Accessibility analysis by public transport in Ulaanbaatar, Mongolia Technical Memo June 2024 ACKNOWLEDGMENT The development of this report was made possible through the funding provided by the Quality Infrastructure Investment (QII) Partnership Trust Fund. The World Bank Group and the government of Japan established the QII Partnership in 2016 to raise awareness and scale up attention to the Quality Infrastructure Investment Principles endorsed by the G20. These include maximizing the positive impact of infrastructure, raising economic efficiency in view of life-cycle cost, integrating environmental and social considerations, building resilience against natural disasters, and strengthening infrastructure governance. The QII Partnership accomplishes this goal by providing grant support to prepare and implement infrastructure projects in developing countries, conduct analyses, and disseminate knowledge about the application of the QII Principles. For additional details on these partnerships, please visit their respective websites at https://www.worldbank.org/en/programs/quality-infrastructure-investment-partnership The report's preparation was undertaken by the consultant Tainá A. Bittencourt, under the commission of the World Bank. Khaliun Bat-Orig (Transport Analyst, Co-TTL) was the task lead of the activity, providing technical guidance, coordinating client consultations, and facilitating capacity-building activities. The team comprised Joanna Moody (Transport Specialist), Ok Stella Namkung (Transport Consultant), and Clemens Portenlaenger (Sr. Transport Specialist), with special acknowledgments to Angar Enkhtur for her administrative support throughout the activity's execution. This analytical work is an integral component of the "Mongolia Transport Development Support" Programmatic Advisory Services and Analytics (PASA), conducted by the World Bank. It serves as a complement to other studies, guidelines, and technical notes produced under the PASA, with the goal of enhancing the technical expertise of the Municipality of Ulaanbaatar (MUB) and ensuring the high-quality execution of the Ulaanbaatar Sustainable Urban Transport Project (USUTP). This endeavor aims to foster more sustainable, resilient, safe, and inclusive urban transport within Ulaanbaatar. The report is a product of the collaborative efforts with the UB city government, particularly the Project Management Office of the USUTP. The authors extend their gratitude to Bulgaa Khurelbaatar, the coordinator of USUTP, and USUTP staff. STANDARD DISCLAIMER This volume is a product of the staff of the International Bank for Reconstruction and Development/The World Bank. The findings, interpretations, and conclusions expressed in this paper do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. COPYRIGHT STATEMENT The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development/The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. Please cite the work as follows: World Bank. Accessibility Analysis by Public Transport in Ulaanbaatar Technical Memo. 2024. Washington, DC. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA, telephone 978-750-8400, fax 978-750-4470, http://www.copyright.com/. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA, fax 202-522-2422, e-mail pubrights@worldbank.org. REPLICATION The procedures undertaken in the analyses presented in this technical note, such as the creation of the General Transit Feed Specification (GTFS), data and indicator processing, and the development of maps and graphs, are thoroughly detailed in the step-by-step replication guide included in the annex. 1 CONTEXT Ulaanbaatar is the capital of Mongolia, with 1.64 million inhabitants 1 unevenly distributed over 470.4 km², concentrating half of the country’s total population 2. This reflects the city’s rapid urbanization, ranging from 2-4 percent per year in recent decades, with a large influx of rural migrants seeking better opportunities. This urban growth has strained the city's infrastructure and services, leading to challenges such as housing shortages, traffic congestion, and air pollution. This technical note analyzes the accessibility to jobs, schools, and hospitals by public transportation in Ulaanbaatar. It provides insights into further actions to improve the population's access to transportation and essential opportunities and services. LAND USE AND OCCUPATION The spatial grid used in this technical note consists of regular hexagons measuring 500 meters in diameter, each covering an area of 25.25 hectares. Travel and employment data by traffic analysis zone (TAZ) are sourced from the 2022 Household Travel Survey3, while data on schools and hospitals are obtained from the Capital City Road Development Agency (RDA). Population data was obtained from WorldPop (2020), providing more granular information on where people live in the city with a regular grid of 100 meters, which is significantly smaller than the available travel data by TAZ 4. Ulaanbaatar has a very contrasting population density landscape. Most of the population lives near the city center in vertically oriented buildings concentrated alongside Peace Avenue (parallel to the railway) and other large roads and highways that pass through the city (Figure 1). On the other hand, the areas further from the major roadways consist of low-density horizontal households with an irregular and broken street network. The spatial distribution of job opportunities and schools follows the same pattern, with a higher concentration within a 2 km radius from the railway in the east-west axis. Hospitals, on the other hand, are less numerous and mostly concentrated in the city center. 1 National Statistics Office, Population as of 2023. 2 National Statistics Office, Total population as of 2023 was 3.396,788 3 Governor’s Office of the Capital City, Road Development Agency of the Capital City, and Think Tank, 2022, “15,000 household survey for OD matrix in Ulaanbaatar city” report. 4 Although population by khoroos is available, WorldPop is preferred in this case as it provides a more detailed understanding of the population's spatial distribution, with 9,288 spatial units compared to the 171 units in Khoroos. Additionally, WorldPop's data is based on the 2020 Mongolian Census and satellite imagery, ensuring greater accuracy and relevance. 2 Figure 1. Spatial distribution of the population (top left), jobs (top right), school vacancies (bottom left), and hospitals (bottom right) in Ulaanbaatar. Source: Data from WorldPop (2020) and the Household Travel Survey (2022). We were unable to find updated disaggregated data on income in an editable format. Nonetheless, a previous study on poverty in Ulaanbaatar, conducted by the World Bank, estimated household income and multidimensional poverty by small territorial units, as shown in Figure 2. The available data shows that the railway line strongly influences the concentration of poverty. Large low-income areas are found in the Songinokhairkhan district, in Khan-Ul near the railway extension in the south, and in the east of the Bayanzurkh district. 3 Figure 2. Multidimensional poverty clustering map. Source: (World Bank, 2017). Urban Poverty in Ulaanbaatar: Understanding the Dimensions and Addressing the Challenges. ACCESS TO PUBLIC TRANSPORT Access to public transportation involves multiple individual, social, economic, and spatial factors. In this report, we refer to the proximity of residential or commercial areas to a bus stop, with the expectation that individuals can easily reach the bus stop within a short walking distance. This concept is relevant in urban planning and transportation studies as it indicates the extent to which people have convenient access to public transit services, which can influence their travel behavior and overall mobility patterns 5. In many urban areas, 300 meters is considered a reasonable walking distance for most individuals to reach a bus stop from their residence, typically representing a 5-minute walk. Providing access to public transportation within this radius aims to encourage transit use, reduce reliance on private vehicles, alleviate traffic congestion, and promote sustainable modes of transportation. It helps identify areas with adequate transit access as well as gaps where improvements may be needed to enhance mobility options and promote equitable access to transportation resources. In Ulaanbaatar, most bus lines are radial, passing through the main roads and connecting different neighborhoods to the city center. Consequently, people living far from the main roads may face 5Schwanen, T., & Mokhtarian, P. L. (2005). What Affects Commute Mode Choice: Neighborhood Physical Structure or Preferences Toward Neighborhoods? Journal of Transport Geography, 13(1), 83–99. 4 difficulties accessing public transportation services. As depicted in Figure 3, around 54.4% of Ulaanbaatar’s residents live 300 meters or less from bus stops, while 78.7% have a bus stop within a 500-meter radius of their homes. This means that a large share of the population must walk a considerable distance to access public transport services, increasing transit travel times and reducing comfort while traveling. This is particularly problematic during the winter and at night, when the cold and darkness may discourage or prevent people from walking. Figure 3. People more distant than 300 (left) and 500 meters (right) of a bus stop. Compared to other cities with similar population sizes, the percentage of people with long walking distances to bus stops in Ulaanbaatar is large. In Porto Alegre, Brazil, 93% of the city’s residents live within 300 meters of bus stops 6. In Zhongshan, in China, 93.8% of people in the central city and 83.8% in suburban areas live within 500 meters of a bus stop 7. From Figure 3, we can highlight some areas worthy of attention. The first one is in the north of the city, near Yargait, Yargaitiin Bogino Ovoo, and Sharga Morit (between the Chingeltei and Sukhbaatar districts). Although people live in those neighborhoods, they would have to walk up to 2 kilometers to reach a bus stop on Sanzai Road. Another specific area is in the southwest, near the 10th and 12th khoroos in the Khan-Uul district. Despite the high population density, the distance to bus stops can reach up to 2.5 kilometers. This radial structure of Ulaanbaatar’s public transport system is also evident in the bus frequency maps in Figure 4. Bus lines on Peace Avenue (east-west) and E. Galdanboshgot Street (north) have lower headways, reaching less than 3 minutes during peak hours. At night, public transportation service drops significantly, with fewer line options and lower frequency. Figure 4. Headway of transit lines at 8 am and at 9 pm in Ulaanbaatar. 6 AcessoCidades - FNP (2022). Diagnóstico quantitativo de acessibilidade e mobilidade urbana com enfoque de classe, raça e gênero. Available at: https://www.redus.org.br/acesso-cidades/biblioteca/pasta/d5d7ff5a-8d46-463e-8588- 795a023f9968 7 Hao, Z.; Peng, Y (2023). Comparing Nonlinear and Threshold Effects of Bus Stop Proximity on Transit Use and Carbon Emissions in Developing Cities. Land, 12, 28. 5 Due to this configuration and the irregular topology of the street network, there are few peripheral bus lines, making traveling between neighborhoods more difficult. This contributes to the monocentric structure of the city and the extensive use of private cars, especially for non- mandatory trips. According to Household Travel Survey (2022) data, most trips are made by car, which aligns with the high car ownership rates: 76% of all households own a private car. On the other hand, although 72.8% of all interviewees own a public transportation card, only a small share of all trips is made by public transport. Even khoroos with a large percentage of poor households have higher shares of car trips than public transport trips, as depicted in Figure 5. 6 Figure 5. Spatial distribution of transit (left) and car (right) trips Source: Data from the Household Travel Survey (2022). ACCESSIBILITY Urban accessibility refers to the ease with which people can reach desired destinations within a city or region, typically using various modes of transportation such as walking, cycling, driving, or public transit 8. It encompasses the factors of transportation infrastructure and services and the location of residences and opportunities, such as jobs, education, healthcare facilities, etc. Metrics of urban accessibility can vary depending on the specific context and goals of a study or analysis. However, one of the most common metrics is location-based cumulative accessibility. Cumulative accessibility refers to the total level of accessibility to various destinations accumulated within a given spatial area, typically measured over a specific time period. It takes into account the accessibility of multiple destinations rather than focusing on individual locations or trips 9. In sum, it calculates the sum of all opportunities reached by residents of an area within a transit ride shorter than a specific time threshold (60 minutes, for instance) divided by the number of all opportunities available in the city. This concept is often used in urban and transportation planning to evaluate the overall accessibility experienced by residents within a city or region. To assess the accessibility of Ulaanbaatar’s residents, we calculated the percentage of accessible opportunities (jobs, school vacancies, and hospitals) within 20, 30, 45, 60, and 90 minutes by transit, factoring in access, in-vehicle, and egress time. A General Transit Feed Specification (GTFS) 10 was built based on data provided by the public transportation authority in April 2024. In total, four travel time matrices were created, with one 8 Levinson, D., & Krizek, K. J. (2008). Accessibility and the Journey to Work. Journal of Transport Geography, 16(3), 191–203. 9 Geurs, K. T., & van Wee, B. (2004). Accessibility Evaluation of Land-Use and Transport Strategies: Review and Research Directions. Journal of Transport Geography, 12(2), 127–140. 10 The General Transit Feed Specification (GTFS) defines a common format for public transportation schedules and associated geographic information. A GTFS feed is composed of a series of text files collected in a ZIP file. Each file 7 matrix generated every 15 minutes from 8h to 8h45 am. The consolidated travel matrices were derived from the median of the four travel time matrices generated during the morning peak. Maps illustrating access to jobs, schools, and hospitals, as shown in Figure 6, exhibit significant similarities. This is because opportunities are concentrated in the city center, and accessibility is particularly influenced by the availability and connectivity of public transport systems. Since transit services primarily operate along main roads and avenues, accessibility is higher for residents living near these areas and lower for those residing in more distant regions characterized by lower density, rugged topography, and an irregular street network. models a particular aspect of transit information: stops, routes, trips, and other schedule data. The details of each file are defined in the GTFS reference. 8 Figure 6. Accessibility to jobs (top), school places (center), and hospitals (bottom) by transit considering a 30 (left), and 60 (right) threshold. Table 1 shows that half of the population can access half of the jobs within a 60-minute public transport trip, but only 13.6% access ¾ of the total jobs within this same time threshold. Schools, being more spatially distributed than jobs, allow a higher share of the population to access half of 9 the schools’ capacity within 60 minutes. In contrast, hospitals are more concentrated, requiring people to travel to the city center to access them. Table 1. Percentage of the population with access to at least 25%, 50%, 75%, and 100% of opportunities within 60 minutes by transit. Jobs Schools Hospitals 25% 885,190 (71.0%) 910,094 (73.0%) 937,833 (75.2%) 50% 649,703 (52.1%) 757,560 (60.7%) 783,880 (62.8%) 75% 169,976 (13.6%) 490.078 (39.3%) 517,026 (41.4%) 100% 0 0 0 Compared to other cities with similar population and area, such as Curitiba (1.7 million inhabitants and 432 km²) and Fortaleza (2.4 million inhabitants and 314 km²) in Brazil, Ulaanbaatar presents lower levels of accessibility. 52% of Ulaanbaatar's residents have access to 50% of job opportunities within 60 minutes by transit, compared to 78% in Curitiba and 81% in Fortaleza. Within 60 minutes, almost 14% of the Ulaanbaatar population have access to 75% of jobs, compared to 39% in Curitiba and 52% in Fortaleza (see Appendix 7 for further details). The areas with the lowest levels of accessibility are shown in Figure 7, considering the populated hexagons in the bottom 10% (Figure 7, left) and the bottom 25% (Figure 7, right) for access to job, school, and healthcare opportunities. Priority areas are concentrated in the 19th khoroo of the Chingeltei and Sukhbaatar districts, near the Selbe river; in the 20th khoroo of Sukhbaatar, near Belkh Street; in the 20th khoroo of the Songinokhairkhan district; and in the 23rd, 28th, and 20th khoroos of the Bayanzurkh district. 10 Figure 7. Population in the bottom 10% (left) and 25% (right) areas of access to job, school and hospital opportunities. The temporal analysis of planned accessibility, illustrating its variation throughout the day in Figure 8, shows that accessibility increases during peak hours (morning, lunchtime, and evening) due to the higher frequency of transit services. In off-peak periods, accessibility experiences a modest decrease, mainly due to longer waiting times. Figure 8. Average 60-minute job accessibility by transit at different times of the day. Since this analysis is based on scheduled transit times, it does not account for delays due to traffic and congestion, which are greater during peak hours. Using the average speed of bus lines from operational data, we recreated the GTFS and re-ran the accessibility analysis for the morning peak 11 at 8 am. Table 2 shows that the average 60-minute job accessibility drops from 44.85% to 16.98% when the operational data is considered. Table 2. Average job accessibility at 8 am considering planned timetables and the operational speed. Time threshold Planned timetables Operational speed 20 minutes 1.21% 0.81% 30 minutes 5.68% 2.41% 45 minutes 21.88% 7.80% 60 minutes 44.85% 16.98% 90 minutes 75.98% 41.98% Ideally, everyone should have access to transportation and, through transportation, to urban opportunities in a reasonable amount of time and cost. However, in the context of limited resources and the need for prioritization, a few indicators can be considered. One approach is to maximize the number of people who benefit from transport investments (Figure 9, left). In this case, new transport services should aim to improve accessibility in areas with high population density and low current accessibility (light pink). Specifically, areas such as the 8th, 23rd, and 24th khoroos in the Khan-Uul district in the south could be highlighted. Another approach is to minimize socio-spatial inequalities and enhance transit accessibility for those who rely heavily on public transportation (Figure 9, right). In the absence of more robust indicators, priority may be given to areas with a high proportion of public transport users and low current accessibility (light pink), predominantly located in more distant zones in the south and north. 12 Figure 9. Bivariate LISA Maps 11 of population and 60-minute job accessibility (left) and share of public transport trips and 60-minute job accessibility. 11 The local Moran statistic, porposed by Anselin (1995) in "Local Indicators of Spatial Association — LISA" (Geographical Analysis, 27), is used to identify local clusters and spatial outliers. This method calculates the correlation between a variable in a specific spatial unit and another variable in its neighbors. More details on this statistic are available at GeoDa Center. 13 CONCLUSIONS AND RECOMMENDATIONS Urban Structure and Public Transport Accessibility ● Ulaanbaatar's urban structure exhibits notable disparities in access to public transport and essential opportunities such as jobs, schools, and hospitals. The uneven development of the city highlights the need for immediate action to enhance public transport coverage, especially in low-density areas beyond the expanded city center. ● An urgent priority is to expand the coverage and reach of public transportation into lower-density areas outside the expanded city center. Public Transport Solutions ● To improve transport coverage, one proposed solution is the implementation of fixed transit routes operated by smaller vehicles such as vans or microbuses. ● Additionally, Demand Responsive Transit (DRT) can provide flexible and efficient transportation tailored to user requests. o There are various DRT models globally12, with an interesting example being Barcelona. In areas challenging to access with large buses, Barcelona employs microbuses and vans to maintain service frequency. In more remote areas, Barcelona's DRT service, El meu bus 13, utilizes fixed stops but generates routes based on user requests, optimizing travel efficiency (see Appendix 4 and Appendix 5 for details). First and Last Mile Connectivity ● Enhancing the first and last mile of transit trips is crucial, which involves improving sidewalks and integrating public transport with bicycle infrastructure. o Examples from Omaha and Grenoble, illustrated in Appendix 6, showcase how intermodal transport hubs can improve connectivity and convenience for commuters. Perimetral Bus Lines and Connectivity ● Exploring the viability of perimetral bus lines can enhance the connectivity of Ulaanbaatar's public transport system. The current radial structure connects neighborhoods to the city center but lacks efficient connections between different neighborhoods, exacerbated by the city's fragmented street network and rugged topography. 12 See an interesting analysis of some existing solutions in Perez et al (2021) Demand Responsive Transit Understanding Emerging Solutions, a study conducted by WRI Mexico and available at https://es.wri.org/publicaciones/demand-responsive-transit. 13 More information about the on-demand service in the metropolitan area of Barcelona in https://www.tmb.cat/en/barcelona/bus-on-demand#informacio 14 Modal Shift to Sustainable Transport ● Encouraging a shift from private car usage to more sustainable transport modes is also crucial. Policies should incorporate both incentives and disincentives, such as parking restrictions and improvements in public transport quality. Despite being unpopular, awareness campaigns and infrastructure upgrades are pivotal for successful implementation. o Addressing the primary reasons people prefer private cars is essential. According to the 2022 Household Travel Survey, convenience and shorter travel times are among the main factors driving car preference. ● Effective transport policies should also prioritize improving accessibility for underserved areas over introducing new systems in already accessible regions. Prioritizing enhancements for those with limited access to transport and opportunities ensures equitable development. 15 APPENDIX A1. Accessibility to jobs by transit considering a 20- (top left), 30- (top right), 45- (bottom left) and 90-minute threshold (bottom right). 16 A2. Accessibility to school places by transit considering a 20- (top left), 30- (top right), 45- (bottom left) and 90-minute threshold (bottom right). 17 A3. Accessibility to hospitals by transit considering a 20- (top left), 30- (top right), 45- (bottom left) and 90-minute threshold (bottom right). 18 A4. Example of bus line 116, in Barcelona, which runs from a hill near Parc Güell, and it is operated by small buses. A5. Service areas of the “El meu bus” in the metropolitan area of Barcelona. A6. Interchange transport hubs with bike sharing stations in Omaha, United States (left), and with bike parking stations in Grenoble, France (right). 19 20 A7. Accessibility comparison between Ulaanbaatar and other cities. Ulaanbaatar Fortaleza Curitiba Rio de Janeiro Mongolia Brazil Brazil Brazil Population 1.5 mi 2.4 mi 1.7 mi 6.3 mi Area 470 km² 314 km² 432 km² 1,200 km² 2,332,393 1,599,725 2,824,113 25% 885,190 (71.0%) (95.6%) (91.8%) (45.2%) 1,983,273 1,361,212 1,458,636 50% 649,703 (52.1%) (81.3%) (78.1%) (23.3%) 1,276,344 75% 169,976 (13.6%) 675,204 (38.8%) 0 (52.3%) 21