your comment about a smaller pool of riders, I think this also depends on how you design a neighborhood. This is the exact point of what BART’s survey data shows, and this was the exact reason why I went through the trouble of excerpting those maps and including them in the post. Ridership at 16th/Mission is several times higher than what’s at North Concord, even though most riders at 16th/Mission are coming from within one-half mile of the station (see the map in the post). This implies that the way to increase ridership is exactly by creating dense neighborhoods in the immediate vicinity of stations. Providing lots of parking is by no means an guarantee of ridership (see, e.g. tons of empty parking at Millbrae). Density, on the other hand, is a far surer guarantee. In the long-term, the system will be better-used if we prioritize urban density near stations.
I would really suggest you click that Arlington link that eddo provided above. WMATA and BART are similar in many respects — about the same miles of track, similar age (WMATA is just slightly younger and learned from our mistakes). Two big differences between BART and WMATA? WMATA has way more stations, and it has essentially created urban downtowns next to its metro stations — and this is critical — even where the surrounding area is suburban. (Cross-reference: I also wrote about this topic last year on this post.) The point is: this approach, which we have mostly not followed in the Bay Area, creates a much larger pool of riders than a parking lot. And for doing that, WMATA’s ridership is about double BART’s, for about the same number of track miles. We have a long way to go on this.
Posted by Eric | 11 May 2009, 9:04 amAs usual, Eric, a fantastic post! Even though I don’t always comment, I follow your blog regularly. It’s my favorite out of all the Bay Area transit/land use blogs. Keep up the good work, and I look forward to reading your future posts.
BTW, does anyone know what ever happened to the 30th St and Mission Station? I read parts of the study from the BART website, but there is no indication as to whether they are moving forward with this project or simply have it eternally on the back burner.
Posted by Danny | 11 May 2009, 10:46 amDC Metro geys more riders because it goes mor places–several different routes thru downtown, likewise river crossings to VA–direct to National Airport from the begining of the Yellow line. Its as if BART ran out Geary, crossed the Golden Gate, had a west side of SF N-S trunk connecting the Richmond/Matin directly to Daly City and south., aand a Van Ness to BVHP via the Potrero, Bayshore corridor, and the Muni Metro had remained a BART tunnel.
Further thoughts, Bethesda was already a busy area–even 40+ years ago when I was a high schooler, IBM already had its Federal Systems Division offices on Wisconsin not very far from the Farm Women’s Coop–nice time warp to today’ chic farmewrs’s markets.
Infill? yes, And fgar better cost benefit than even a mile of the the SJ boondoggle. Along w/ a station east of LM, there is fallow land there for a mini yard/telay track installation to hold OOS trains, lay up trains to serve events @ Coliseum–which needs to be three tracks two platforms.
Posted by david vartanoff | 11 May 2009, 12:03 pm@Danny
The 30th St station is gathering dust pending the political leadership to put it forward
@david vartanoff
The Van Ness/Potrero line is the old Muni H streetcar. I completely agree that this should be brought back and connected via C Chav or Evans to HP
Posted by Josh | 11 May 2009, 1:41 pmDanny, thanks for reading (and for emerging out from lurkdom to comment!).
Re: 30th/Mission, it’s in limbo until we make it a priority. BART and MTC are wrapped up in extension projects (BART to Warm Springs, BART to SJ, eBART, Oakland Airport Connector), and aren’t making the urban core a priority. This philosophy is actually changing, but it’s somewhat embryonic and theoretical at this point. And so much prior planning has already been done on those extensions that we aren’t seeing real steps being taken yet to step up urban service to where it needs to be. Basically, the suburban projects are the ones being funded (or vying for funding) right now, while the urban improvements are being postponed until the next generation of projects. This is why you see the sort of backwards approach, in which all the expensive rail projects are in the suburbs, while the high-ridership urban corridors are basically only getting cheaper BRT projects (Van Ness, Geary, and Telegraph/East 14th). The Central Subway is the notable exception … but then, you probably already know my thoughts on that. And then there’s Caltrain DTX, which is the rare gem in the bunch.
Posted by Eric | 11 May 2009, 1:51 pmFascinating read. I lived in Arlington for 6 years from 1994-2000. I watched the development along Wilson Blvd sprout during this time and after when I returned for a visit. A couple points to mention…
The orange line was in place along this corridor by the early 80s. Each station and its surrounding area was to have a particular focus…government (Courthouse), education (Virginia Square), urban village/retail (Clarendon), etc. Over the course of two decades development grew organically around these stations, rather than dozing a 2-mile swath and putting up towers and condos. It took a while for the empty lots to disappear. Today, there are still some empty patches, but the flow from Rosslyn to Ballston is mostly uninterrupted.
On the flip side, the development around Potomac Yards pales in comparison. There is no metro station within walking distance (the plans for a privately-funded Potomac Yard station fell through several times). Most people rely on the auto to get around as bus service on Route 1 is slow and unpredictable because of high traffic volume. The center attraction is a huge suburban-type strip mall surrounded by condo towers. As a result, the county has to backtrack to figure out how to incorporate better means of public transit to this area.
Back to the Bay Area…what I’ve noticed with transit villages around the BART stations is that they don’t assimilate into the existing neighborhoods. At Fruitvale you have this behemoth structure that adjoins the commercial district on Intl., but it feels like it doesn’t want to associate with the rest of the neighborhood. The dismal development at EC Plaza earlier in the decade showed the lack of vision of creating a sustainable commercial/residential addition to the community. So much promise for that area and all you got was a strip mall with styrofoam architectural detailling. Very auto-focused. Does not blend in with the community along San Pablo Ave. ZERO housing built.
There are lessons to be learned throughout the process. The most important ones are learned from others’ mistakes.
Posted by Mark | 12 May 2009, 9:13 am