Towards True Multimodal Transportation Accessibility: Data, Measures, and Methods

Transcription

1 Towards True Multimodal Transportation Accessibility: Data, Measures, and Methods Ivana Tasic* Graduate Research Assistant Department of Civil and Environmental Engineering, University of Utah 0 Central Campus Dr., Rm. 0 Salt Lake City, Utah Phone: (0) - ivanat@trafficlab.utah.edu *Corresponding Author Claire Bozic Senior Analyst Chicago Metropolitan Agency for Planning South Wacker Drive, Suite 00 Chicago IL 00 Phone: () - cbozic@cmap.illinois.gov Eric Hanss Principal Pedestrian Planner, Active Transportation Alliance Consultant, CDOT Division of Project Development 0 N. LaSalle, Ste 00 Chicago IL 00 Phone: () - eric.hanss@activetrans.org Word Count: + ( Figures) = Prepared for the Transportation Research Board Annual Meeting 0 Re-Submission Date: November, 0

2 Tasic, Bozic, and Hanss 0 0 ABSTRACT The measure of accessibility becomes particularly relevant in the era when applications and user information based on real-time data feed become factors of major influence on traveling choices and behavior. From the likelihood of choosing alternate modes of transportation, to travel time reliability, and the overall ability to reach desired destinations, transportation accessibility is the measure that truly represents the success of complex multimodal transportation systems from both users and transportation practitioners perspective. This paper presents spatio-temporal measures of accessibility for pedestrians, bicyclists, and transit users, using the case study of the City of Chicago. The paper summarizes the current challenges in the area of accessibility measurement methodologies, the impediments to implementing accessibility as a performance measure, and the benefits of achieving multimodal accessibility. The goal of this paper is to demonstrate how multimodal accessibility can be used to characterize the quality of transportation service on the city-wide level, and identify the potential benefits and challenges to developing and implementing accessibility measures as a part of the transportation performance measurement efforts that have a direct influence on transportation policy development. The results show the presence of inequity and clear lack of integration of pedestrian, bicyclist, and transit options, particularly outside of the downtown core area of the city. Keywords: Accessibility, Multimodal Transportation, Urban Environment, Performance Measurement

3 Tasic, Bozic, and Hanss INTRODUCTION Transportation shapes the cities, which in return require transportation as the engine of their economic, environmental, and social development. This interaction between cities and their transportation systems is continuous, where the change of one always requires further change of the other. Neglecting this relationship may lead to transportation problems commonly faced by urban environments today, mostly resulting from urban policies that favor one mode over the other modes of transportation (-). The success of the implemented urban transportation projects is evaluated using transportation performance measures, which also influence and form long-term transportation policies for ensuring more sustainable solutions. With the emerging need to incorporate multimodal transportation users in short-term and long-term policy consideration, a new set of performance measures is needed to serve as an indicator of transportation performance in multimodal environments. Accessibility is one such measure, as it is linked to the core purpose of transportation to access the desired destination, but also has the potential to provide relative evaluation of how various users needs are met within a transportation system of a specific design and integration with the environment. The measure of accessibility in transportation enables us to effectively identify neighborhoods and regions that lack multimodal transportation options and target the parts of our transportation network that fails to provide service. Particularly in large urban centers, accessibility matters, because cities would not be able to function without the ability to access transportation options and the opportunities they connect us with. This paper is focused on multimodal accessibility in major urban centers. It starts with a review of the implementation of accessibility in transportation policy and practice. It then provides the overview of the existing accessibility measures, followed by more recent applications of accessibility research. In its core this paper builds on the previous research methods and showcases the implementation of accessibility measures for multimodal users on a large scale network of a City of Chicago. Pedestrian and bicyclist measures are opportunitybased, while public transit accessibility measures account for spatio-temporal variation of transit service. The goal of this paper is to demonstrate how multimodal accessibility can be used to characterize the quality of transportation service on the city-wide level, and identify the potential benefits and challenges to developing and implementing accessibility measures as a part of the transportation performance measurement efforts that have a direct influence on transportation policy development. ACCESSIBILITY IN TRANSPORTATION POLICY, PRACTICE, AND RESEARCH On the national level, transportation policy started to focus on introducing and improving transportation accessibilities in the early nineties. One of the core factors for consideration in the process of transportation planning, established by the Transportation Equity Act for st Century (TEA-) signed in includes increasing accessibility and mobility options available for people and freight. This legislation was followed by the Safe, Accountable, Flexible, Efficient Transportation Equity Act: A Legacy for Users (SAFETEA-LU) signed in 00, continuing to enforce providing efficient access to jobs, services, and centers of trade. The Moving Ahead for Progress in st Century Act (MAP-) signed in 0 also brings up the need to facilitate multimodal connectivity and accessibility.

4 Tasic, Bozic, and Hanss These policies and legislation found their way to be implemented on the state-wide, regional and local level across the United States. The Wasatch Front Regional Council, in its Regional Transportation Plan for 00, emphasizes the need to encourage accessibility from residential zones to other destinations to increase the amount of walking, biking, and active travel in the region. The need to improve access to transit stops by walking or biking is defined as particularly relevant for achieving long-term goals related to more sustainable transportation systems. The Puget Sound Region Council puts the adequate integration of land use and transportation at the core of resolving the issues with proximity to various opportunities in the transportation system. The long-range transportation plan clarifies the link between the land use, transportation infrastructure connectivity, and multimodal transportation options. Similarly, the Southern California Association of Governments transportation plan for 0 requires improved access to opportunities such as jobs, education, and health care and acknowledges how the existing gaps in transportation network produce gaps in access to certain parts of the region. The link between multimodal infrastructure and accessibility, specifically provided through quality transit service and fair treatment of different modes, is a part of Knoxville Transportation Mobility Plan for 00. The Maricopa County Association of Governments addresses the MAP- requirements through plans for encouraging access to opportunities and providing multimodal options. Similar efforts have been taken on the city-wide level for major cities across the country. New York Regional Transportation Plan is highly multimodal. San Francisco Municipal Transportation Authority strategic plan emphasizes that sharing our cars, rides, and bicycles will grow and provide greater access for many more residents, workers, and visitors. The City of Chicago that serves as a case study here, also has adopted transit access targets for the period 00-00, acknowledging serious inequities in access to opportunities and assets on the regional level. The Existing Measures of Accessibility Several types of accessibility measures related to transportation are developed in the existing research: ) Cumulative or opportunity-based accessibility measures evaluate accessibility in terms of the number of opportunities or activity locations that can be reached within the given travel time from a defined reference location (, ). Cumulative accessibility is a function of proximity, connectivity, and mobility (-). ) Gravity-based accessibility measures assign specific weights to the opportunities depending on the distance, travel time, and cost required to reach those opportunities or activity locations. With gravity-based measures, accessibility increases with proximity and affordability of opportunities, and decreases as those opportunities become more distant and their costs increase (-). The available literature emphasizes two disadvantages of these measures, as they require assigning weight to a wide range of destinations, and there is a need for an impedance factor that represents distance, travel time, and cost of the weighted opportunities (-). ) Utility-based accessibility measures incorporate traveler preferences, which affect the weight of opportunities in terms of access. These measures calculate the utility of the chosen opportunity relative to the utilities of alternative opportunities (, -). ) Some measures related to network connectivity in urban areas are also good indicators of accessibility, since denser, better connected networks make destinations easier to reach and increase the number of reachable destinations in general (0, ). Higher

5 Tasic, Bozic, and Hanss connectivity indices improve accessibility up to a certain point, but it does not always guarantee the optimal transportation performance (, ). ) The composite accessibility measure incorporates temporal constraints in addition to spatial constraints for a more complex measurement approach (0-). One of the most powerful techniques for space-time accessibility measurements is the space-time prism (STP). The STP-based accessibility measures determine a feasible set of locations for travel and activity participation, considering spatial and temporal constraints that affect individual s behavior (0). Some earlier STP-based accessibility measures had the disadvantage of treating travel time as static rather than dynamic. After empirical research proved that temporal constraints have a significant impact on an individual s ability to reach desired destinations, the STP-based accessibility measurement methods have been updated to account for this (0, ). Applications of Accessibility Measures in Transportation The measures of accessibility are incorporated in long-term transportation and land use planning, and it is considered that they impact the amount and nature of travel that occurs on various levels (-). Accessibility measures are widely used in the field of travel demand modeling and forecasting, particularly when predicting mode share or conducting traffic assignment (0-). Studies focused on network design and optimization also use accessibility to recommend the network structures that result in maximum level of service for various user types (-). Indicators from accessibility theory are sometimes used to improve the representation of multimodal exposure and capture the presence of multimodal transportation options and the overall access to multimodal infrastructure (-). The most recent advancements in opensource tools for transportation service ratings brought attention to the importance of multimodal accessibility measurements. The most recent implementation of accessibility metrics in transportation includes application development, for different categories of multimodal users, ranging from walking, to bike share, car share, and public transit. Challenges to Measuring Multimodal Accessibility The prerequisite for quality multimodal accessibility measurements are data (, ). It is very challenging, and in the first place time consuming, to collect data about pedestrian networks, as data about crosswalks and sidewalks are very difficult to obtain (). The complexity of multimodal transportation networks in major cities is such that maintaining and updating data inventories becomes a very challenging task. This resulted in scarcity of studies that consider multimodal accessibility. Several existing studies focus on pedestrian accessibility measurements and applications (,). There is an apparent lack of studies focusing on the accessibility provided for bicyclists (). When compared to other modes, studies focusing on transit accessibility measurements are more frequent (,, ). The summary of reviewed research indicates that the scope of accessibility needs to be expanded to include the array of modes, particularly in urban environments. This paper further presents multimodal accessibility measurements using Chicago as a case study, and evaluating the ability to access opportunity by walking, biking, and public transit. CASE STUDY SITE AND DATA COLLECTION The methodology presented in this paper builds on the previous research methods demonstrated on a smaller neighborhood-based networks, and applies the developed accessibility measurements on a large-scale network of the City of Chicago. The complexity of multimodal

6 Tasic, Bozic, and Hanss 0 network in Chicago, the multilayered transportation system, and the presence of true multimodal diversity, were the factors in identifying this major city as an adequate case study. The data collection required to implement the developed methodology included road network data, multimodal infrastructure data, functional classification and speeds, and land use data. Road network data are provided by the City of Chicago open database, with the origin from the Chicago Department of Transportation (CDOT). Data files take the form of digital geospatial polyline features of road centerlines in Chicago, with the most recent update for the year of 0. Several agencies provided multimodal infrastructure data: CDOT, Chicago Transit Authority (CTA) data imported through the Google Transit Feed Specification (GTFS), and the City of Chicago. Multimodal infrastructure included sidewalk areas, bike lanes and bike racks, transit line network, stops, GTFS data about transit schedule and service, rail lines and rail stops. Variables related to multimodal transportation features in Chicago were an important addition to the dataset, as they are rarely present in the previous research at this level of detail. The Illinois Department of Transportation (IDOT) provided functional classification data, while CTA provided transit speed estimates. Chicago Metropolitan Agency of Planning provided the most recent land use data for the year of 00 in the form of parcel data. Figure provides the visualization of multimodal infrastructure in Chicago, spatial distribution of daily vehicle miles traveled (DVMT), and the percentage of street network serving all four modes in each census tract. 0 DVMT NC_Car_WTB < 0,000 < % 0,000-0,000 % - 0% 0,000-00,000 0% - % 00,000-0,000 % - 0% > 0,000 > 0% Daily Vehicle Miles Traveled Percentage of Network Completed FIGURE Road network, multimodal infrastructure, dvmt, and share of street network serving all modes in the City of Chicago.

7 Tasic, Bozic, and Hanss ACCESSIBILITY MEASUREMENT METHODOLOGY Accessibility in transportation varies across the modes of transport, and while access for all modes is a function of infrastructure features and land use characteristics, nonmotorized modes and transit are much more sensitive to the way activities are distributed in space and time than private vehicles. In this paper, accessibility measures are developed and applied for three types of transportation users: pedestrians, bicyclists, and public transit users. Accessibility for Pedestrian Mode The accessibility measurement approach for nonmotorized modes builds on a previous, smaller scale case study, used to quantify pedestrian and bicyclist accessibility (). For nonmotorized modes, a specific destination is considered accessible if there is a connection between the origin and destination, if that connection is within a defined distance, and if it is reachable within a defined time frame. The first step in a pedestrian accessibility assessment was to define the potential origins and destinations in a study area of Chicago. Land use parcels were used as both origins and destinations for pedestrian trips in the city. This resulted in total of, origins defined for pedestrian trips, and just as many destinations. Defining each land use parcel as both an origin and a destination, made the computational process very exhaustive, but significantly contributed to more precise measurements of pedestrian accessibility. To authors knowledge, no previous studies measured pedestrian accessibility at a city-wide scale using such high level of data aggregation in terms of trip origins and destinations. In addition to defining adequate origins and destinations, another challenge in calculating pedestrian accessibility was to properly define pedestrian network in the city. The available sidewalk area data were used to edit the street network of Chicago and include only those streets in Chicago that have sidewalk in the pedestrian accessibility analysis. This is a standard way of manipulating the street network data to ensure that freeways and ramps are excluded from the final pedestrian network (). This approach, however, does not account for all pedestrian paths in the city, as some pedestrian routes that cut through parks and public spaces were not incorporated in the shortest path search between the origins and destinations. While excluding the off-street pedestrian paths from the analysis could be a limitation for this study, using the entire street-based pedestrian network provided a good approximation for the possible pedestrian routes in the city. For the defined origins and destinations, an origin destination (OD) matrix was created for all possible OD combinations with the following questions assessed for each pair (): Is there a feasible walking connection between origin and destination? Is the distance between origin and destination adequate for pedestrians? Is the time needed to reach the destination adequate for pedestrians? To answer the first question, possible connections between each O-D pair were identified as uninterrupted paths between an origin and a destination, on the terrain appropriate for pedestrians, with AASHTO guidelines for the adequate path width of.ft (AASHTO, 00). One-quarter mile and half a mile distances were adopted as the criterion for acceptable walking distances as suggested by the AASHTO Guide for Planning and Design of Pedestrian Facilities (). The distances between origins and destinations were measured in ArcGIS and calculating shortest paths for pedestrians. By applying MUTCD guidelines for average pedestrian speed of feet per second to this distance, the one-quarter mile distance criterion suggests that visitors are not willing to walk more than 0 minutes to reach their destination (). Pedestrian speeds range from. feet per second to feet per second, and the average pedestrian speed is usually related

8 Tasic, Bozic, and Hanss to the age of the population in the observed area. The pedestrian speed of feet per second was used to determine travel times between origins and destinations. Several time buffers were used to account that pedestrians might be willing to walk longer to certain destinations, including time frames from minutes to 0 minutes walking time with minute increments (). This study primarily measured cumulative pedestrian accessibility defined as the total number of destinations accessible to pedestrians within the defined time frames. In the case of cumulative measures, the same weight is used for all destinations, acknowledging this as a limitation that should be addressed in potential future research efforts, as all destinations do not have the same level of attractiveness, and visitors might be willing to walk longer to some destinations than others. Using the described guidelines and assumptions, the cumulative number of opportunities for pedestrians was calculated as following: Ped i = i j{d ij N T ij T} Equation () Where: Ped i total number of destinations accessible from origin i within time T d ij destination j accessible from origin i within time T ij N total number of destinations T ij time needed to reach destination j from origin i T available time budget (, 0,, 0,, or 0 minutes) In addition to cumulative pedestrian accessibility, weighted accessibility is also calculated by incorporating the travel time impedance function into the equation for cumulative accessibility measures: Ped iw = {d i j ij N T ij T} Equation () T ij Where: Ped iw weighted pedestrian accessibility for origin i d ij destination j accessible from origin i N total number of destinations T ij time needed to reach destination j from origin i The results of pedestrian accessibility measurements are presented and discussed in the following section. Accessibility for Bicyclists Accessibility for bicyclists was computed similar to pedestrian accessibility, but with different standards for the acceptable biking distances and travel times. The origins again were defined using all land use parcels (without the parcels referring to vehicular right of way areas) in Chicago, and the destinations were defined in the same way as origins. Defining destinations and origins like this enabled building OD matrices from each land use parcel to all other land use parcels in Chicago. In order to be considered accessible for biking, there should be an uninterrupted connection between the origin and destination, the origin and destination should be within the acceptable distance, and the destination should be reached within the acceptable travel time for bicyclists.

9 Tasic, Bozic, and Hanss 0 0 A connection was defined as an uninterrupted path between an origin and a destination, on the terrain appropriate for bicyclists, but this time with AASHTO guidelines for adequate operating spaces for bicyclists (0). Based on the reviewed literature, acceptable biking distance was defined as miles (0). An average biking speed of miles per hour was used for average biking speed as suggested by the AASHTO Guide for Planning and Design of Bicycle Facilities (0). As for the acceptable biking time for most of the cyclists traveling from an origin to a destination, times of, 0,, and 0 minutes were adopted. Cumulative number of opportunities for bicyclists were then computed by using the following equation: Bike i = i j{d ij N T ij T} Equation () Where: Bike i total number of accessible destinations by bike from origin i d ij destination j accessible from origin i N total number of destinations T ij time needed to reach destination j from origin i T available time budget (, 0,, and 0 minutes) Similar as for pedestrian accessibility, weighted bicycle accessibility is calculated (please see Table for results) by using the following equation: Bike iw = {d i j ij N T ij T} Equation () T ij Where: Bike iw weighted bicyclist accessibility for origin i d ij destination j accessible from origin i N total number of destinations T ij time needed to reach destination j from origin i The results of bicyclist accessibility measured in this manner are presented and discussed in the following section. This particular approach for nonmotorized accessibility was adopted to not only differentiate between the two modes, but to also clarify that even with high level of overall network connectedness, accessibility for nonmotorized modes might still be limited. Transit Accessibility As public transit has unique characteristics among other modes, due to its spatial and temporal constraints, using composite space-time accessibility measures is appropriate for developing transit accessibility indicators. The methodology for transit accessibility measurement builds upon the traffic and transit data from the case study network, and uses transit network and GTFS data. The transit accessibility measurement framework builds on the previously developed method for a smaller scale study network (). The methodology considers network features, acceptable walking time, available time budget, transit schedule variability, and spatial constraints as impact factors in accessibility measurements. The City of Chicago road network with nodes, links, census tracts, transit network, and transit stations, imported as GIS shapefiles, was used as a basis for transit accessibility calculations. Transit accessibility, is defined as the average daily number of destinations

10 Tasic, Bozic, and Hanss reachable by transit, using both walking and transit routes, constrained by spatial characteristics of the case study network and temporal dimension determined by transit service and traffic characteristics. In order to execute transit accessibility measurements, the transit station areas are defined as trip origins, while all land use parcels in the city were considered as potential destinations. Transit lines, stations, and speed data available from CTA were combined with Chicago GTFS to provide the information on spatial and temporal distribution of transit services. The GTSF from Google includes the following (): Calendar that specifies when service starts and ends, including the days of the week when service is available Calendar dates with possible service exceptions Routes or groups of trips displayed to riders as single service Shapes or rules for representing transit routes on the maps Stop times or arrival and departure times for each individual trip Stops or passenger pick up and drop off points Trips or sequences of stops for each route The stop time records that include a sequence of stops along each trip (trip identification, arrival and departure time, stop identification, and stop sequence) were used to determine how many times within each -minute period over the course of one day is each transit station accessible within the various combinations of time needed to walk to/from a transit station and time needed for a trip by public transit. A total of, transit stations with up to stop time records per station were included in the analysis that resulted in average daily accessibility to destinations by transit on the census tract level. While the resulting measures of transit accessibility appear similar to those of pedestrian and bicyclist accessibility, using the data on transit stop time and transit schedule helped to incorporate daily temporal variations of transit service into census tract-level transit accessibility measurements. This inclusion of time-dependent transit availability dynamics is the key difference between accessibility measurements for non-motorized modes and transit mode, as transit travel times to destinations vary with both space and time. The idea to combine spatial and temporal changes in transit service in transit accessibility measurements is rooted in the composite space-time accessibility measures based on Miller s STP concept (). The STP is a set of locations in space and time that are accessible to an individual, given the locations and duration of fixed activities, time budget, and transportation speeds. The STP-based accessibility measures account for both individual sequence of trips and spatio-temporal constraints, calculating the amount of space that an individual can reach at specific combinations of times and locations. In the case of public transit, travel times are not only affected by traffic conditions, but also by the time needed to access the transit stop, waiting time which depends on familiarity with the timetable, potential stops and transfers, and the time needed to reach the destination from the final transit stop. The accessibility models were developed based on the dynamic potential path calculations, but also to account for pedestrian connectivity, transit stop accessibility, and scheduled service variability as elements that specifically relate to transit mode. Instead of using a simple, radius-based service coverage, the actual transportation network was used. Walk to transit was adopted as the mode used to access transit stations. Calculations and assumptions for different space-time constraints, applied to compute average daily number of destinations accessible by transit is as follows:

11 Tasic, Bozic, and Hanss PT i = d ij(walk)+d ij (transit) f it i j t Equation () nt Such that: T ij = t ijwalk + t ijtransit T 0 Where: PT k daily average number of destinations accessible by transit from origin i d ij (walk) destination j accessible from origin i by walking d ij (transit) destination j accessible from origin i by transit f it frequency of transit stop time records at station i during time period t (time period t in this case is a -minute period) nt total number of time periods t within a working day N total number of destinations T ij time needed to reach destination j from transit station i (min) t ijwalk total walk time included in the trip between origin i and destination j (acceptable walking time includes, 0,, and 0 minutes in this case) t ijtransit total time spent in public transit between origin i and destination j (acceptable time spent in public transit does not exceed 0 minutes) T available time budget (up to 0 min) The defined origins and destinations with the public transit network of Chicago were uploaded in ArcGIS platform, and shortest path calculations were conducted between each OD pair for all determined constraints related to acceptable time budget. Shortest path is calculated between transit stations and destinations accessible within 0-minute walk distance and pairs of transit stations located within 0 minutes distance traveled by transit regardless of the number of transfers. The accessible number of destinations is then calculated as a sum of all destinations accessible by combining walking and public transit within 0 minutes of travel time budget. This measure of accessibility is calculated for each -minute time period during the daily period of the public transit service, resulting in time variable transit accessibility for each origin over the course of a working day. Weighted transit accessibility was also calculated, using the approach similar to the one used for nonmotorized accessibility, based on the travel time impedance: PT iw = PT i j t ijwalk +t ijtransit i Equation () Where: PT iw weighted transit accessibility in from origin i PT i daily average number of destinations accessible by transit from origin i t ijwalk total walk time included in the trip between station i and destination j t ijtransit total time spent in public transit between station i and destination Evaluating public transit is always more complex than any other mode of transportation, and selecting adequate accessibility measures is also a challenge. Several factors that impact space-time constraints were included in transit accessibility analysis. Service variability refers to the frequency of transit service and service span in general. Walking distance is the acceptable

12 Tasic, Bozic, and Hanss walking distance to transit stops. Available time budget defines the time that individual has to access activity locations from the given trip origin. Transit accessibility measurement results are presented and discussed in the following section of the paper. CASE STUDY RESULTS AND DISCUSSION This section presents the results and discussion of results obtained from the multimodal accessibility measurements based on the methodologies from the previous section. Figures,, and present the results of pedestrian, bicyclist, and transit accessibility calculations. Since there are, origins and destinations used to conduct the measurements, the resullts presented in Figures through are aggregated on the census tract in the City of Chicago, to provide better visualisation of accessibility variations for multimodal users. Figure shows the cumulative pedestrian accessibility, calculated by using the Equation for the defined time frames of, 0,, 0,, and 0 minutes of the assumed acceptable walking time. As expected, census tracts with the highest number of destinations accessible within the defined time frames are mostly located in city center and the North side of the city, as these areas have higher densities and more developed pedestrian networks. Some census tracts on the South side also show higher cumulative pedestrian accessibility, mostly where current multimodal infrastructure investments are concentrated. Figure shows the spatial distribution of cumulative bicyclist accessibility in Chicago. When compared to Figure, Figure shows significantly higher cumulative accessibility for bicyclists than for pedestrians, due to obvious differences in the amount of time users are willing to spend walking and biking. Even the highest considered biking time budget of 0 minutes does not render the entire area of the city highly accessible for bicyclists, particularly some census tracts alongside the lake Michigan. As expected, the downtown area as well as census tracts along the so-called diagonal avenues in Chicago show a high number of destinations reachable by bike. The gaps in nonmotorized accessibility are visible in Figures and, as they occur if the network of pedestrian and bicyclist facilities are disconnected with many interruptions and segments for vehicular mode only, then due to large distances between origins and destinations, and finally due to any traffic conditions that might cause delay for pedestrians and bicyclists on the way to their trip destinations. Figure shows spatial distribution of the average daily transit accessibility on the census tract level. The results show relatively good coverage of the entire city by transit service, where the downtown area of Chicago has the best access to destinations by transit, as expected. The variation in transit service is noticeable across the entire city, as it is highly concentrated along the major elevated train lines, and then dispersed in the suburbs of Chicago. The maps provided on Figure show how diverse the city itself is in terms of land use patterns, ranging from high density to sprawling neighborhoods. Transit accessibility also appears to be lower in the South Side of the city, where most of the lower income neighborhoods are concentrated, and this could be the topic for future research efforts. What was emphasized in this paper is how different transit is from other modes, due to the way it is constrained in terms of space (limited to station areas) and time (limited to schedule and service frequency). The total number of transit stop time records in Chicago ranges between a 000 and 000 stops for all transit stations within each -minute period over the course of a day. This variation in stop frequency influences overall city-wide transit accessibility, in addition to the influence that other factors such as walking time and transit travel time have on the ability to reach destinations by transit. Figures and provide a more detailed insight into daily

13 Tasic, Bozic, and Hanss 0 0 variations of transit service in Chicago. Figure a presents the daily dependence of transit service on walk time, and shows how the service patterns remains consistent, but the overall accessibility to service would increase for users willing to spend longer time walking to transit stations. Similar are the results shown in Figure b, presenting the dependence of transit accessibility on the available users time budget, where accessibility increases for users who can afford more time riding transit to their destinations. Accessibility between each pair of transit stations in the city was calculated as a part of average accessibility calculations for public transit mode. The percent of accessible transit stations is then extracted as a ratio of transit station pairs accessible within a given time over the total number of transit station pairs. These results are provided in Figure a. As expected, the percent of accessible transit stations increases as the users travel time budget increases. This is further confirmed with the results provided in Figure b, where transit accessibility incremental change is presented dependent on the various combinations of walk time and time budget. All these and other factors that may influence transit accessibility in urban environments are explained in a more detailed manner in a previous study (Tasic et al., 0). Comparing the spatial variation in accessibility for pedestrians, bicyclists, and transit from the Figures through, even though the measures used for nonmotorized and transit users are based on different methods, it is possible to notice the inconsistency that exists in terms of service quality for various modes within the same areas. The access to opportunities via multimodal options is fairly consistent only in the downtown area. The other areas of the city show poorer integrations of multimodal options. It is particularly noticeable that some areas with better transit accessibility, still have lower pedestrian accessibility, which would force transit users to either bike or use private or shared vehicles, particularly when users are not willing to spend more than minutes walking to transit stations. While the developed accessibility indicators provide the basis for a general multimodal accessibility indicator, calculation of such an indicator would require making assumptions about user preference in terms of mode choice that are not strongly supported by the data available in this study. The development of a single indicator of multimodal accessibility remains a part of future research efforts.

14 Tasic, Bozic, and Hanss Ped_D Ped_D0 Ped_D < < < > 0 > 0 > 00 minute Walk Time 0 minute Walk Time minute Walk Time Ped_D0 Ped_D Ped_D0 < < < > 00 > 00 > minute Walk Time minute Walk Time 0 minute Walk Time FIGURE Destinations accessible within the given walking time.

15 Tasic, Bozic, and Hanss Bike_D Bike_D0 < 00 < > 000 > 000 minute Biking Time 0 minute Biking Time Bike_D Bike_D0 < 00 < > 000 > 0000 minute Biking Time 0 minute Biking Time FIGURE Destinations accessible within the given biking time.

16 Tasic, Bozic, and Hanss Transit_D Transit_D < 00 < > > minute Travel Time by Transit 0 minute Travel Time by Transit Transit_D Transit_D < 00 < > 000 > minute Travel Time by Transit 0 minute Travel Time by Transit FIGURE Destinations accessible within the given travel time by transit.

17 Average Number of Destinations Accessible by Transit :00 :0 :00 :0 :00 :0 :00 :0 0:00 0:0 :00 :0 :00 :0 :00 :0 :00 :0 :00 :0 :00 :0 :00 :0 :00 :0 :00 :0 0:00 0:0 :00 :0 Average Number of Destinations Accessible by Transit :00 :0 :00 :0 :00 :0 :00 :0 0:00 0:0 :00 :0 :00 :0 :00 :0 :00 :0 :00 :0 :00 :0 :00 :0 :00 :0 :00 :0 0:00 0:0 :00 :0 Tasic, Bozic, and Hanss Transit Accessibility and Walking Time Time min Walk Time 0 min Walk Time min Walk Time 0 min Walk Time FIGURE a Daily variations of transit accessibility for all census tracts dependent on preferred walk time within the 0-minute available time budget. Transit Accessibility and the Available Budget Time min budget Time 0 min Budget Time min Budget Time 0 min Budget Time 0 min Budget Time 0 min Budget Time FIGURE b Daily variations of transit accessibility for all census tracts dependent on the available budget time within the -minute walk from the transit stations.

18 Number of Accessible Destinations Percent Change in Accessibility Number of Transit Stop Pairs Tasic, Bozic, and Hanss 0000 Transit Stop Accessibility 00% 0000 % % 0000 % 0 0% Access Time between Pairs of Transit Stops (min) FIGURE a Number and percentage of transit stops accessible within the given amount of time budget. Incremental Accessibility Change 000 0% Available Budget Time (min) 0% 0% 0% 0% 0% 0% 0 min Walk Access 0 min Walk Access min Walk Access 0 min Walk Access FIGURE b Incremental change in transit accessibility depending on the available time budget and walking distance.

19 Tasic, Bozic, and Hanss CONCLUSIONS AND RECOMMENDATIONS This paper used the multimodal transportation infrastructure and land use data from the City of Chicago to demonstrate how multimodal accessibility can be measured for pedestrians, bicyclists and public transit users. Nonmotorized accessibility measures are developed from the opportunity and gravity-based methods, while transit accessibility measures account for spatiotemporal constraints that characterize transit service. These measures can further be used to identify gaps in transportation service and coverage for each mode individually, as well as for the quality of integration of these three modes on the city-wide level. The measures presented here are built on the previous applications for the case studies of small-town networks, and ensure transferability to various street network and land use patterns. The data collection in this paper was conducted to acknowledge the complexity of multimodal transportation systems. The study provides accessibility measurements using all land use parcels in Chicago as origins and destinations, and detailed data on road, pedestrian, bicyclist, and transit network. This inclusion of multimodal accessibility on a higher level is achieved to ensure higher level of measure accuracy for all users that were accounted for in this study. The detailed data collection effort and the development of accessibility measures for all three alternative modes of transportation for a large-scale multilayered network are the major contributions of this study. The developed measures of accessibility may serve as indicators of success of multimodal transportation systems in terms of accommodating the needs of all transportation users on urban street networks. Future research efforts should incorporate users preference for different land use types, and more robust pedestrian network. Potential directions for future research could also include exploring the associations of transportation accessibility with social equity, improved estimates of modal share, and implications for policy development and implementation. REFERENCES. Vuchic, V. R. (). Transportation for livable cities. New Brunswick, NJ: Center for Urban Policy Research.. Jacobs, J. (). The death and life of great American cities. New York: Vintage Books Edition.. Mumford, L. (). The city in history: Its origins, its transformations, and its prospects. New York: Harcourt, Brace & World.. Black, J., & Conroy, M. (). Accessibility measures and the social evaluation of urban structure. Environment and Planning A, (), 0-0. doi:0.0/a00. Handy, S. L., & Niemeier, D. A. (). Measuring accessibility: An exploration of issues and alternatives. Environment and Planning A, (), -. doi:0.0/a. Cascetta, E., Cartenì, A., & Montanino, M. (0). A behavioral model of accessibility based on the number of available opportunities. Journal of Transport Geography,, -. doi: Chen, Y., Ravulaparthy, S., Deutsch, K., Dalal, P., Yoon, S., Lei, T.,... Hu, H.-H. (0). Development of Indicators of Opportunity-Based Accessibility. Transportation Research Record: Journal of the Transportation Research Board,, -. doi:0./-0. Handy, S. (00). Accessibility vs. mobility enhancing strategies for addressing automobile dependence in the U.S. Prepared for the European Conference of Ministers of Transport, Paris, France.. Alam, B., Thompson, G., & Brown, J. (00). Estimating Transit Accessibility with an

20 Tasic, Bozic, and Hanss Alternative Method. Transportation Research Record: Journal of the Transportation Research Board,, -. doi:0./-0 0. Cheng, J., & Bertolini, L. (0). Measuring urban job accessibility with distance decay, competition and diversity. Journal of Transport Geography, 0, doi: El-Geneidy, A., Levinson, D., Diab, E., Boisjoly, G., Verbich, D., & Loong, C. (0). The cost of equity: Assessing transit accessibility and social disparity using total travel cost. Transportation Research Part A: Policy and Practice,, 0-. doi: Hansen, W. G. (). How accessibility shapes land use. Journal of the American Institute of Planners, (), -. doi:0.00/000. Scheurer, J., & C. Curtis. (00). Accessibility Measures: Overview and Practical Applications. Department of Urban and Regional Planning, Curtin University.. Ben-Akiva, M., & S. R. Lerman (). Disaggregate travel and mobility choice models and measures of accessibility. In D.A. Hensher & P.R. Storper (Eds.), Behavioral travel modelling (pp. -). London: Chrom-Helm.. Ben-Akiva, M., & Bowman, J. (). Integration of an activity-based model system and a residential location model. Urban Studies, (), pp. -. doi:0.00/000. Geurs, K.T. & Ritsema van Eck, J. R.. (00). Accessibility measures: Review and applications. Bilthoven, Netherlands: National Institute of Public Health and Environment (RIVM).. Kuzmyak, J., Baber, C., & Savory, D. (00). Use of Walk Opportunities Index to Quantify Local Accessibility. Transportation Research Record: Journal of the Transportation Research Board,, -. doi:0./-. Alba, C., & E. Beimborn (00). Analysis of the effects of local street connectivity on arterial traffic. Proceedings of the th Annual Meeting of the Transportation Research Board. Washington, D.C.: Transportation Research Board of the National Academies.. Tasic, I., Zlatkovic, M., Martin, P. T., & Porter, R. J. (0). Street connectivity versus street widening. Transportation Research Record: Journal of the Transportation Research Board,, -. doi:0./-0 0. Kwan, M. (00). Space-time and integral measures of individual accessibility: A comparative analysis using a point-based framework. Geographical Analysis, 0(), -. doi:0./j.-..tb00.x. Miller, H. J. (00). Measuring space-time accessibility benefits within transportation networks: Basic theory and computational procedures. Geographical Analysis, (), -. doi:0./j.-..tb000.x. Neutens, T. (0). Accessibility, equity and health care: review and research directions for transport geographers. Journal of Transport Geography,, -. doi: Tong, L., Zhou, X., & Miller, H. J. (0). Transportation network design for maximizing space time accessibility. Transportation Research Part B: Methodological,, Part, -. doi: Alam, B., Thompson, G., & Brown, J. (00). Estimating Transit Accessibility with an Alternative Method. Transportation Research Record: Journal of the Transportation Research Board,, -. doi:0./-0

21 Tasic, Bozic, and Hanss Allen, W. B., Liu, D., & Singer, S. (). Accesibility measures of U.S. metropolitan areas. Transportation Research Part B: Methodological, (), -. doi: Cao, J., Liu, X. C., Wang, Y., & Li, Q. (0). Accessibility impacts of China s high-speed rail network. Journal of Transport Geography,, -. doi: Chandra, S., Bari, M. E., Devarasetty, P. C., & Vadali, S. (0). Accessibility evaluations of feeder transit services. Transportation Research Part A: Policy and Practice,, -. doi: Fransen, K., Neutens, T., Farber, S., De Maeyer, P., Deruyter, G., & Witlox, F. (0). Identifying public transport gaps using time-dependent accessibility levels. Journal of Transport Geography,, -. doi: Ryan, S., & McNally, M. G. (a). Accessibility of neotraditional neighborhoods: A review of design concepts, policies, and recent literature. Transportation Research Part A: Policy and Practice, (), -0. doi: 0. Beimborn, E., Greenwald, M., & Jin, X. (00). Accessibility, Connectivity, and Captivity: Impacts on Transit Choice. Transportation Research Record: Journal of the Transportation Research Board,, -. doi:0./-0. Bhat, C., & Zhao, H. (00). The spatial analysis of activity stop generation. Transportation Research Part B: Methodological, (), -. doi: (0)000-. Case, R. (0). Accessibility-Based Factors of Travel Odds. Transportation Research Record: Journal of the Transportation Research Board,, 0-. doi:0./-. Owen, A., & Levinson, D. M. (0c). Modeling the commute mode share of transit using continuous accessibility to jobs. Transportation Research Part A: Policy and Practice,, 0-. doi: Bigotte, J. F., Krass, D., Antunes, A. P., & Berman, O. (00). Integrated modeling of urban hierarchy and transportation network planning. Transportation Research Part A: Policy and Practice, (), 0-. doi: Daganzo, C. F. (00). Structure of competitive transit networks. Transportation Research Part B: Methodological, (), -. doi: Tong, L., Zhou, X., & Miller, H. J. (0). Transportation network design for maximizing space time accessibility. Transportation Research Part B: Methodological,, Part, -. doi: Kim, K., Pant, P., & Yamashita, E. (0). Measuring influence of accessibility on accident severity with structural equation modeling. Transportation Research Record: Journal of the Transportation Research Board,, -0. doi:0./-0. Kim, K., Pant, P., & Yamashita, E. (00). Accidents and accessibility. Transportation Research Record: Journal of the Transportation Research Board,, -. doi:0./-0. Mondschein, A., Brumbaugh, S., & Taylor, B.. (00). Congestion and accessibility: What s the relationship? Presented at the th Annual Meeting of the Transportation Research Board, Washington, D.C. 0. Sathisan, S., & Srinivasan, N. (). Evaluation of Accessibility of Urban Transportation Networks. Transportation Research Record: Journal of the Transportation Research Board,, -. doi:0./-