Comprehensive Bus Network Redesigns (2019)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

9 2.1 Overview A national and international literature review was conducted using articles and documents to assess the current state of the industry regarding comprehensive bus network redesigns. The literature review focused on both peer-reviewed and self-published articles in the following topic areas: • Case study research about bus network redesigns, including studies of ridership impacts, customer perceptions, and approaches to network planning. • Industry resources and current overviews of industry trends, including Transit Cooperative Research Program (TCRP) reports and reviews of multiple transit agency redesigns. • Self-published documents from transit agencies that describe their comprehensive bus net- work redesign projects. A full listing of reviewed documents and a brief description of each is provided in Appendix A. Several previous TCRP reports have touched on the concept of comprehensive bus redesign— TCRP Synthesis 10: Bus Route Evaluation Standards (Benn 1995), which discusses the prevalence and contents of documented bus route standards by transit agencies, includ- ing standards for route and bus stop spacing, route deviations and branches, and require- ments for population and employment density. These factors—along with others regarding network connectivity, limitations on the number of passenger transfers, and service area coverage—are all key components of the current era of comprehensive bus network redesigns. Later, TCRP Report 95: Traveler Response to Transportation System Changes Handbook (3rd ed.), Chapter 10, Bus Routing and Coverage (Pratt and Evans 2004), provides an overview of how riders tend to react, on average, to different types of bus routing and coverage changes. The authors discuss many different types of service changes, including “service restructuring,” which is the closest type of change to what this report considers a comprehensive bus network redesign. This synthesis report builds upon these previous ones, adding information from other litera- ture and case studies, adding current relevance to the subject as comprehensive bus network redesigns become ever more common as a way to refresh older bus networks. 2.2 History of Bus Network Redesigns Although there is no definitive history of bus network redesigns, some evidence of them can be found in the literature. For example, TCRP Report 95 (Pratt and Evans 2004) gives many examples of previous bus network “restructurings.” Some of these C H A P T E R 2 Overall Framework: A Review of Existing Research

10 Synthesis on Comprehensive Bus Network Redesigns restructurings would likely constitute network redesigns under this study. The following are examples: • In the late 1970s, Boston, Massachusetts, made significant changes to its bus network by providing more direct service to the Downtown Crossing area and to new rapid transit stations instead of running into downtown. • In 1977, the radial routes serving southeast Portland, Oregon, underwent a major realign- ment, with more continuous north-south and east-west routes and new crosstown service. The changes helped to increase ridership. • In 1992, HARTline in Tampa, Florida, implemented a bus network reorganization resulting in increased ridership. • In 1994, Sacramento Regional Transit District in California executed a significant network restructuring by reinvesting resources from low-productivity routes into higher-performing areas and routes; the changes helped increase ridership. • Between 1994 and 1996, C-TRAN in Vancouver, Washington, transitioned from a low- frequency timed-transfer system to a grid system with improved headways. These changes (along with other factors like improvements in the local economy) helped increase ridership. • In 1995, the Orange County Transportation Authority implemented a network restructur- ing by eliminating unproductive routes, improving headways and headway consistency, and adding neighborhood shuttles and rail feeders. This network redesign broke ridership records while reducing operating costs. • In 1995, SunTran in Albuquerque, New Mexico, restructured its bus system into a grid-like network, increasing ridership. • In 1996, Boise, Idaho, restructured its network into a hybrid hub-and-spoke and grid system, with mixed results. • Also in 1996, King County Metro in Seattle, Washington, implemented an extensive restruc- turing of its bus network. The changes included a hub-and-spoke network emphasis, improve- ments to core routes, and improved off-peak headways, resulting in increased ridership. Although many of the above-described network redesigns were associated with ridership increases, Pratt and Evans (2004) point out that ridership increases were sometimes likely also related to local factors (e.g., improvements in the economy) and other transit changes (e.g., fare reductions or service expansions). In addition, Pratt and Evans point out that several net- work restructurings resulted in ridership losses—again likely due partly to the restructuring itself and partly to other factors; it is very hard to attribute ridership changes directly to the redesign with many other factors always at play. Stanley’s (1998) review of successful ridership-growing initiatives found that service rede- signs (not necessarily complete network redesigns) were part of a package of changes that often resulted in ridership growth. Stanley also found that transit agencies were equipped with detailed marketing and survey data to aid service restructurings, likely contributing to the prevalence of network restructurings in his sample of transit agencies. Between 1996 and 2011, other bus network redesigns were likely implemented; however, there appears to be no definitive source of information that cataloged these efforts. The first signs of the recent “resurgence” of bus network redesigns came in 2011, when StarMetro in Tallahassee, Florida, implemented all route changes for its bus network redesign in one day (Bhattacharya et al. 2014). In 2014, the Jacksonville Transportation Authority (JTA) implemented their net- work redesign, known as the Route Optimization Initiative also enacting all network changes in one day (Jacksonville Transportation Authority 2016). Then, in 2015, Houston METRO followed Jacksonville’s model and implemented all service changes recommended by its System Reimagining Plan in one day (Houston METRO n.d.). Houston METRO’s Reimagined 5-Year Transit Service Plan, which supported the opening of

Overall Framework: A Review of Existing Research 11 three new rail lines in 2013 and 2014, was one of the largest one-day service changes ever enacted by a transit agency in the United States (Houston METRO 2014). Many additional network redesigns were implemented by various transit agencies between 2015 and today, including the following: • In 2015, Omaha Metro implemented the first phase of network changes from its Network Evolution Plan (Jaffe 2015). • Between 2015 and 2017, the Central Ohio Transit Authority (COTA) in Columbus, Ohio, implemented minor changes from its Transit System Redesign, a study that started in 2013. COTA implemented the first phase of major network changes in 2017 (Central Ohio Transit Authority 2016). • In 2017, Halifax Transit in Halifax, Nova Scotia, Canada, implemented initial network changes from its Moving Forward Together Plan, which commenced in 2014 and was approved in 2016 (Halifax Transit 2016a and 2016b). Halifax Transit plans additional, ongoing changes in subsequent phases. • In 2017, Cherriots in Salem, Oregon implemented a redesigned regional bus network, including rebranding the regional bus network from CARTS to Cherriots Regional (Cherriots 2016). • In 2017, the Maryland Transit Administration (MTA) in Baltimore, Maryland, implemented a bus network redesign branded as BaltimoreLink, changing over 60 bus routes in one day (Maryland Transit Administration 2017, 2018). • In early 2018, Dallas Area Rapid Transit (DART) implemented the first service changes that are part of its COA, which includes several network-level changes (DART 2016 and 2017). • In June 2018, the Capital Metropolitan Transit Authority (Capital Metro) in Austin, Texas, implemented changes to more than half of its bus routes and added eight routes as a part of its Connections2025 plan (Capital Metro n.d.). Recently, bus network redesigns were described as the hottest trend in transit (Vock 2017). Vock suggests that transit agencies are making significant changes to their bus networks in response to dropping ridership and to changes in the land-use, demographics, and density of cities. The increase in bus network redesigns may also be partly fueled by the increasing availabil- ity of data and geospatial processing tools that allow for relatively easy analysis and visualization of demographic, land-use, and ridership data. The next sections provide more detail about the results of the literature review, including information about the key objectives and motivations, planning approaches, outcome measures, and engagement strategies used by transit agencies during their bus network redesigns. 2.3 Key Objectives and Motivations There are many possible reasons a transit agency may want to embark on a bus network rede- sign and just as many ways these motivations could be translated into talking points or slogans. According to the literature, the key objective for most bus network redesigns is to improve the transit service for both current and potential riders. For example, the network redesigns in Tallahassee, Halifax, and Columbus all explicitly sought to improve the system’s appeal or “attractiveness” to infrequent and discretionary riders (Bhattacharya et al. 2014, Central Ohio Transit Authority 2016, Halifax Transit 2016a) and to make transit more competitive with private vehicles (Halifax Transit 2016b). The Halifax and Columbus redesigns also explicitly sought to improve the experience of existing riders in abstract terms (Central Ohio Transit Authority 2016, Halifax Transit 2016b), while the Staten Island express bus redesign was spe- cifically implemented to help improve system reliability—a much more measurable concept (NYCT MTA 2017a).

12 Synthesis on Comprehensive Bus Network Redesigns Several redesigns sought to improve riders’ experiences by better matching the transit net- work to both current and forecasted travel patterns being changed by shifting land-use and development (Bhattacharya et al. 2014, Dallas Area Rapid Transit 2016, Houston METRO 2014, Jacksonville Transportation Authority 2016, NYCT MTA 2017a, Vock 2017). Changes in transit networks themselves, for example adding high-capacity transit such as light rail, also led transit agencies to take a comprehensive look at their overall bus network structure. Both Dallas and Houston cited changes in their rail networks as part of the justification for their bus network redesigns (Dallas Area Rapid Transit 2016, Houston METRO 2014). Changes in land-use and development patterns, particularly in re-emerging urban areas, may lead cities to seek improvements in the bus network as a way to support urban sustainability and to reduce congestion (Kalantari et al. 2014, Vock 2017). For example, Omaha’s bus network redesign aimed to support sustainable, active transportation lifestyles in Omaha (Transporta- tion Management & Design 2013), and Currie and Tivendale’s (2010) work in Melbourne, Aus- tralia aimed to build the community and political buy-in needed for funding service expansions to increase ridership. Ridership increases—or counteracting ridership losses—was also part of the motivation behind the bus network redesigns in Houston, Dallas, and Jacksonville (Dallas Area Rapid Tran- sit 2016, Houston METRO 2014, Jacksonville Transportation Authority 2016). Although several bus network redesigns’ first phases were implemented as operating-cost neutral changes (i.e., there was no substantive change in vehicle hours), many bus network redesigns were efforts to curb or control operating costs. For example, the main point of the bus network redesign study in Greensboro, North Carolina, was to improve network efficiency, saving resources to allow for expanding transit coverage without additional cost; due to Board sensitivity to public opposition the proposal was tabled (Boyle and Rey 2012). (Public opposi- tion stemmed from a lack of input early in the planning process, making the public feel like the plan was being “forced” on them.) Other redesigns can be prompted by financial hardship in the transit agency, and redesigning the network is seen as the best way to save resources while still providing the maximum level of service for as many riders as possible, e.g., in Lleida, Spain (Trapote-Barreria et al. 2016). Overall, most bus network redesigns sought to improve the quality and utility of the transit network for the most possible riders—both current and potential. However, the approaches taken by transit agencies to make their networks better were often quite different. 2.4 Approaches to Network Redesigns Although most transit agencies started their bus network redesigns to improve the quality and utility of their transit systems, transit agencies exhibited significant variability in how they undertook their redesigns. This section discusses strategies and tactics across four topics related to undertaking a network redesign: • Planning approaches and network design policies; • Outcome metrics used to evaluate or promote the changed network; • Implementation of the service changes associated with the network redesign; and • Engagement before, during, and after implementation. Planning Approaches and Network Design Policies There are many possible bus network designs for local bus networks, but most can be clas- sified as centralized and radial, hub-and-spoke, grid, or combinations of one or more of these

Overall Framework: A Review of Existing Research 13 types. In addition to the network design, bus networks can also have tiers of bus services, for example, high-frequency core routes, low-frequency coverage routes, express routes, circula- tors, feeders, and even peak-period commuter routes. When transit agencies begin their bus network redesigns, they must decide on the best network design and the most appropriate types of service. Even more foundationally, transit agencies should decide how to approach develop- ing the best network scenario. Is it best to start with a “blank map” and begin redrawing the bus network? Or, is it most important to try to preserve the existing route network to the extent pos- sible? Or should a computer algorithm find the optimal network design? In reality, most agencies end up somewhere on the spectrum between a blank slate and keeping the network as close to the current network and also use a combination of quantitatively driven and qualitatively driven approaches to developing the new service plan. The use of computer algorithms to design optimal transit networks appears to be quite rare within the United States. None of the bus network redesigns at U.S. transit agencies used pure computational approaches to lay out their network designs. However, research in transit net- work optimization through computational models continues to grow and seek application to real-world transit networks. For example, Trapote-Barreira et al. (2016) applied a model devel- oped by Badia et al. (2014) to design the optimal bus network for Lleida, Spain during its eco- nomic downturn. The optimal network was then tweaked and revised based on operational and public feedback. Kalantari et al. (2014) developed an analytical approach to optimize a transit network while keeping the network as similar as possible to the original network, and they propose that this approach is often more palatable to riders and more likely to result in actual implementation of the network changes because the changes are less drastic than would occur if the transit agency started with a blank map. In 2017, Viggiano proposed a new approach to network-level public transportation planning that combines OD-level analysis with new spa- tial aggregation methodologies to develop a framework for identifying corridors for new bus service. The challenge of making big changes has not deterred many transit agencies from taking a blank-map approach to planning their networks or from entertaining complete conversions of their network designs from one type to another. For example, Boyle and Rey (2012) supported Greensboro, North Carolina’s blank-map network redesign in which the network would have been converted from a centralized radial system to a multiple hub-and-spoke network (although the changes were not implemented). Jacksonville, Florida’s bus network redesign also touted its “from scratch” approach (Jacksonville Transportation Authority 2016). Bus network redesigns in Tallahassee, Florida; Jacksonville, Florida; Omaha, Nebraska; and Houston, Texas all involved a conversion from one network type to another. Tallahassee’s centralized radial network was converted to a decentralized, grid-like network to better match Tallahassee’s decentralized pat- tern of development (Bhattacharya et al. 2014). Omaha switched from a complex, interlined radial crosstown network to a grid network geared toward supporting easy movement around the urban core (Transportation Management & Design 2013). Barcelona, Spain’s complicated set of low-frequency lines was converted to a high-frequency transfer-based grid system (Badia et al. 2017). In fact, the creation of high-frequency networks (grid or otherwise) was an emerging theme among many bus network redesigns. Examples include Jacksonville, Houston, Baltimore, and Dallas (Jacksonville Transportation Authority 2016, Houston METRO 2014, Maryland Transit Administration 2017 and 2018, Dallas Area Rapid Transit 2016). However, the choice of network design and service types all depend on the network rede- sign’s goals and objectives. Houston, Halifax, and Columbus all embarked on their redesigns by establishing a governing-body-adopted goal that determined how much of the transit agency’s resources should be focused on high-frequency services and areas and how much should be used to provide coverage to areas with less ridership potential. Although the goal to increase the

14 Synthesis on Comprehensive Bus Network Redesigns proportion of resources allocated toward high-ridership services in Halifax was vague (Halifax Transit 2016a), Houston set its goal at 80% of resources devoted to high ridership routes and 20% for coverage (Houston METRO 2014), and Columbus set its goal at 70% for ridership and 30% for coverage (Central Ohio Transit Authority 2016). These explicit goals helped to guide the design of the bus networks and to prioritize the use of operating and capital resources. Another difference among bus network redesigns was whether or not the service changes were accompanied by other initiatives, for example, bus stop optimizations (i.e., consolidation or balancing), new passenger bus stop amenities and transfer facilities, or improvements in technology and passenger information. Most of the network redesigns had some form of accom- panying initiative, however, the breadth and promotion of these initiatives varied greatly across transit agencies. Bus stop optimizations were part of the network redesigns in Tallahassee, Jacksonville, Baltimore, and Staten Island (Bhattacharya et al. 2014, Jacksonville Transporta- tion Authority 2016, Maryland Transit Administration 2017 and 2018, NYCT MTA 2017b). In Jacksonville, 30% of bus stops, or approximately 1,200 stops, were removed as part of their optimization in an effort to speed up their bus routes. Although many of the redesigns likely had one or more accompanying initiatives, Jacksonville, Halifax, and Baltimore’s (Maryland DOT Maryland Transit Administration [MDOT MTA]) redesigns included the most detail. The following are examples: • Jacksonville increased route supervision and security and safety. • Jacksonville and the MTA both replaced all the bus stop signs in their networks. • Jacksonville launched an improved real-time passenger information system during implementation. • MDOT MTA built new transfer hubs and worked with the local municipality to implement transit signal prioritization and bus-only lanes in key corridors. • MDOT MTA rebranded the entire system with a new name and color scheme, including completely new bus stop signs and vehicle paint scheme. • Halifax Transit also implemented queue jumps, transit lanes, and transit signal prioritization along with other initiatives. Network Scenario Outcome Measures No matter how transit agencies approached their network redesign projects, in order to pro- mote and gain approval for the new networks, transit agencies needed to demonstrate to the public and stakeholders whether the new network was better than the old. For example, in Lleida, Spain, the transit agency developed a balanced scorecard with key performance indica- tors that allowed transit planners and the public to see how different network scenarios per- formed (Trapote-Barreira et al. 2016). Transit agencies also often enacted follow-up studies to see whether the changes resulted in the intended improvements. The following examples highlight some of the outcome measures commonly used by transit agencies before and after implementation of their bus network redesigns. • Transit accessibility scores, which calculate how many places (usually jobs) are accessible by transit and walking within a given period of time (e.g., 45 minutes), were used by Tallahassee, Houston METRO, and Baltimore (Bhattacharya et al. 2014, Houston METRO 2014, Maryland Transit Administration 2017 and 2018). • Rider travel times along a corridor (including walk time to stops) was used by Tallahassee to show that, even with route and stop consolidation, most riders’ trips were faster under the new network (Bhattacharya et al. 2014). A similar analysis was performed by Houston METRO (Houston METRO 2014) and in Baltimore (Maryland Transit Administration 2018). • Boyle and Rey (2012) analyzed the changes in walking distance to transit and the changes in direct transit access to major retailers and to downtown Greensboro, North Carolina.

Overall Framework: A Review of Existing Research 15 • A unique example of a network measure is found in the Staten Island express bus redesign, in which NYCT calculated the reduction in on-street mileage of bus routes, the reduction in turns at intersections, and the increase in non-stop running time as positive outcomes to help improve bus reliability (NYCT 2017). • A unique example of a postimplementation evaluation can be found in Badia et al.’s (2017) analysis of the network change in Barcelona, Spain. They examined ridership trends and transfer rates and saw evidence of increased ridership across the network of routes that out- paced ridership growth related to the addition of or improvements to individual routes. This led them to believe that the network design was successful in accomplishing easier anywhere- to-anywhere travel in the city. Figure 2. Example measurement of access to high-frequency transit (DART), graphic courtesy of DART. Figure 3. Example measurement of access to high-frequency transit, graphic courtesy of MDOT MTA.

16 Synthesis on Comprehensive Bus Network Redesigns Lastly, for those bus network redesigns that implemented high-frequency bus networks, transit agencies typically measured the number of people and destinations (e.g., jobs, hospitals, or schools) that had access to frequent transit routes before and after the network redesign (Central Ohio Transit Authority 2016, Houston METRO 2014, and Maryland Transit Admin- istration 2017 and 2018). See Figures 2 through 5 for examples. Implementation Although redesigning a bus network requires a lot of careful study, analysis, and customer engagement, actually implementing the changes recommended takes at least the same if not more effort. Transit agencies that have implemented their bus network redesigns usually fall into two categories: (1) implementing all the changes at once (usually on one day) or (2) implement- ing the changes in phases. 475,000 788,000 643,000 998,000 Source: 2010 US Census Data; American Community Survey Existing Reimagined Existing Reimagined Frequent Access Frequent AccessW ith in 1/ 2 M ile W ith in 1/ 4 M ile Total METRO Service Area Employment: 1.8 Million +66% +55% Frequent Access: Jobs Figure 4. Example measure of access to high-frequency transit, graphic courtesy of Houston METRO. Figure 5. Example map of extent of high-frequency transit service, graphic courtesy of NYCT.

Overall Framework: A Review of Existing Research 17 Transit agencies identified in the literature that have implemented their service changes in one day are listed in Table 1. The remaining transit agencies that took a phased implementation approach are listed in Table 2. Whether the implementation of planned changes was done in one day or over multiple phases, transit agencies often found the need to implement additional changes or revisions that were not in the original service plan (e.g., Brown et al. 2013, Maryland Transit Administration 2017). In fact, Lleida performed a 3,000-person phone-based survey one year after the initial network change implementation to help understand riders’ concerns about the new network (Trapote- Barreira et al. 2016). Feedback from the survey was used in the “final fine-tuning analysis that could lead to the final changes.” Engagement Transit agencies that embark on bus network redesigns all follow different engagement pro- tocols and timelines and use different tools, depending on their local needs and the stage in the redesign process. However, network redesign projects require extensive engagement, before the project begins, during network planning, and when educating the public on the final, approved Transit Agency City Date Source Metro Transit Madison, WI One day in 1998 Brown et al. (2013) StarMetro Tallahassee, FL July 11, 2011 Brown et al. (2013) Lleida Area Territorial Transport Authoritya Lleida, Spain April 2013 Trapote-Barreira et al. (2016) Jacksonville Transportation Authority Jacksonville, FL December 1, 2014 Jacksonville Transportation Authority (2016) Houston METRO Houston, TX August 16, 2015 Houston METRO (n.d.) Cherriots Salem, OR One day in Spring 2017 Cherriots (2016) Central Ohio Transit Authorityb Columbus, OH May 1, 2017 Central Ohio Transit Authority (2016) Maryland Transit Administration Baltimore, MD June 18, 2017 Maryland Transit Administration (2017 and 2018) Notes: aThe changes in Lleida, Spain were very fast and implemented over a “short period of time;” however, the authors are unsure whether the changes were implemented in exactly one day. bColumbus phased in minor changes over three years; however, they implemented all of the major changes at one time. Table 1. Transit agencies that implemented their bus network redesigns in a single day.

18 Synthesis on Comprehensive Bus Network Redesigns changes. The sheer number of changes—sometimes involving every single route in the system— means that engagement is more complex and more important. This section discusses some highlights from engagement strategies used by transit agencies during network redesigns or recommended by research. The highlights are categorized into three groups: public meetings and focused public engagement, public online engagement, and transit agency in-reach (i.e., engaging transit agency staff). Transit Agency City Phasing Approach Source Fresno Area Express Fresno, CA Continually added decentralized routes to network Brown et al. (2013) Tompkins Consolidated Area Transit Ithaca, NY Gradual consolidation of three separate transit systems Brown et al. (2013) Spokane Transit Authority Spokane, WA Phased improvements, including creating new, crosstown services Brown et al. (2013) Sun Tran Tucson, AZ Added suburban transfer centers and created decentralized routes Brown et al. (2013) Transportes Metropolitanes de Barcelona Barcelona, Spain Changes were phased in once a year and completed in the fall of 2018 Badia et al. (2017) Halifax Transit Halifax, Nova Scotia, Canada 5-year phased changes Halifax Transit (2016a) Omaha Metro Omaha, NE 3 phases; first phase cost neutral, remaining phases require additional investment Transportation Management & Design (2013) Dallas Area Rapid Transit Dallas, TX Will be phased over an uncertain number of years and prioritized as part of the long- range plan Dallas Area Rapid Transit (2016) Table 2. Transit agencies that implemented their bus network redesigns in phases.

Overall Framework: A Review of Existing Research 19 Public Meetings and Focused Public Engagement All bus network redesigns had one engagement strategy in common: all involved the public at least at the point when the network design was finalized for public comments. However, in the case of Greensboro, North Carolina, waiting to engage the public until the network design was done ultimately undermined the implementation of the changes (Boyle and Rey 2012). In most redesigns, transit agencies engaged the public when the study launched, during the transit planning, and finally prior to service change implementation. Here are a few examples: • In Tallahassee, StarMetro held around 100 public meetings during the 2 years leading up to the service changes (Bhattacharya et al. 2014). • Several transit agencies reported creating and holding regular working sessions with a spe- cialized stakeholder committee of local leaders and representatives of key constituent groups (e.g., Central Ohio Transit Authority 2016, Halifax Transit 2016a). • Meetings with community organizations, civic associations, groups representing seniors or people with disabilities, and other special-interest groups were also very common methods for engaging the public (Central Ohio Transit Authority 2016, NYCT 2017, Maryland Transit Administration 2017 and 2018). • Several transit agencies also conducted surveys of riders prior to forming bus network rec- ommendations (Central Ohio Transit Authority 2016, Halifax Transit 2016a, Jacksonville Transportation Authority 2016). Some surveys were conducted at bus stops (NYCT 2017) and onboard buses or at transit facilities (Halifax Transit 2016a, NYCT 2017), while other surveys were conducted online (Halifax Transit 2016a). • Another common type of engagement was to meet customers where they were. These events, sometimes called “pop-up” events were used in Halifax, Omaha, Jacksonville, Baltimore, and Salem (Halifax Transit 2016a, Transportation Management & Design 2013, Jacksonville Trans- portation Authority 2016, Maryland Transit Administration 2017 and 2018, Cherriots 2016). – The Maryland Transit Administration used a retired articulated bus as an “InfoBus,” which it stationed at various locations throughout the city as a traveling public meeting during the planning phases. The bus was also used in-service on a variety of current routes and was staffed by people knowledgeable in the new network, and offered free rides in addition to important redesign information. – Representatives from Cherriots in Salem, Oregon attended community events like markets and fairs in addition to holding traditional public meetings and staffing an information table at a downtown transit center. • NYCT also had a unique example of public engagement, in which it co-hosted, along with a local university and non-governmental organization, a hackathon for its Staten Island express bus reimagining (NYCT 2017a). The hackathon had 150 attendees who worked with NYCT archived GPS data, passenger origin-destination tables, passenger boarding and alighting data, and NYCT’s general transit feed specification (GTFS) data. • Due to the large number of Baltimore City school students that utilize public transit to travel to and from school, the Maryland Transit Administration developed a specialized outreach campaign for students and their families (Maryland Transit Administration 2017 and 2018). • Lastly, Currie and Tivendale (2010) describe in detail the inclusive design process used during a bus network review in Melbourne, Australia. The process used both data analysis and com- munity and operator views to inform network option development and avoided extensive bus network demand modeling. Online Engagement In addition to in-person, face-to-face engagement, transit agencies also used online and web- based tools. Many transit agencies likely used standard tools like a webpage on the transit agency’s

20 Synthesis on Comprehensive Bus Network Redesigns website; however, only a few transit agencies specifically featured their online engagement in the available literature. For example, NYCT accepted public feedback through its regular online customer feedback system and through a special email address set up for the study (NYCT 2017). Halifax Transit used an online engagement tool at MakeTransitBetter.ca and also held Twitter town halls at which people could ask transit agency staff questions or submit comments about the network redesign (Halifax Transit 2016a). Dallas Area Rapid Transit used an online platform specifically designed for public engage- ment in local government projects called MindMixer (now MySidewalk) (Dallas Area Rapid Transit 2016). Social media was also generally used by most transit agencies (Central Ohio Transit Authority 2016 and Maryland Transit Administration 2017 and 2018). Transit Agency In-Reach Because bus network redesigns impact almost all parts of a transit agency’s internal staff, tran- sit agencies may also need to develop strategies to educate staff and to generate buy-in. One of the key practices highlighted in TransitCenter’s Redesign Workshop proceedings is to make the most of drivers’ expertise and daily contact with riders through focus groups and ride-alongs. Although most transit agencies likely had some form of in-reach, only a few transit agencies highlighted their in-reach efforts in the available documentation. Staff who have significant involvement in daily operations and customer interactions may often glean important insights about what needs to be changed in a bus network and also have ideas about how customers may react to proposed changes. The network redesigns in Jackson- ville, Columbus, and Dallas all used feedback from bus operators, supervisors, or customer service representatives obtained through town halls, interviews, or other types of forums (Cen- tral Ohio Transit Authority 2016, Dallas Area Rapid Transit 2016, Jacksonville Transportation Authority 2016, Maryland Transit Administration 2017 and 2018). The Maryland Transit Admin- istration took very explicit steps for bus operator in-reach, developing and implementing a detailed and robust input process at the bus divisions for the year before network implementa- tion and surveying bus operators after implementation to understand their experiences and the public perception of the service changes. 2.5 Common Practices and Lessons Learned Transit agencies offered many insights and lessons learned that would benefit other transit agencies during their contemplation of, planning for, and implementation of a bus network redesign. The lessons learned are grouped into three areas: engagement, network design, and implementation. Engagement Lessons Learned Transit agencies offered many lessons learned regarding engagement and working with the public during bus network redesigns. From the TransitCenters Untangling Transit: Bus Network Redesign Workshop Proceedings, comprising input from more than 30 agencies, key suggestions are to frame the problem and case for redesign from the outset and to ask for input before com- ing to riders with a proposed plan (2017). These approaches make the process transparent and engender trust and engagement by the public. The agencies also suggested using data and metrics to support recommendations and counter anecdotes, as well as using advocates and local stake- holders who support the plan as voices for the agency to leverage the trust already established by these other parties. Two very interesting studies on the bus network change in Tallahassee, Florida, found that riders from different communities impacted by the network changes did not perceive the

Overall Framework: A Review of Existing Research 21 network changes the same way (Bhattacharya et al. 2014, Brown et al. 2013). The studies exam- ined the perceptions of the network changes of riders in two different communities (a student community and a low-income community). Riders’ perceptions of the network changes did not match the analyzed results of the network changes. For example, although both communities actually had higher levels of transit accessibility (i.e., access to more destinations in fewer minutes) and lower bus travel times, riders in the low-income community perceived the network changes as negative, while riders in the student community viewed them as positive. The negative per- ceptions in the low-income community actually exhibited themselves in reduced transit use. Bhattacharya et al. (2014) reported that the perception in the low-income community was that their concerns were not addressed or that the outreach had been inadequate. These studies lend evidence to the mantra that perception is reality when it comes to network changes and suggest that transit agencies should enact strategies for meaningful public engagement in order to make sure that concerns are being heard, acknowledged, and addressed, either by modifications to the proposed plans or by a clear explanation as to why plans are not modified, to ensure that trust is earned in the planning process and ridership is not negatively impacted. Brown et al. (2013) recommend early and frequent engagement with people with disabilities and seniors and their related advocacy and special interest groups in order to actively work to address their concerns. Brown et al. reminds readers that engagement needs to be early, com- prehensive, and exhaustive and that public participation has to be actively encouraged. In fact, Boyle and Rey (2012) cite a lack of early public engagement as a significant obstacle to imple- menting the bus network redesign in Greensboro, North Carolina. In their case, the public did not know the agency was working on a network redesign until the agency shared the proposed network design at public meetings. The public felt that the network was being forced on them, and riders did not understand the need for a “blank slate” approach—why erase everything and start all over when the network seems to be working for many riders’ trips (Boyle and Rey 2012)? On the other hand, Currie and Tivendale’s (2010) inclusive planning process helped to increase stakeholder buy-in and even resulted in positive upward pressure through the local political structures for increased funding to support the transit improvements recommended by the planning process. Currie and Tivendale recommend that leaders organize bus network qualities and service characteristics into an easy-to-understand hierarchy to facilitate discussions with the public. Trapote-Barreira et al. (2016) recommend that transit agencies apply best practices and the principles of change management to a network redesign. Stakeholders must possess a sense of urgency and understand why change is needed. The transit agency should clearly communicate the change and work to produce small wins to generate confidence among stakeholders. Trapote-Barreira et al. (2016) and Jacksonville Transportation Authority (2016) recommend using performance measures and well-defined goals and objectives during the planning and implementation phases of a bus network redesign. Emphasizing goals, objectives, and measures can help to focus conversation, maintain support and understanding from public officials, and reduce the likelihood that pushback will result in substantive changes to the optimal network design. Jacksonville also encourages transit agencies to clearly define the goals and objectives and what their service implications are at the outset of the project—being completely transparent about the objectives and their likely real-world impacts on riders will help ensure the continued support of governing boards and public officials. Network Structure Lessons Learned The proceedings from TransitCenter’s 2017 Bus Network Redesign Workshop suggest that a key element to designing the right network structure is to establish why the agency is doing a

22 Synthesis on Comprehensive Bus Network Redesigns redesign before they start—such as whether such a huge undertaking is needed, or whether the goal is to reduce costs or increase ridership. Although many of the transit agencies that converted their centralized radial systems into grid or hub-and-spoke networks saw ridership increases after implementation (Pratt and Evans 2004), not all transit systems did see increases in rider- ship. Decentralized bus networks may work in some cases, but not in all, depending largely on local population and land-use patterns. Thompson and Matoff (2003) found that decentralized regions with multiple strong origins and destinations were better served by multi-destinational (i.e., decentralized) transit networks rather than by radial, centralized networks. Brown et al. (2013) also caution that grid networks only work well when the routes in the grid have high- frequencies to facilitate easy transfer between routes. Additionally, transit agencies implement- ing network restructurings that might force passenger transfers or longer walks to bus stops may also need to carefully consider the infrastructure required to facilitate a high-quality pedestrian experience (e.g., sidewalks, shelters, cross-walks, and protected street crossing signals) (Brown et al. 2013). Implementation Lessons Learned There are many lessons learned from transit agencies that have implemented bus network redesigns. The agencies that participated in TransitCenter’s Bus Network Redesign Workshop (Transit Center 2017) suggested • Organizing a task force to coordinate all internal and external communications. • Providing the public, stakeholders, and bus operators with clear messaging about why the redesign is happening. • Arming ambassadors, bus operators, and the public with information to make the transition as painless as possible. • Providing free rides during an initial period to ease the pain of the transition. • Preparing tactical teams to address the inevitable issues that will crop up with new bus routes and stops. The only document reviewed that provided pointers and lessons learned about implementing a transit network redesign came from the Jacksonville Transportation Authority (Jacksonville Transportation Authority 2016). Selected recommendations include • Implementing all the changes at once is important to project success—multiple phases may delay implementation and create additional rider confusion. • Including a full portfolio of improvements to wrap around and augment the service changes— replace bus stop signs, add shelters, improve customer service, implement new technolo- gies, etc. • Holding post-launch operator workshops to help develop system refinements quickly and address safety issues and customer concerns. Later chapters in this report will provide a much more extensive discussion of lessons learned based on results from the transit agency surveys and case examples and address implementation ideas on a variety of topics. 2.6 Summary The literature review covers bus network redesign case studies, articles on theories of network redesign approaches, bus route and network planning industry resources, and self-published network redesign documents from transit agencies. Although bus network redesigns are not solely a recent phenomenon, they certainly have become more prevalent and more publicized

Overall Framework: A Review of Existing Research 23 within the last decade, likely a result of increased access to data and a broader approach to com- munication and information sharing than ever before. Bus network redesigns can range in scale from small to large; they can be implemented in stages or all at once; and they may require fewer, the same, or more operating and capital resources. However, several themes emerged as constants from the literature review: • Bus network redesigns are usually conducted to improve the quality and utility of transit service by better meeting the travel needs of both current and potential riders. They are a mix of moving from one type of service provision to another and taking a blank-map approach to developing a new network. • Measurement and quantification of anticipated and actual improvements from bus network redesigns can be a key tool in obtaining buy-in for the plan and making decisions between different network scenarios. • Bus network redesigns require extensive public and external outreach at project inception, during network planning, and prior to implementation. Continued outreach after implemen- tation was also highly recommended. • Bus network redesigns can experience significant public pushback, but this can be mitigated by involving stakeholders early; having a strong public champion; setting clear goals, objec- tives, and measures; and clearly communicating the need for the change in terms stake- holders can identify with. • In-reach to transit agency staff is just as important as outreach to the public. Generating input, excitement, and buy-in among staff, especially those on the front-line, is critical to success.