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.
48 Using the models described in the previous chapter, a draft toolkit was developed. This chapter will provide a discussion of what has been included and whyâthe actual toolkit is entirely included on the accompanying CD-ROM that uses Microsoft® Word and Excel files to enable the user to define potential routes, estimate ridership, and make adjustments to those estimates. Format Given the desire to make the toolkit easy to use, the study team was faced with two key issues: (1) Both model approaches involve the use of population data, which can be a complication to find and compile, and (2) both involve the need to plug the data into a formula. Initially there were thoughts about pro- viding hypertext links to Census population websites to allow the user to access data, but it was decided to create a computer- based package that would include all the needed population data and would have the models embedded so that the answer would be calculated for the user, based on inputs. The CD includes the required U.S. Census population data, with populations for all urbanized areas (over 50,000), urban clusters (2,500 to 50,000), and Census-designated places (under 2,500) from the 2000 Census. A user-friendly inter- face tells the user to make sure that this model is appropriate for their proposed service, provides some background, provides instructions, and then allows the user to select the stops for their proposed route. The user indicates whether the service connects to an airport and whether it is provided by an inter- city bus operator. The model selects the appropriate regional rate for the trip rate model, based on the state selected. Once the data is input, an output page presents the predicted rider- ship based on each technique. This ridership is characterized as an initial estimate of potential demand. Links to text describe possible means of adjusting the result based on the availability of other services or particular adjustments (post-processing techniques). Finally, a box presents basic service characteristics and ridership for the four most similar routes in the database, along with links to more descriptive information for each of these routes (case studies). All of thisâthe user interface, the data, all formulas, and the instructional textâis on a single CD. The user sees some instructional text, some entry boxes (set up to allow the user to select place names, rather than type them in), and a number of links to specialized text. The instructional text is presented here in the order in which the user will encounter it on the CD. Toolkit Development The development of the toolkit involved the collection of dataâtype of carrier, fare levels, frequency of service, route length, and stopsâon many rural intercity bus routes across the country. This service information was combined with data regarding the populations and types of key destinations served. Many of the routes were identified by contacting state departments of transportation to identify routes that have received funding under the Section 5311(f) program of rural intercity transit assistance; however, unsubsidized routes with rural stops were also included in the database. The database used to calibrate and validate the models included in the toolkit had the following characteristics: ⢠Route length: 50 to 453 miles ⢠Frequency (round-trips per week): 1 to 21 ⢠Average corridor population (not including destination city): 1,196 to 275,108 ⢠Annual ridership (boardings): 877 to 64,434 ⢠Number of routes servicing airports: 34 out of 57 ⢠Number of routes operated by intercity bus carriers: 38 out of 57 Not all Section 5311(f)âfunded routes were included in the database used to calibrate and test models, as some have characteristics that differ from the intercity definition used to C H A P T E R 6 The Toolkit
classify routes, or there were issues with the available data on services or ridership. More information on the database and the development of the toolkit is available in the previous chapters of this report. The toolkit includes information on alternative methods for estimating ridership: ⢠Previous or current ridershipâuse of data from previous or current services on the route. ⢠Data and models for two alternative methods for develop- ing an initial estimate of ridership ⢠A discussion of potential adjustments to the initial estimate ⢠A data set of rural intercity route characteristics and rider- ship, to facilitate consideration of ridership on services that are analogous to the proposed service Users should understand that the estimation of demand for any transportation service is one step in an overall planning process, and that demand estimates by their nature have a high level of uncertainty. The actual ridership on any service can be significantly affected by many variables that are not quantified in a model or technique. This is particularly true for intercity services because the actual ridership may be affected by variables such as the continuity of service, the possibility of overhead traffic (riders using the route to travel to and from points beyond the ends of the segment being analyzed), competing services or modes (other parallel intercity routes, intercity rail passenger services, or even airport ground transportation services), the quality of station facilities, or even the particu- lar abilities of the sales staff at agencies on the line. Toolkit Applicability The applicability of the toolkit is limited to rural intercity services: ⢠Rural routes (or stops on routes) that provide service to places with populations under 50,000 (though they may also have stops in areas with higher populations, such as urbanized areas over 50,000) ⢠Intercity services, defined as services that â Serve two or more urban areas (defined in Section 5311(f) as places over 2,500 persons) not in close proximity â Have limited stops â Provide capacity for passengers to carry baggage â Offer a connection to the national network of intercity bus services This toolkit is not appropriate for: ⢠Commuter bus service, defined as peak-hour, peak- direction, low-frequency service, weekdays, often with commutation tickets providing multiple rides ⢠Local or regional transit, characterized as having many local stops, low per-mile flat or zone fares, multiple frequencies, often weekdays only, and without connectivity to the national intercity bus network ⢠Rural public transit offering area-wide demand-responsive service or rural fixed or route-deviation services that may be low frequency or serve areas with a population under 50,000, but have many local stops and low flat or zone fares and do not offer connections to the national intercity bus network (in terms of serving common stops or having coordinated schedules) Steps in Applying this Toolkit Step 1: Define Proposed Route The initial step requires the user to define the service under consideration. Earlier steps in the planning process should help in defining potential service options. In this initial step, the user should obtain a map and determine stops along the route. For purposes of estimating operating costs or develop- ing schedules, the distance between each stop may be needed, but only the length of the entire proposed route is needed for this toolkit. This length can be easily determined by using an Internet-based mapping site. Also, the user should determine if the proposed operator is an intercity bus company that is a member of the NBTA or a regional provider (public, private non-profit, or for-profit). A technique or warning has been included regarding routes that link two large population centers. The model application in the toolkit defines the largest population centers. The model application in the toolkit defines the largest population city on the route as the âDestinationâ and does not include that population in the estimation of demand, as those places typ- ically have a lot of additional intercity bus service. However, if there are two large population centers, the lower popula- tion center would still be included, and it could overstate the demand considerably. In those cases, it is recommended that one of the two large population centers be excluded in the selection of locations served while using the tool to estimate intercity demand. Step 2: Determine if there is Current Service The potential demand for new service could be affected to a large extent if there is other service linking residents of the areas served by possible stops with intercity services. This other service could involve intercity bus services that are not known to the analyst that directly parallel the proposed service or services that offer alternative routes from an origin to the proposed destination. It could include other types of service that may be alternatives to intercity bus services, such as airport ground transportation providers or intercity rail passenger 49
services. It could include transit services operated by local or regional public transit agencies. While the precise impact of such services can be difficult to determine without doing surveys and additional market research, this information is useful in evaluating ridership estimates provided by this toolkit and in making adjustments to those estimates. Also, existing providers may be willing to share data on current ridership or unmet demands for alter- native services that can be useful in estimating ridership. For example, a planner may check common sources of informa- tion on intercity bus service and determine that a particular corridor has no intercity service. However, additional research may reveal that an airport ground transportation provider serves that corridor and offers connections to an intercity bus station as an alternative destination. If that operator is willing to provide some ridership information, it could be compared to estimates provided by the models calibrated on rural inter- city data to see what proportion of the overall estimated market might already be served. Potential sources of information on existing services include: ⢠Greyhound schedule and route information (http://www. greyhound.com/revsup/schedules/). ⢠Trailways schedule and route information (http://www. trailways.com/). ⢠Amtrak intercity rail passenger service and connecting Amtrak thruway bus services (http://www.AMTRAK.com). ⢠Russellâs Official National Motor Coach Guide (http://www. russellsguides.com/mainpage.shtml or 329 10th Avenue, S.E., Cedar Rapids, Iowa 52401) is a monthly publication that provides timetables for many (but not all) scheduled intercity bus routes in the United States and Canada. It is indexed by stop, so a user can determine which timetables involve service to a particular place. This publication has been available since 1927, and earlier issues can provide information about previous routes and schedules. ⢠American Intercity Bus Rider Association map (http:// aibra.org/). ⢠Airport providers (contact individual airport websites for ground transportation information). If there are existing operators on the proposed route, contact them to try to obtain existing ridership at points that are currently served. Ask about customer service requests that they may have had regarding service similar to the proposed route or about unmet needs. Step 3: Determine Historic Ridership, if there has been Intercity Bus Service on the Route in Recent Times Rural intercity bus service has seen major service abandon- ments over the past 30 years, and information on ridership for services abandoned long ago will not likely be available or useful. However, the unsubsidized private intercity bus industry has been undergoing route and service restructuring over the past decade, and in many cases the routes or stops proposed for analysis have seen operation within the past decade. In some cases, ridership or revenue per mile data may be available from the previous carrier or its former represen- tatives. Again, this information can be useful in evaluating new ridership estimates, making adjustments, or as a basis for estimating potential demand. Step 4: Generate Ridership Estimates Data Entry. The toolkit includes two models that will generate estimates of annual ridership, based on user inputs regarding the characteristics of the proposed service. The information necessary to generate the initial estimate includes the following: ⢠The name of the state in which the largest portion of the route will be operated ⢠Whether or not the proposed route will serve one or more airports with commercial air service ⢠Whether or not the proposed service will be operated by a national or regional bus carrier that is part of the national intercity bus network ⢠The names of the urbanized area, urban cluster, or Census- designated places that will be designated stops for the proposed service. Note that all these place names and their associated populations are already part of the toolkit, and the user will only need to scroll through and select or enter the place names. There is also a place to enter population data the user may have for any stop locations for which there is no data provided in the toolkitâfor example, places too small to have population counts from the Census. Note that it is up to the user to enter stop locations that make sense as a potential routeâthe toolkit will not limit users to stops that are located in any one state or adjoining states, nor will it prevent other entry errors. Demand Estimates. Once this information is entered, the user can click on the âOutput/Resultsâ button, which will advance to a new page of the toolkit showing the results of the demand estimation. On this page the input characteristics selected by the user are presented, along with annual rider- ship estimates developed by two different models. Models Used. The regression model is a statistical equation that was estimated using the ridership data and route charac- teristics of the 57 routes included in the study database. The equation estimates the annual ridership based on the average 50
population of the stops served by the route (excluding the largest population stop, which is assumed to be the destination), the length of the route, whether the route serves an airport, and whether the route would be operated by a carrier that is part of the national intercity bus network. The output page pres- ents a 95 percent confidence interval (CI) for this estimate, meaning that there is a 95 percent probability that the mean ridership of all routes with these characteristics is within the upper and lower bounds of the CI. The trip rate model is a different approach using data on the number of long-distance trips (over 50 miles in length) made by rural residents using public transportation modes. This data was collected as part of the National Household Travel Survey. Data on the number of such trips and rural population were used to calculate a long-distance trip rate for each region of the country. In this model the region is selected based on the âstateâ designation selected by the user. The results are generated based on the population of the points served by the route and the trip rate. An adjustment is made to these results based on the residual error term of a regression model that includes the length of the route, the number of stops, whether the route serves a correctional facility, the average population of the origin stops, and the total population served. Accuracy of the Estimates. The two models are presented to provide the user with more information about the potential ridership estimates. Table 6-1 presents information about the potential accuracy of the two approaches, which is similar. Comparable Routes. The output page also provides a table entitled âComparable Routes.â This table presents summary data on ridership and route characteristics for the four routes in the database that are most similar to the proposed route. This allows the user to further evaluate the results of the two demand models by comparing the proposed route to peer services. A link to additional descriptive information is pro- vided for each of the comparable routes, in case additional detail is desired. Advantages of Each Model. The results of two different models are presented because they are each sensitive to dif- ferent factors. The regression model will present different results based on the type of carrier and whether service is provided to an airportâfactors that cannot be considered in the trip rate model. In addition, as a statistical equation, a Confidence Interval can be calculated and presented to help put bounds on the range of potential ridership estimates. The trip rate model is driven only by the population of the stops that are served, so there is an estimated demand for each stop. This can be modified, if the user has additional information that could affect the potential demand arising at that location. Perhaps there is a large university in one stop, not reflected in the population count, or a stop may also be served by other intercity bus routes, potentially dividing the market for intercity bus trips from that location. For that reason, the Output page provides for a link to a page allowing for manual adjustments to the trip rate model estimate. Potential adjust- ment techniques are presented in the next step. Step 5: Adjustments to the Trip Rate Model While the user can accept the results of the two models at this point, there are some circumstances in which it may be advisable to take advantage of the flexibility of the trip rate model to incorporate additional information to improve the accuracy of the estimates. There are two situations in which further adjustment is most advisable: (1) if there are locations on the route (other than the destination city) that have other intercity bus service and (2) if there is competing service on the same route as that proposed. In addition, the analyst may wish to modify the trip rate estimates for a particular stop based on additional information about institutions that could generate intercity bus ridership, such as a university or a military base. Additional information on possible adjustments can be found by clicking on the âAdjustment Worksheetâ button on the Output page. User Inputs Locations Served The user selects the locations that the proposed route will serve from the drop-down list. The user can scroll through the drop-down list to select the city or begin typing the name of the city in the box and it will auto-fill. Some city names ap- pear in more than one state, so the user should make sure that the selected city is in the desired state. For example, there is an Aberdeen in Maryland, Mississippi, South Dakota, and Washington. 51 Regression Model Trip Rate Model with Adjustment Within 50% of actual ridership 59.6% 54.4% Within 10% of actual ridership 17.5% 15.8% Within 5% of actual ridership 5.3% 5.3% Table 6-1. Accuracy of regression and trip rate models.
The cities listed are those that are recognized by the 2000 Census. If the user does not find a city or town in the drop- down list, it is either: ⢠Not recognized in the 2000 Census or ⢠Part of an urban cluster or urbanized area that has a different name. If the city or town is part of another urban cluster or urban- ized area, then select the urban cluster or urbanized area that includes it. In Figure 6-1, Town Z overlaps Urbanized Area A. Therefore, Town Z will not be listed in the drop-down list and Urbanized Area A should be selected as the location. Population The population will automatically be generated for each of the places selected in the Locations Served column. The population figures are from the 2000 Census. Route Length The route length is the one-way length of the route (in miles). The user will need to calculate this beforehand and enter it into the input sheet. Definition of Intercity Bus (ICB) Service For the purposes of this toolkit, intercity bus service is considered to be rural intercity routes provided in the tra- ditional intercity model, generally with low frequencies (one daily round-trip or less), comparable distance-based fares ($0.10 to $0.17 per passenger-mile), interline ticketing (through the National Bus Traffic Association), information about connections through national bus information systems (Russellâs Guide, Greyhound telephone/on-line information, etc.), and generally operated by private for-profit firms. Generally, if the toolkit user anticipates that the likely operator could be Greyhound, a member of the Trailways National Bus System, or a regional provider of regular-route intercity bus services such as Indian Trails, Jefferson Lines, Black Hills Stage Lines/Arrow, Southeastern Stages, Coach USA, Coach America, etc. then âICBâ should be checked on the input page of the toolkit. The ICB designation affects only the results of the âRegres- sionâ model. If âICBâ is checked, the predicted ridership will be higher, likely reflecting the impact of the additional con- nections possible through this network and the information availability for inbound passengers from outside the region. The following types of service are not considered intercity bus service (in terms of the data used to construct the toolkit): ⢠Local rural transit routes connecting to other rural intercity routes that do connect to the national intercity network are not considered intercity in themselves as they do not connect directly to the national intercity bus network. With multiple transfers to reach the national network, it is unlikely that many of the passengers on these connecting local services are making intercity trips, even if such linkages are technically possible. ⢠Intercity routes that do not serve rural intermediate points but have a majority of the stops located in the urbanized areas of the origin and destination endpoints do not nec- essarily establish rural intercity service because the service is not accessible to residents of the non-urbanized area. They are not included in the definition of rural intercity services. ⢠Commuter services, defined as peak-hour, peak-direction weekday services, are not considered rural intercity services. Airport The user should select the âAirportâ box if the proposed service will be serving one or more airports with scheduled commercial air service. The proposed route can offer direct service to the airport terminal or it can offer a rider the oppor- tunity to reach the airport with a single transfer. If the proposed route would serve a terminal that has scheduled service to the airport such that a passenger on the proposed intercity service could reach the airport with a single transfer, the user may select this box even if the proposed intercity bus service does not directly serve the airport. The user should select âAirport Serviceâ in situations similar to these: 52 Figure 6-1. Hypothetical example of overlapping Census areas.
⢠Proposed Service: City A to F, via B, C, D, and EâThe proposed intercity bus service terminates in City F at the intermodal terminal or major transfer point that is the hub of the local transit system. A local transit route originates at the intermodal terminal and terminates at the City F Regional Airport. ⢠Proposed Service: City W to City Z, via X and YâThe pro- posed intercity bus service between these points stops at the Greyhound station in City Z and continues on to a final scheduled stop at the City Z International Airport. The user should not select âAirport Serviceâ if the situation is similar to these: ⢠Proposed Service: Green Valley to Capital CityâGreen Valley has no airport, and in Capital City the route would end at the intercity bus station. The station is served by a local bus route, Route A. Route A crosses Route E near downtown, allowing a transfer. Route E services the Capital City Airport. An intercity bus passenger trying to reach the airport would have to transfer twiceâonce from the inter- city bus to Route A and a second time from Route A to Route E. ⢠Proposed Service: Green Valley to SpringfieldâGreen Valley has no airport, and in Springfield the intercity bus station is not on a fixed-route, fixed-schedule transit route. An intercity bus passenger trying to reach the Springfield Inter- national Airport would have to take a taxi from the intercity bus station. There is no transit connection to the airport. ⢠Proposed Service: Green Valley to Stone MountainâGreen Valley has no airport, and the airport in Stone Mountain does not have commercial air service, but only air taxi, charter, or executive aircraft. In developing and testing rural intercity bus alternatives, the planner can try running the model with the same inputs selecting âAirport Serviceâ one time, and then not selecting it the next, to get a sense of the potential additional ridership that would result. If the additional ridership is significant, it may warrant designing the service to provide that connec- tion, either on the proposed route or by making sure that a convenient local connection is available. Correctional Facility The user should check the âCorrectional Facilityâ box if the proposed route serves a town or city with a state or federal correctional facility. Correctional facilities generate inter- city bus demand from relatives of inmates who may take the bus for visitation and from release of inmates. In the case of state or federal facilities, both types of trips may involve intercity travel, as inmate populations are not local. Do not check the box if the route serves only local municipal or county jails, as the resulting travel demand is likely to be local in nature. Information on federal prison types, populations, and locations is available from the Federal Bureau of Prisons website (www.bop.gov). That website provides information on the characteristics of the different prison typesâin general the minimum security institutions known as Federal Correc- tional Institutions (FCIs) have small populations and would not warrant checking the box to indicate that the proposed route provides service to correctional facilities. State correc- tional facility location, type, and population data for most states is available from individual state websites. In planning rural intercity bus routes, the presence of a state or federal correctional facility may warrant a stop in that town or affect the choice of routing to provide service to that location. Adjustments to the Trip Rate Model While the user can accept the results of the two models at this point, there are some circumstances in which it may be advisable to take advantage of the flexibility of the trip rate model to incorporate additional information to improve the accuracy of the estimates. There are two situations in which further adjustment is most advisable: one, if there are locations on the route (other than the destination city) that have other intercity bus service, and two, if there is competing service on the same route as that proposed. In addition, the analyst may wish to modify the trip rate estimates for a particular stop based on additional information about institutions that could generate intercity bus ridership, such as a university or a military base. Additional Intercity Bus Service Adjustment The initial step in the process called for an assessment of existing intercity bus services at the locations to be served on the proposed route. Based on that inventory, the analyst should be aware of locations that have additional intercity bus routes. If a location has additional intercity bus services, the estimated demand for that location should be allocated between the proposed route and the other existing services. If there is no available data on the ridership of the existing routes at that point, an allocation can be made by: ⢠Determining the number of daily departures on all existing routes from available timetables, ⢠Adding the number of daily departures on the proposed route to create a total number of daily departures from that location, 53
⢠Dividing the number of daily departures on the proposed service by the total to get the percentage of the total inter- city bus departures that would be on the proposed service, and then ⢠Multiplying the trip rate modelâs predicted ridership for that stop by the percentage of the total departures repre- sented by the proposed service, and inserting that number into the appropriate blank on the âAdjustmentsâ page. If there is existing service on the proposed route that would directly compete, the best adjustment would be to subtract the ridership on that service from the estimated demand. However, if ridership data for the existing service is not avail- able, the technique presented above can be applied at each of the origin stops to allocate the predicted ridership at that stop between the two services based on the relative frequency. Adjustments for Additional Generators of Intercity Bus Ridership If a proposed stop has a potential generator of intercity bus trips such as a college or university, military base, correctional facility, or other population that either is not reflected in the Census population estimates or is likely to have a higher propensity to use the bus (because of lower auto ownership, lower incomes, parking restrictions, higher frequency of trips, etc.), the analyst can use the âAdjustmentsâ page to manually add trips estimated through other means. This process would be part of the overall planning effort for a rural intercity routeâ the toolkit âAdjustmentsâ page simply provides a place for such additional data to be entered into the overall demand estimates. Possible ways of developing such ridership adjustments vary with the type of potential ridership generator: ⢠University or College: Data on the overall undergraduate student enrollment can be obtained from the College Board website (collegesearch.collegeboard.com/search/adv_typeof school.jsp). If possible, determine what percentage of the undergraduate student body lives on campus, as commuters from the immediate vicinity are not likely to need intercity bus service to visit home. Multiply the undergraduate population living on campus times an assumed number of annual long-distance trips (one round-trip for fall break, one round-trip for Thanksgiving, one round-trip for winter break, one round-trip for spring breakâassuming that moving in for fall term and moving out at the end of spring will not be by bus). Many students have cars, so not all these trips will be by bus. Unless there are known campus policies that prevent undergraduates living on campus from having personal vehicles, multiply the estimated total trips by a conservative bus mode share of 2 percent to get an estimate of the potential additional ridership at that stop, which can be added to the model estimate and inserted on the âAdjustmentsâ page. ⢠Military Base: The National Park Service provides a map of military bases in the contiguous United States (www.nps. gov/nagpra/documents/basesmilitarymap.htm), which is accompanied by an index. This map can be used to identify military bases in a particular state. Once identified, addi- tional information on the location, and often the number of persons stationed at that location, is generally available from the baseâs own website. A planner seeking to add mil- itary base populations to the demand estimate would likely need to obtain additional information about the particular transportation needs at a base, generally by contacting staff at the base. ⢠Correctional Facilities: Correctional facilities generate intercity bus trips from visitors and from released inmates. In planning rural intercity bus routes, the presence of a state or federal correctional facility may warrant a stop in that town or affect the choice of routing to provide service to that location. Information on federal prison types, pop- ulations, and locations is available from the Federal Bureau of Prisons website (www.bop.gov). That website provides information on the characteristics of the different prison typesâin general the minimum security institutions known as Federal Correctional Institutions (FCIs) have small pop- ulations and would not warrant further efforts to estimate demand. State correctional facility location, type, and pop- ulation data for most states is available from individual state websites. Again, the planning process may warrant efforts to contact staff at the correctional facility to deter- mine the numbers and timing of inmate releases, and any information that may be available regarding the need for intercity bus services for visitor needs. In some cases cor- rectional facilities operate their own van services for these purposes and may have data that can be used to improve the overall demand estimate. Example of Model Application with Adjustmentsâ Baltimore, Maryland, to Morgantown, West Virginia A planner working for the Maryland Transit Administration has performed a statewide assessment of existing intercity bus needs and noted there is no longer any intercity bus service in the state west of Hagerstown, Maryland. For planning purposes an estimate of potential ridership is needed. The planner uses an old Greyhound schedule to identify places that were formerly served and also reviews some demographic data from the statewide assessment to identify a set of poten- tial stops. The planner uses the toolkit to estimate potential ridership: 54
1. To make sure that this toolkit is appropriate, some potential service parameters are checked: ⢠Proposed route length: 210 miles, one-way ⢠Proposed daily frequency: one round-trip, 365 days per year ⢠Proposed fare level: assume standard NBTA fare levels, approximately $0.13 per mile ⢠Proposed operator: an intercity bus carrier member of the NBTA ⢠Proposed connectivity with the national intercity bus network: service into the Baltimore Greyhound station, the Frederick MARC/Greyhound station, the Hager- stown Greyhound station, and the Morgantown Moun- tain Transit transfer point All of these elements are comparable to the routes used to develop the toolkit, as indicated by the informa- tion on the links shown on the CD. 2. The trip rate and regression models are both utilized. Figure 6-2 presents the data input screen with the selected stops. The planner does not have to look up any data at this point; all of the population data is included on the CD. 3. Figure 6-3 presents the initial output screen with results for the regression model and the trip rate model. Both models use population data, but the regression model is a corridor-level model, and it also reflects the type of provider and whether an airport is served. In this case the planner assumed an intercity bus provider, but no airport connection. 4. Also in Figure 6-3, there is a box that presents ridership for the four routes in the database that are closest to the proposed route in terms of basic characteristics. As can be seen, the two model estimates are comparable. 5. Adjustments: a. Other Intercity Bus Service: Some of the stops shown have other intercity bus service, so it is likely that not all of the intercity bus ridership generated at those points will be using just this route. The model processing elim- inates the Baltimore population from the calculation. To make additional adjustments to reflect this alternative 55 Figure 6-2. Data input screen with selected stops.
service, the planner goes to the Greyhound link and finds that Hagerstown has three existing intercity bus sched- ules per day each way, and the proposed route would add a fourth. So the proposed route might capture 25% of the estimated Hagerstown demand. This demand is shown in the Adjustment Worksheet (Figure 6-4), which shows ridership by stop for the trip rate model (the regression model cannot be adjusted in this manner). The same situation is true for Frederick, so a similar adjustment is made to the estimated ridership for that stop. A new sum total estimated ridership is provided in the Adjustment Worksheet table. b. Destinations: In this case one key destination being served is Frostburg State University in Frostburg. This is one of Marylandâs larger public universities, with a significant residential population. According to the College Board website, the student population at Frostburg is 4,755. The Census population of 10,916 should include the student population, but it is likely that the usage would be higher than that of a town of similar population without the university. To adjust upward for the University, one might apply different rates to the student population and the town population. Subtracting the 4,755 students from the overall population leaves 6,161, times the trip rate of .055 equals 339 annual trips. Applying the adjust- ment factors to the 4,755 students, if 2 percent of each student takes eight intercity bus trips per year (as sug- gested in the section of the toolkit on Adjustments), this results in an estimate of 761 student bus trips. Added to the 339 town-based trips, the estimated total becomes 1,100, rather than the original 390. This amount is entered in the Adjustment Worksheet as well. c. Combined AdjustmentsâOther Service and a Key Destination: Morgantown is the home of West Virginia University with 21,720 students (according to the College Board website), and it also has Section 5311(f) intercity bus service linking it with Clarksburg and Pittsburgh, Pennsylvania. Mountain Line Transit Authority operates 56 Figure 6-3. Initial output screen with results for the regression model and the trip rate model.
two round-trips per day, so a third intercity connection might potentially receive a third of the trips generated by the overall population. Applying the same student body adjustment as for Frostburg State, the estimated West Virginia University ridership would be 3,475 (2 percent times eight trips per year), added to estimated town ridership of 1,885, for a total of 5,360 annual trips instead of the 2,010 estimated on population alone. Then, apply- ing the adjustment for the additional intercity bus service, the proposed route would receive a third of the estimated total trips, or 1,786 trips. This amount is entered man- ually on the Adjustment Worksheet. d. Figure 6-5 presents the estimated adjusted ridership for the route, which is 6,600 annual trips. Comparables/Route Descriptions The data table used for the lookup of comparable routes and the descriptive information for each route are on the CD. Other Issues That Were Addressed in the Final Version of the Toolkit The final version of the toolkit included some additional user information identified by the project team: ⢠It is clearly stated that this set of tools provides an âinitial planning estimateâ of potential demand and that a full demand analysis requires the adjustment process described in the toolkit and likely some additional expertise or local knowledge on the part of the user. ⢠A warning states clearly that neither technique includes or addresses ridership that might arise from the location of a route in a broader network. The actual ridership on an intercity bus route includes both the riders generated at the stops served by the route, and the additional riders that are fed onto that route from other locations on the network. If a route connects other routes so as to reduce travel time or make a connection that did not previously exist, it may 57 Figure 6-4. Adjustment Worksheet.
generate ridership that has nothing to do with the popula- tion along the route. The ridership on these âbridgeâ routes will be a function of the connections at either end and could only be estimated with a general national intercity bus network model or by obtaining data from the operators at either end regarding the likely feed traffic coming from their connecting services. A route with little or no population will have a very low predicted ridership in the toolkit, but if it plays a particular role in the network it might well have much higher ridership than predicted. ⢠Some discussion is included regarding the two different models, their differing results, and which factors a user might consider in selecting one over the other. The regres- sion model incorporates population, but also includes the type of carrier and airport connections as factors. The trip rate model is entirely population based and really must be used as the base for making the kinds of adjustments sug- gested in the toolkit, as these are done on a stop-by-stop basis. ⢠A technique or warning has been included regarding routes that link two large population centers. The model applica- tion in the toolkit defines the largest population city on the route as the âDestinationâ and does not include that pop- ulation in the estimation of demand, as those places typi- cally have a lot of additional intercity bus service. However, if there are two large population centers, the lower popu- lation center would still be included and the demand could be overstated considerably. 58 Figure 6-5. Estimated adjusted ridership.
