Atlanta Light Rail Streetcar

Readying Atlanta for Its Bright Future

» Resurgent city plans major transit expansion in the form of streetcar and light rail lines, but will they serve its growing population well?

When Atlanta hosted the Olympics in 1996, it had the chance to project itself as the epitome of progress in the New South. But the city was sick. Despite massive growth in its suburbs, Atlanta had been losing population since 1970. The Games were marred by poorly planned transit service and a bombing that killed two people and injured 100 others. The year before, the city had been rated the most dangerous in the United States.

Atlanta rebuilt itself quickly into the capital of the South after it was destroyed during the Civil War by General William T. Sherman, who described it later: “Behind us lay Atlanta, smouldering and in ruins, the black smoke rising high in air, and hanging like a pall over the ruined city.” The Atlanta of the early 20th century was dense and active, vibrating with business and public enterprise at the height of contemporary modernity. Now more than ever, the metropolis has the potential to be remade into the walkable and exciting city it once was. Its population reached 540,000 last year, a historic peak. Crime is down dramatically, and redevelopment has taken hold in neighborhoods from Buckhead to the West End. The city’s three skylines in downtown, midtown, and Buckhead are bursting with activity. But this growth needs to be put in order to create a livable city through the effective use of public transportation.

Last week, the city of Atlanta applied for $300 million in stimulus funds for the construction of a new streetcar in downtown and midtown. The U.S. Department of Transportation will make a decision on whether to allocate money for the project by February, and if it is approved, it could be complete by early 2012. The first phase of the Peachtree Corridor Streetcar would extend roughly three miles north-south on Peachtree and 1.5 miles east-west between downtown and the Martin Luther King, Jr. Center in Sweet Auburn. The program is to be implemented by a group called Georgia Transit Connector, which expects 12,000 to 17,700 daily trips.

Though the streetcar project would be the first major transit expansion within Atlanta city limits since the Bankhead Station on the MARTA Proctor Creek Branch opened in 1992, the city has a number of other plans under active consideration. Most prominent is the Beltline, which would bring light rail to 22 miles of (mostly) inactive railroads ringing the city’s core. The $2.8 billion project has the strong support of current Mayor Shirley Franklin and most of the city’s other politicians. Its primary objectives would be to leverage brownfields adjacent to the improved line for new development (including 5,600 units of affordable housing) by improving public transportation, building 30 miles of trails, and creating 1,200 acres of new green space. It is one of the most innovative plans for using transit as a redevelopment tool in the country.

That’s not it, of course: MARTA has studied a number of proposals including a new bus rapid transit line on I-20, a light rail corridor along Clifton Road, and several suburban expansions of the existing heavy rail lines. The recent Connect Atlanta plan plans for a city with a population of 800,000 by 2030 with densities in the city core of 12,000 people per square mile compared to 3,500 today. It recommends a 5-mile east-west line on the north side of the city, the covering of freeway interchanges downtown, and the better connection of roads to make a more efficient street grid.

These ideas, however, do not do enough to take advantage of the city’s existing resources.

Atlanta Future Transit Map

The north-south segment of the proposed streetcar, the project city officials have pushed forward as their top priority, is the most problematic, because it duplicates existing service on MARTA’s main north-south line between Five Points and Arts Center. Its three miles of stops would be just two blocks away from MARTA’s closely spaced stations, which are only about half a mile apart in this part of the city. While the streetcar would encourage growth on the already dense Peachtree Street corridor, it should not be the first new transit line for Atlanta.

Indeed, a streetcar line running north from Arts Center station to Brookhaven-Oglethorpe station along Peachtree Street and Road would have the advantage of providing public transportation to currently under-served neighborhoods and increasing connectivity along MARTA’s existing routes. This has been mentioned as a second phase of the first streetcar line. Such a line would run through the heart of rapidly expanding Buckhead, both north and south of the existing MARTA station, and reinforce plans to remake the neighborhood into a walkable place. It would connect to Piedmont Hospital where it could link with future Beltline service, and it would provide direct service to the Amtrak station north of Arts Center. This seven-mile line could extend Buckhead’s growth south and east and push Midtown to the north, allowing for the replacement of several downtrodden strip centers along the street.

Although the east-west streetcar line proposed by city officials would improve connectivity among Atlanta’s top tourist areas, its 1.5-mile length makes it too short to be of much use. Indeed, what if it were extended north to serve the intown communities west of the Downtown Connector highway? A four-mile extension would bring the streetcar north on Luckie Street past Coca-Cola headquarters, along Tech Parkway through the center of Georgia Tech, and north on Northside Drive, where there a number vacant properties waiting for new use. Most importantly, though, this line would return east on 16th or 17th Streets to Arts Center, passing through Atlantic Station, one of the country’s largest urban development projects.

Connected, these proposals would create a 12.5-mile streetcar that would provide a valuable transit expansion in the inner city, acting as a counterpart to both the Beltline and existing MARTA lines. A Peachtree line between midtown and downtown could come later.

But these plans will need adequate funding. As Secretary of Transportation Ray LaHood put it last week during a visit to Atlanta, “There has to be a commitment by state government that transit is important.” When Georgia decides to bring mass transit to the front burner, and when the city finds itself willing to devote more tax funds to transportation improvements, the city’s potential as a modern, livable city will be reached once again.

36 replies on “Readying Atlanta for Its Bright Future”

Someone from Atlanta once told me that MARTA basically covers all of Atlanta inhabited by black people. The state has never funded MARTA and the suburban counties don’t want it anywhere near them. I think that has changed somewhat in the last ten years, but I don’t foresee tons of money flowing into anything “socialist” like mass transit. If they want to quadruple their downtown density, they are going to need more than one streetcar route, especially since the rail stops are spaced so far apart. I’d think 3 or 4 crisscrossing through downtown would be the minimum, something like what they are doing in Houston. Are Georgia and the Atlanta suburbs capable of supporting that level of investment over the next 20 years?

AlexB –
That person from Atlanta didn’t know what he or she was talking about: while MARTA isn’t the most extensive of rail systems, it serves plenty of diverse areas in the region. The communities on the northern sections of the North-South lines, such as Buckhead, are majority white.

Meanwhile, as I wrote in the most, the stations between Five Points and Arts Center are just about 1/2-mile apart, which is as close as you’re going to get. I’d also like to note that development in Atlanta is very linear compared to that in other cities, meaning that while it might be justifiable to see a second north-south heavy rail line going through downtown, more than that would probably be over kill.

AlexB, I’d say you were misinformed about MARTA. MARTA covers Atlanta inhabited by both white and black people. If anything, it favors the “white” parts of town too much. Since moving to Atlanta a few months ago, I’ve noticed that the demographics of MARTA riders during peak hours are roughly similar to the demographics of Atlanta as a whole. During off-peak hours MARTA mostly serves those with fewer choices.

Yonah, I share your concerns about the Peachtree Streetcar. It seems to me that it’s mostly a pet project of downtown and midtown developers. Could you do some analysis on the beltline? I’ve heard a great deal of skepticism about it since I moved here, both from transit supporters and transit haters. It seems many people think that the mechanics of connecting it to MARTA are too complex and that it’s potential usefulness relies too much on potential dense development along the line, mostly opposed by surrounding neighborhoods. Any input you have would be appreciated.

Atlanta seems like a great place to try deploying PRT. The clusters of density could be linked with the commuter lines and each other effectively with destination to destination service while achieving better cost recovery, better service, and comparable or higher volumes to light rail transit. I’d like to see PRT put to a serious, scalable test in ATL.

PRT is ridiculous. It has never worked because its main idea is flawed. You cannot have low density vehicles (3-5 people per pod), on such an expensive guide way. Also during peak hours the amount of pods needed is prohibited by the amount of track available. The idea is probably technically feasible. It lacks any sort of positive track record. The south is generally very wary of mass transit, supporting PRT will do more to hurt urbanity in the South than help.

I agree that the southern portion of the Peachtree Streetcar should not be a priority; it’s too close to the heavy rail line. But I also do not support the northern portion of the line. Buckhead is under-served only because MARTA has been cutting back bus service along Peachtree due to budget constraints. And this area does not need any potential development stimulus that a streetcar might bring. In my opinion improving bus service is all that is needed in this area (signal priority, more frequent service, etc.) Priority for new rail lines should be given to the Beltline or to the west-side line that you describe (although the alignment should be shifted a bit east, so that it would go through Georgia Tech; Tech Parkway is the western edge of the campus).

two points – enough about PRT already. save it for the glamor pet projects in the middle east. It’s not a realistic alternative. we already have a great PRT system – it’s called the automobile.

Also, is there really a part of Atlanta called Blandtown? that’s great. You’d have thought some branding expert would’ve had them change it to “exciting town” by now…

I don’t think it’s worth dismissing PRT out of hand. You solve the problem of choosing among multiple alignments (something Yonah is grumbling about constantly – see Portland, this article just in the last week), you increase WTP by reducing 95% of headways to less than 1 minute and by offering a more comfortable, faster, nonstop trip. The economics works out pretty well where there isn’t sufficient room for parking (okay, maybe that’s not Atlanta so much but DC, SoCal, Bay Area, and other corporate campus places come to mind), and the capacity matches that of LRT, except the operating costs are lower than for scheduled transit. Medical centers, corporate campuses, universities, and even major malls might all give it a look because the cost of a spur is probably worth the opportunity to replace parking with usable office space. APM are well tested and the concept is not much different – just a little more guideway and more complex collections. There’s more experience with these complex systems all the time. And the marginal cost of expanding the network is lower while the marginal benefits are potentially much higher than for LRT.

the capacity matches that of LRT

Says who? Certainly not reality, where PRT headways are limited to those of cars, making a two-track PRT system equal in capacity to a two-lane highway.

PRT headways aren’t necessarily limited to those of cars, depending on what rail safety boards will approve, and typical highway lane capacities about match those of LRT with average headways. If PRT has higher load factor (you’d need some sort of solution like they now use for elevators to bundle peak hour travelers with the same destination into cars together – although too many stops would create too many DOF), then it’s like a HOV lane at capacity. Lower speeds than highways (45mph vs 65 mph) mean shorter headways, which you could reduce even more b/c of the automation. I’d like to see a model of the optimal car size, station spacing, and headways (how long you have wait for a departure on average) to see whether it leans towards resembling LRT or it’s more like what most PRT vendors currently offer.

LRT can move 15,000 people per track per hour. A highway lane can move 2,400.

Anyway, re: Atlanta, the most damning part of your map isn’t the streetcar, but the Beltline LRT. When you build a circumferential line without easy transfers to the radial lines, you get New York’s G train. Hopefully they’ll either make major revisions to the routing, or at least build infill stations on the existing lines to let people make transfers.

Presently Atlanta is an example of building transit and urban development around the stations without any concern for pedestrian friendly urban design. blank walls, wide high speed streets, ugly above ground structured parking, dead ground floors along the street, skybridges, privatized public space, multi-block megastructures. I’m skeptical that TOD can be done in Atlanta, the developers only know how to do “edge city” development and the customers only go to malls or mall-like places.

then you have faux-urban atlantic station a good distance from MARTA metro where the rent a guards go ape sh*t if you take a photo of the project. so much for being ‘urban’ if you have draconian rules against something like photography.

you want PRT?… try “bixi”

Also, is there really a part of Atlanta called Blandtown?
I thought that was called Peachtree Center :)

So would the belt line be a full loop or 2 separate lines… 1. Lindburgh Center to Inman Park via Piedmont Park and the eastside and 2. Lindburgh Center to Inman Park via Bankhead, Blandtown and the westside?

I agree it connects terribly into the metro rail lines, they needs some serious deviations to Bankhead, Ashby and the West End, a more direct route from the Carter Center to Inman Park and definitely some infill metro stations.

I was here for the Olympics. The bombing and lives lost and injured were tragic. But apart from that, the games were incredible and the sense of goodwill in the city was awesome. We rode Marta on several occasions without incident and met interesting people from all over the world who were also happy to be here.

LRT can move 15,000 people per track per hour. A highway lane can move 2,400.

No U.S. LRT system comes close to having that capacity. The two highest ridership systems in the U.S. both manage 3 minute headways and Muni is capable of running 3 car trains. Unfortunately, both Boston and San Francisco are stuck with fairly narrow tunnels which require narrow trains, so their LRV’s only handle about 160 riders each. This gives a capacity of (160 people/LRV)*(3 LRV/train)*(20 trains/hour) = 9600 people/hour. Not too shabby, but not 15,000 people.

Well, what about a modern system? Surely they can do better. Let’s look at the Los Angeles Blue Line – the busiest single line in the country (SF’s and Boston’s high frequencies are due to many lines converging on a single point). Los Angeles runs modern LRVS mostly on a dedicated ROW and with 82,000 daily riders is facing fairly severe crowding. Their vehicles can carry 230 people each, giving 690 people per 3 car train, but due to the large number of intersections it crosses, it is limited to 12 trains per hour, giving a capacity of 8280 people per hour. Decent, but not close to 15,000 passengers per hour.

While you can, in theory move 15,000 people per hour on a light rail line, 10,000 seems to be the realistic limit for any system once you actually build it. As for a highway lanes, there are several HOV-3 lanes that manage to move over 5400 people per hour without doing anything especially special. While it is true that you CAN move more people with in a corridor of a given width LRTs the actual capacity difference is a bit less than a factor of 2, not the factor of 7 that LRT advocates like to claim. The difference is even more dramatic when you compare the real world performance of the two, but this is something for another day.

I was just in Atlanta. It’s a disaster if you goal is a coherent city, but it has nice neighborhoods, if you don’t mind driving. With regards to the tram proposals. Ponce De Leon through Donald Lee Hallowell Pkwy would seem to be a good bet, because there is no cross-town option in Midtown. As the article mentions, the area around Peachtree is really dense, and really close to Marta, but if you go out about Five Blocks it stops being a city. The Monroe Ave./ Blvd. Ave corridor is another area that would greatly benefit from TOD. This Could run from the Beltline (South), through Grant Park, and up to Piedmont Park. Constructing these two lines would greatly improve mobility. However, ATL really needs to think about how to create more complete pedestrian neighborhoods. Midtown is not a paragon of a livable urban neighborhood.

No U.S. LRT system comes close to having that capacity.

The US isn’t the only country in the world. In Calgary, whose light rail is the busiest in North America, the current capacity is 15,000, and with currently built upgrades to 4-car trains it will increase to 19,000.


Yes, it can be done, just as car lanes can be made to carry 8000 people per hour (HOV-4 lanes on a metered freeway would do that without difficulty). I never disputed that. However, the fact remains that 15,000 people per hour per direction isn’t a typical capacity for a light rail line as advocates would have you believe. Most actual systems are in the 6000 people/hour/direction range and even the busiest systems in the U.S only have a capacity of 10,000 people/hour in each direction.

HOV-4 lanes on a metered freeway would do that?

If you can find 2,000 4-person carpools per hour, perhaps.

But there’s a reason that HOV-4 lanes are pretty rare–besides transit vehicles, the sort of vehicles which typically use these are families with children, not commuters carpooling to work. Nothing against hauling the kids around, of course–but 4 persons per carpool will probably deter carpooling, unless there is a significant penalty for otherwise.

And in which case, the logistics involved will probably encourage transit use instead. :)

However, the fact remains that 15,000 people per hour per direction isn’t a typical capacity for a light rail line

Light rail where? In the US, where transit builders have all the competence of a retarded giraffe? Outside the US, 15,000 ppdph is standard for light rail. The 2,000 cars per hour per lane limit for highways is absolute, however.


You would also have a hard time getting 15,000 people/hour/direction on a light rail line. That’s a portion of my point.

Winston: Four-car trainsets at three minute headways will do it just fine; LRT systems generally have little problem with such headways. And that’s at typical design loads, not crush loads. The exception is when the LRT has lots of closely-spaced stops.

(For instance, MAX in Portland runs three-minute headways each direction across the Steel Bridge; albeit in two-car trainsets, so the capacity is closer to 7000 pphpd. In order to get that performance, though, they had to add a second line through the downtown street network, which operates in the street grid and can’t do better than 5-minute headways or so. Short block lengths in Portland limit trains to two cars each, but that’s an issue specific to Portland, many LRT systems routinely run longer trains).

The only way a pavement gets that many people through on a lane is if its a busway–or if every vehicle is a minivan, loaded front to back with passengers.

Practically speaking–and this is to address Alon’s point–there’s not many places in the US where there is that level of DEMAND. The reason is simple–we’ve extensively built out and subsidized our road network, and for many trips, the car is a more “rational” choice in the US, especially if you already own one. Comparing transit builders (do you mean transit engineers, planning professionals, managers, or whom exactly?) to “retarded giraffes” is a bit unfair–while planners certainly do stupid things, they also operate under cost and political constraints that don’t exist in Europe.

I don’t believe the cost constraints are the real problem. Again, look at Calgary – it’s not only the busiest light rail system in North America, but also the cheapest to construct per route-km. The original construction cost $15 million per route-km. When you make sure to minimize costs, you can build light rail surprisingly cheaply.

What does Calgary do to minimize costs, that is not being done in other North American locales? Did the government buy up land a long time ago, and doesn’t have to acquire ROW? ROW is frequently a big expense in constructing a LRT line, especially in places where eminent domain is expensive (and affected property owners have avenues of legal recourse).

Conversely, in many developing parts of the world, landowners will find their land seized for public projects (including transit) for little or no compensation, and with little or no legal recourse.

It’s interesting to compare CTrain (Calgary’s system) to MAX. MAX has nearly double the amount of revenue track–54 miles (84 km) vs 30 miles (48 km) for CTrain. Both operate modest, self-service stations with automated ticket vending, and use the proof of payment technique for fare enforcement. CTrain doesn’t have the Streetcar-ish stop density and street-grid running downtown that MAX does, but is still a mostly at-grade system. Yet CTrain gets nearly 3x the number of riders that MAX does.

I suspect the difference in ridership is not in the competence of the transit planners, or in any particular failure of Tri-Met to adequately design the system. As speculated before, the main difference is that Portland has a robust freeway network whereas Calgary does not. Portland’s freeways are smaller in scope and less modern than the 8-lane behemoths found in places like Southern California; but you can get to most any part of town on them. Calgary is one of a handful of North American cities (especially excluding Mexico) where transit is more convenenient for a significant number of trips than driving; Portland, despite not embracing the full-on auto culture of many other US burgs, is not on that short list.

I agree that 4 can trainsets of modern LRVs with 3 minute headways will do it, but that’s actually kind of difficult to achieve, either because your train length is limited by the city street grid (many cities, such as Portland have blocks less than the 328 foot length of a train) or because train frequency is limited by the desire not to interfere with traffic on cross streets (as is the case with the blue line in L.A.) or by track configuration (as is the case in Sacramento). I’m not asserting that light rail isn’t able to move more people per square foot than a freeway lane or even an HOV lane. I am saying that when talking about light rail people should talk about what can actually be achieved in a specific corridor given the real constraints that exist. In most cases it isn’t 15,000 riders per hour per direction.

I would also say this is usually due to GOOD decisions on the part of the designers of the light rail system. There are very few light rail lines that actually operate near capacity and the cost savings from avoiding grade separations and from single tracking are often worth the capacity hit that you take.

In Calgary’s case they had reserved corridors for high capacity transit since the 1960s and the cost of the ROW is not included in the cost of building the system. Alon’s numbers are also not inflation adjusted. Looking at Wikipedia’s cost numbers for Calgary’s light rail system and inflation adjusting them through the BLS inflation calculator gives an inflation-adjusted cost of $30.7 million/mile, which is low for a light rail system but is understandable given Calgary’s far-sighted decision to preserve the right of way. Calgary also benefits from not having Portland’s short blocks along its route and from having a less extensive road network than many north American cities. As a side note, but related to another thread, Calgary IS an example of a light rail system that has extensive freeway median operation, something that U.S. transit activists often oppose.

One problem is that in North American English (we’ll ignore the UK, Australia, New Zealand, etc. for now), we’ve got numerous terms which describe different types of rail line used in transit applications. Some terms (“subway”, “el”, “streetcar”) describe the location of the track; some terms (“circulator”, “commuter rail”, “rapid transit”) describe the intended purpose of the service.

And some terms are used to describe the relative “level” of service, as a combined function of route topology, station placement, etc. “Streetcar” is used in this manner as well, as is “metro” and “light rail”.

The trouble is, there are more different “levels” of service than there are terms to describe them in English. There are at least five different levels of service for rail transit that I can think of:

1) In-street, mixed-traffic running local service, generally with frequ,enty stops.
2) Next-to-street running–often with no physical separation between trains and other traffic (cars, pedestrians), and often without signal pre-emption at crossings.
3) Adjacent to a street (either alongside or in the median), with frequent crossings and stops, but some physical separation between the trains and other traffic.
4) In a dedicated right-of-way apart from a road, with more limited crossings (and crossing gates or other barriers installed), but generally at-grade.
5) In a physically protected (either grade-separated or fenced-off ROW) with no (or few) at-grade crossings and no pedestrian access to tracks or train other than at platforms. May have third-rail electrification, driverless operation, or very high operating speeds (>80MPH or so), all of which are incompatible with 1-4.

The first level of service generally gets called “streetcar”, and the last is often referred to in English as a “metro” (when more specific terms like “subway” aren’t used). But “light rail” runs the gamut between 2-4, and many systems called “light rail” even have significant stretches of 1 and 5. MAX, for instance–the transit mall line and Steel Bridge crossing are #1, the route through downtown Portland and downtown Hillsboro generally is #2, Interstate MAX and the Blue Line along Burnside is #3, the Blue Line between Beaverton and Hillsdale and in parts of Gresham, where MAX uses former freight ROWs is #4, and the freeway-running parts of MAX and the West Hills tunnel are #5.

It might be useful to have terminology which better distinguishes between these. Metro rail has no problem with 15,000 pphpd (and can do many times that), #4 light rail generally has no problem either. But it’s a stretch for street-median LRT, and probably a no-go for #2 and #1–both of which Portland has in spades downtown.

Similar analogies apply to BRT, which depending on the context can refer to any of the above levels of service, albeit with caveats appropriate to bus. And one of the main advantages of BRT is that a route can “drop” to #1 level service for short stretches without difficulty, whereas running trains in mixed traffic severly limits their operating characteristics (length and speed).

Calgary’s less extensive road network wasn’t an accident. As early as the 1960s, the city was reserving ROW for future transportation expansion; once it made the decision to go with light rail instead of BRT or freeways, it used those strips of land for rail only instead of slapping highways on them. It’s a basic competence issue – you don’t try to help the competition.

Calgary engaged in other cost-saving schemes, such as avoiding tunnels and elevated sections at all costs, including the cost of grade crossings. It also made sure to avoid sharing ROW – the downtown transit mall is reserved for light rail, buses, and emergency vehicles. Right now the city is planning to underground the transit mall, but that’s after decades of successful operation. Other cost-saving decisions were operational (no automatic train operation, no air-conditioning on trains) or managerial.

The original $15 million/km figure is not inflation adjusted. However, unlike many other cities, Calgary managed to keep cost growth below economic growth. C-Train’s website lists the current cost of construction as $15 million/km on the surface, $30 million/km elevated, and $35 million/km underground.

Alon wrote:

It’s a basic competence issue – you don’t try to help the competition.

Therein lies the rub for many transit agencies–the “competition” is utterly out of their control, and in many cases is not only competing with transit for users, but for captial as well.

For most transit planners, “lets close down the freeways” is simply out of scope.

That’s an issue of urban competence, not planner competence. But even within the parameters of transit planning, the C-Train cost less than most comparable projects in the US (or elsewhere in Canada). By non-US standards, it wasn’t that cheap – e.g. the Lyon tramway, constructed in the 1990s and 2000s, came in at €19 million/km (link; the costs are stated in francs, but in 1999 the value of the franc was fixed at €0.1524).

Some of the methods used by Calgary can be described in this paper, by the way – e.g. Calgary limited park-and-rides, which are expensive to construct.

I am sorry I made an erroneous comment about racial segregation in Atlanta, I was misinformed. My main points were: 1) Georgia is not the best when it comes to funding transit. 2) 12,000 people/sq mile is a lot and will require a whole network of streetcars. Perhaps if the city is organized linearly, it might not need many new lines, but it will need more than one new streetcar. What specifically is the area of the map that they want to densify? How many square miles of 12,000/sq mile do they want?

Leave a Reply