Transportation Deployment Casebook/Lifecycle Analysis: Ridership of San Francisco BART system

As a rapid transit system, Bay Area Rapid Transit (BART) has served the San Francisco Bay Area for 40 years from its beginning operation at September 11, 1972. Cities of San Francisco are highly connected by the heavy-rail public transit subway system form East bay and suburbs in northern San Mateo County. In those four decades BART has carried more than 2.7 billion people around the San Francisco Bay Area. In this case, the life cycle of ridership of BART between 1973 and 2012 is analyzed, and future trend is estimated. Then the birth of the system is studied, especially focus on the market and policies at the beginning operation. After that, the growing phase is discussed and a hypothesis on its future development is provided.

Life cycle modeling
The annual ridership data of BART is obtained from the official report of Bay Area. The data is used for descript the S-Curve and estimating the birthing, growth and maturity of the system. In the dataset, the annual ridership of BART from the first year operation (1973) to the present year (2012) is extracted. It is described below in Table 1 and Figure 1:

Then the data is used to estimate a three-parameter logistic function as follow:

S(t) = K/(1+exp(-b(t-t0))                    (1)

In this equation, S(t) is the status measure, t is time (in years), t0 is the inflection time (year in which 1/2 K is achieved), K is saturation status level, b is a coefficient (K and b are to be estimated). In this case, S(t) is the annual ridership of BART(short for R(t)).

Then we change the equation as follows:

exp(t-t0)= (R(t))/(K-R(t))                   (2)

And

Y=Ln ((R(t))/(K-R(t)))=bt-bt0                (3)

Form the equation 3, single liner regression is used to study the relationship between the annual ridership and the year. The independent variable of the function is fiscal year and the dependent variable is the natural log of annual ridership (R(t)) over the value of saturation ridership (K) minus annual ridership. However, the K value should be estimated first. Different K value based on the growing trend of annual ridership should be tested in order to get the best regression. In the K value test, I choose 15 K values for regression, from 111,000,000 to 170,000,000 (the interval of first 10 values is 1,000,000, and of the last 5 values is 10,000000), and part of the regression summary is as table 2.



To get a better regression result, the values of R-square and t-stat are important. For the value of R-square, it is better to close to 1 and for the value of t-stat, the greater the better. The following chart gives us a clear view of the fluctuation of R-square and t-stat basing on different K.



From the table 2 and figure 2, it is clear to find out that the optimal regression result could be get when K value is 116,000,000, which means the saturation annual ridership of BART is 116,000,000, as estimated.

Hence, when K=116,000,000, The regression result is as follows:



From the regression result, we can get b=0.101. According to the equation 3, -bt0 =-201.238, so we get t0=int [(-201.240)/(-b)] =1988, which means half of annual ridership of saturation is reached in the year of 1988. Then we put the estimated K, b and t0 to the equation 1, we get the new function as follow:

R(t)= 116000000/(1+exp(-0.101(t-1988))       (4)

According to function 4, the predicted ridership is obtained as Figure 3.



From the Figure 3, we can see the ridership of BART birth in 1973 and reach the inflection time in 1988, and will reach the peak in several decades. At the birth stage of BART, there is a difference between the real ridership and the predicted, partly because that BART was so popular at the beginning. Large demand of travel and current advanced technology contributed the early popularity of BART. While, during the operation, problems of BART emerges and policy changes gradually to cope with the market. At the present day, the BART system approaches maturity. Through the scope for innovation and revolution is small, BART still seeks opportunities to improve itself by adopting proper policy continually.

Prior to BART
In the mid-1950s, nearly 20 years before the birth of BART, there are two independent public agencies providing mass transportation service within San Francisco metropolitan area. One is the Alameda-Contra Costa Transit District (short for AC Transit), which was created in 1956. Operating a connected bus service, the AC Transit has been providing local, express, intercity and transBay service for more than a decade. The other public agency is the Bay Area Rapid Transit District, which was created in 1957 and based on fixed rail technology. Both of them became the major transit providers later and their impact lasts several decades till the present day.

Long before the birth of BART, ferry transit has played a significant role in San Francisco Bay for almost 120 years. Began in 1850, the ferry service established its routes between San Francisco and the Oakland Estuary. It reached its peak in the 1930s and became the second busiest transportation terminals in the world. Every year, there are nearly sixty million people and six million vehicles crossed the bay. For a single day, 250,000 persons travelled through the Perry Building to work or other destinations, and 170 landings by ferry. The success of the ferry transit largely owned to the highly connected bus and rail feeder service.

Actually, the ferry transit, together with the bus and rail service, built a transit system during the first decades of 20th century, called a “Key System”. It was built by a privately-owned mass transit company that operated electric railcars, streetcars and ferries which linked ten East Bay cities and San Francisco. The horse-drawn carriages and steam trains as ways to travel between cities in the area was quickly replaced since the system beginning in 1903. It connected to San Francisco by a train line across the Bay Bridge, and a ferry pier that stretched 16,000 feet over the water. The Key System played a significant role on the city sprawl of Bay area. Travel was limited before the emergence of the system, because cars and horses were needed for the town-to-town travel, which few people could afford. Key System filled in the gap and became popular and achieved great success by 1930s. However, the service level of the system went down since late 1940s because of the financial problem, which resulted in largely lost of ridership. Some section of the rail line could only run five to ten miles an hour because of the aging track. Finally, the Alameda and Contra Costa counties took over the system in 1958, which accelerated the foundation of AC Transit.

The birth of BART
Although began to operate in 1972, the implementation of BART had its start in the 1940s, for Bay Area business leaders were becoming concerned with urban sprawl and lack of central business activity that came with the suburban migration. Decentralization moved fast after the Second World War, hence jobs were never concentrated in the city center. Planners of BART faced the challenge of making urban cores stronger while fostering development outside the city center and along the rail lines. BART aimed to bring peak-hour commuters from their suburban homes to their downtown offices, give an alternative to automobile drivers, and most importantly, be financially viable. In addition, because of the growing and sprawling population, traffic congestion was also a main problem that developers of BART attempted to cope with.

Invention of the technology
Borrowed the operation and management experience from the Key System, BART aimed to build a highly connected system covering the Bay Area (although it was largely achieved by late 20th century). AC Transit provided bus service, including transbay service, in the East Bay counties of BART operation. In San Francisco, the MUNI system (San Francisco Municipal Railway) operates street-cars, trolleys, and bus throughout the city. Moreover, the common railroad format and construction know-how was used in the implement of the BART. For financial concern, some funding was secured in 1959. Modern high-tech was also going to be tested on the new system. Combining all the advantages, BART system was ready to take over the market which originally belonged to the Key System.

Early Market
At the beginning operation, the Concord line crossing the Bay to San Francisco had the highest mode shift of the BART system. This is mainly because of the high corridor volumes and the difficulty of parking at central city travel destinations. It seems that people who cared about travel time, cost and location of origins and destinations would have a trend to choose BART system. A survey study in 1978, five years after the operation of BART, shows that there is a large proportion of BART riders between eighteen and thirty-four. According to the age distribution, further step of survey shows that the two major groups who used the BART frequently were downtown employees and college students. For the downtown employees, the reasons for them to choose BART were because of the congestion along the corridor which made their arrival time uncertain and the difficulty on parking in downtown area. BART also offered an alternating transit mode for college students who could not get better served by the campus commuter. While, for the older people who always made short trip, they preferred to choose the traditional transportation mode, like auto or bus, for travel. Hence, this could explain the steady growth of the ridership at the early time of the operation of BART.

Major constrain
BART had not met all the objectives expressed by local offices, community and business organizations in the five years operation. For example, trains on each line run on twelve minute headways rather than the six minute headways originally planned, and only three of four planned lines are now in operation. Moreover, night service has only operated since 1976, Saturday service only begin in Nov. 1977 and Sunday service is not expected until spring 1978 at the earliest.

Four areas of constraint on BART were summarized as characteristics of unreliability and inconvenience, hours of system operation, fare policy and service priorities. For the first constrain, BART failed to provide a reliable time schedule for the passengers at the early time, which brought the problem to those who relied on BART system for commuting. Moreover, the frequency of the feeder connection was low and uncertain. Those problems together indirectly increased the travel time by extending the delay, waiting time and transfer, and also made the arrival time uncertain. For the second constrain, BART could not provide the night and weekend service at the early time. This constrain made BART lost the ridership of the most of car owners who likely to travel during the night and weekend. Their travel pattern might change if BART offered the operation time they demand (it is proved later). For the third constrain, the fare of the BART made the system noncompetitive compared with other mode. For instance, it is much cost-effective for individual interest to drive a car with two to three passengers than taking BART. The off-peak and group discount is not satisfying. For the last constrain, fairness issue rose because of the inequity in station sitting and hours of operation. Some minority and blue collar employees could not be well served or even not be served by the BART system.

Early policy impact
At the early time of the BART system, transportation scholars could find only small impact of the BART on public policy. For the Bay Area public officials, they just took few policy actions because of BART, which was not expected before the operation. However, within the few action, most of were intended to protect against, rather than take advantage of BART. One of the reasons should be the variety of local government and transportation agency. The operation of BART covered nearly all the Bay Area, but each local government made its own policy on BART. For other public organizations of the Bay Area, they intended to use the existing government policy decision making processes. Briefly speaking, the government seemed to cope with the need of the BART system, rather than creating new policies. Hence little policy impact was gained during the early operation of BART, largely because of the lack of central organization for decision making.

However, BART still had some significant impacts on the land use policy and transit finance policy at the early time. According to the final report of BART Impact Program (Apr.1979), many planning studies, zoning changes, and incentives and public improvement projects were initiated because of the BART program. The BART operation led to the zoning change near the lines and stations in some cities. For example, Mission and Rockridge community groups supported lower density development than was projected and succeeded in reducing zoning in the station area. And for San Francisco, zoning allowances were increased in order to encourage development near BART. For finance policy, new form of investment policy was initiated because BART needed permanent public fund for its operation. This urged the state legislature took measures on two things: first, authorizing additional funds for BART, and second, giving the MTC increased authority for monitoring the use of this fund.

The BART operation, like many transpiration experience, formed the policy making, and inversely, policy on BART would also reshaped its operation experience. The interaction of the policy making and BART operation experience at the early time is important. It could form the operation and development pattern for the future. As we see, the early constrain of BART development seek policy solution for better decision making environment, which means a central policy making group with coordinated strategy. On the contrary, policy on BART somehow was going to led BART on the way of transit-oriented development, which has a far-reaching influence even till the present day.

Steady growth of BART
During the development of the operation, BART tried to make effort on attracting ridership. The effort is separated into two ways: one is to improve the serving level, and the other is to follow the step of the Transit-Oriented Development (TOD).

Service level improvement
At the early time, unfixed schedule, non-service at late night and weekend made the group of people who had an alternating travel mode be away from taking BART. Although BART had its own consideration that off-peak service might not cost-effective for the system operation and the existing ridership was satisfying somehow, it could not work as a regional transit system that accepted by the majority of people. BART recognized the problem and made improvement step by step. Policies were made base on the problems rise. At the beginning, the white collar group felt disappointed about the delay and the uncertain transfer and tired to shift the travel mode back to private cars or other regular transit bus. Five years later, the policy on BART helped achieve the goal of fixing schedule and headways with fundamental priority to peak-hour commuting availability. This improvement benefited the people who had the fixed and regular travel demand on peak hours (especially the employees and students). Since BART ensured their concerns (delay, uncertain of headway and transfer), not only the original group back to BART, BART also attracted infrequent travel with a reliable time arrangement. This fast growth of ridership from the 1980s partly reflects the policy impact. Besides, the goal of late night and weekends service is achieved gradually based on the people’s travel demand. And for long trips, policy on the fare structure made the BART more competitive than other transit mode, which benefits those who care about the cost-efficiency of long travel. Aiming at be adjust to the market, improvement on BART are made step by step. It tried to provide a rapid, convenient and comfortable service to passengers by changing its policies, and the public response is reflected by its fast ridership growth.

Trend of TOD
Although BART is just a transit line operated in the Bay Area, it tends to follow the way of transit-oriented development. According to the California Department of Transportation, “Transit-Oriented 	Development (TOD) is moderate to higher-density development, located within easy walk of major transit stop, generating with a mix of residential, employment and shopping opportunities for pedestrians without excluding the auto”. The reasons that people choose a transit mode are various, but mostly based on the convenient, cost-effective and high service level. If a transit system cannot cover all the three standards, it will not build a large operation scales. By the end of 20th century, BART has achieved the last two standards and tried to build a whole transit system that covered the whole Bay Area and accepted by majority of the people. However, as we know, transit system cannot force people to use, but attract and lead people to use. BART knows what people want, and its Board of Directors adopted a Transit-Oriented Development Policy in July 2005. Two key recommendations underpinned the adoption of the new policy are as follows: the first one is to pursue Transit-Oriented Development, not Joint Development. BART should work proactively with cities to plan for development over a larger area around its stations that is both supportive of transit service and maximizes the value of the land. The second is to shift Access Approach. Developers, cities and funding agencies view BART’s application of a 1:1 parking replacement practice as a significant barrier to joint development and TOD. Refining this replacement practice and developing alternative implementation approaches will enhance development opportunities. Increasing ridership is the prior reason for the transit-oriented development, but it may not everything. BART also wants to create physical connection with the communities and provide service to passengers of various travel demand. Accessibility is important. Ridership grows as more and more TOD projects completed. The station area is not a single station anymore, and it becomes the central place where living and working areas around. The mix-use land for housing, working and retail together with the pedestrian friendly transit stations benefit people living around. BART seems to shape or shift people’ s choice of travel mode by providing adaptive, reliable and accessible service that attract more and more ridership during the system growth.

Current BART and the future
Now BART has become a mature transit system with high-technology support in Bay Area. It operates five lines on 104 miles (167 km) of track with 44 stations in four counties. With an average weekday ridership of 379,300 passengers, and 309,420 weekend day passengers, BART is the fifth-busiest heavy rail rapid transit system in the United States. High technology also contributes the success of the transit system. For instance, the automated fare collection system was first used on BART. It helps save the time and make the ticket process much easier. Computerized train control was also first used on the BART operation and it proved work pretty well. This technology eliminates much uncertain during the operation and enhances the safety and accuracy. Other passenger friendly technological improvements are achieved by BART. For example, BART is the first transit system in the US to offer cellular telephone communication to passengers of all major wireless carriers on its trains underground, and it also offers service of “Library-a-Go-Go ”vending machines that give out books. With those technologies, market opens for those who need them.

According to the estimated life cycle of BART ridership, it will reach the peak in a few years. When the ridership comes to the mature stage, it might only grow steadily with the population growth. Besides, BART system nearly occupied all the new transit technology that has less scope to improve much. However, it does not mean that BART is going to wane in a few years. It is still looking for opportunities to expand its system. Not long ago, BART has adopted a System Expansion Policy, and plan to sprawl in several years. However, we cannot estimate the accurate ridership or impact of BART in the future. But, basically, it will not make a remarkable achievement as several decades ago.

= References =