Long-Distance and Rural Travel Transferable Parameters for Statewide Travel Forecasting Models (2012)

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.

F-1 This appendix covers the following topics related to each statewide model: • Likely Ranges of Parameter Values—Describes sources used for model parameters in existing models, those that are transferable and those that are unique to certain states, and describes the rationale (e.g., geography, population density, available transportation modes, proximity to primary tourist destinations, etc.); • Age of Models—Includes information about when the model was developed, base and fore- cast years available, and current status of the model; • Purposes for Which Models are Applied—This discussion goes beyond the rationale for statewide models to identify how, when, and why each statewide model is used; and • Frequency of Application—It is also important to understand how frequently each model is used and updated, and what level of coordination is done to keep the model current with new assumptions found in urban and regional MPO models. This appendix continues with a state-by-state assessment of rural and long-distance travel from available technical documentation, largely focused on the likely ranges of parameter values. This is followed by a discussion of model applications and uses derived from informal telephone and e-mail contact with state departments of transportation (DOTs) focused on age of models, purposes for which models are applied, and frequency of application. This appendix ends with a summary of conclusions and limitations gleaned from this statewide model document and DOT state review and assessment. F.1 Likely Ranges of Rural and Long-Distance Parameter Values in Statewide, Multistate, and National Models In many cases, states have incorporated methods for forecasting long-distance trips along with shorter regional trips in their statewide models. In most cases, statewide models incorporate truck and auto long-distance trips; however, in some cases, additional modes are incorporated such as air and intercity transit. The threshold for defining long-distance trips also varies among state- wide models, with some states considering trips over 100 miles to be long distance and others considering 50 miles or 75 minutes as long distance. The following discussion is a state-by-state rundown of how long-distance and rural trips are incorporated into their respective statewide models. States without long-distance or rural trip purposes are not included in this discussion. It should be noted that statewide model documen- tation for Arizona, Maryland, and Ohio includes coefficients and constants used in long-distance trip modules; however, it was not believed that these parameters would be transferable in the absence of additional background information on the structure of these models. A p p e n d i x F Review of Statewide Models

F-2 Long-distance and Rural Travel Transferable parameters for Statewide Travel Forecasting Models Arizona The Arizona statewide model (SWM) estimates both long-distance (LD) personal trips and long-distance truck trips. LD personal trips are defined as those trips 50 miles and greater, while the LD truck trips are defined as regional truck flows made between metropolitan areas. The 2009 NHTS/Arizona Add-On was used as a primary data source for model parameters. Although the 2009 NHTS does not include a long-distance survey, it does provide some data on trips longer than 50 miles. As part of the Arizona SWM, long-distance trip data to and from Arizona for all 50 states were included. According to model documentation (Parsons Brinckerhoff, 2011), when the NHTS was conducted, 15 states (including neighboring New Mexico) and Washington, D.C., were left out of the survey. The report states that the NHTS reports origin state for states with a population of 2 million or more and the missing records for smaller states needed to be gener- ated. Due to this omission, records were synthesized by Arizona DOT consultants for each of the missing states based on the records from neighboring states with survey data. Once records were synthesized for the missing states, a regression analysis was performed to obtain nationwide totals for long-distance trips, and three independent variables were tested: population; percent of employment in the service sector; and gross domestic product per capita. The population vari- able had the strongest correlation, and therefore the other two variables were not used. Special generators were also used for major tourist attractions such as the Grand Canyon and other national parks in the state. NHTS data were deemed insufficient for these areas, and instead, data from the parks on visitor travel were developed and used in the long- and short-distance models. California The California High-Speed Rail model is a combination of intraregional and interregional models (Cambridge Systematics, Inc., and Mark Bradley Research and Consulting, 2007). The interregional model is further segmented into short trips (less than 100 miles) and long trips (greater than 100 miles). The 1995 ATS was used to validate the long-distance trips, CTPP was used for short and long commute trips, and the California Statewide Travel Survey was used for short-distance trips. The ATS was used to determine intra-California trips over 100 miles in length. Since the ATS data were from 1995, a 6.9 percent compounded annual growth factor was applied to estimate base year 2000 figures. Applying the growth factor to ATS trips resulted in an estimate of about 350,000 long-distance daily trips in the year 2000 for California. The auto occupancies for short and long trips were about the same, at 1.35 and 1.34, respectively. No average trip lengths were provided. Florida The Florida statewide model is primarily used to forecast long-distance truck trips and inter- city automobile travel. Although freight is modeled outside Florida, passenger trips are only modeled within the state. The average long-distance business trip length is 127.2 minutes in the latest base year 2005 model (BCC Engineering, 2011). Matrix estimation was used to generate long-distance tourist trip tables, for which average trip lengths were not documented. Georgia The Georgia statewide model has the ability to generate external through trips and internal- external trips (Atkins, 2011). The freight model relies on TRANSEARCH data and the passenger model is based on the 2009 NHTS and the Georgia NHTS Add-On. Passenger model trips were separated into two categories, short-distance and long-distance trips. The short-distance trips are considered to be less than 75 minutes and within urbanized areas, and the long-distance

Review of Statewide Models F-3 trips are greater than 75 minutes and interurban. The Georgia passenger model estimated 183,000 internal long-distance trips and 235,000 internal-external long-distance trips for a total of 418,000 average daily base year 2006 trips or 1.34 percent of all trips in the model. The average long-distance trip length is 131.67 minutes. Home-based other long-distance trips are longest at 140 minutes, followed by nonhome-based LD at 135 minutes, and home-based work LD at 120 minutes. Indiana The statewide model network used was a refinement of the I-69 Indiana Statewide Travel Demand Model network (Bernardin, Lochmueller & Associates, Inc., and Cambridge Systemat- ics, Inc., 2004). Data from the 2001 NHTS and the 1995 Indiana Household Travel Survey were used. Long-distance trips were generated as a separate purpose during trip generation. Truck trips were generated separately. There was no indication of the threshold used to classify the trips as long-distance; however, the documentation stated that “long-purpose trips included internal zones in border states.” Louisiana The Louisiana statewide model uses a different approach from most other models. The Louisiana model focuses mainly on urban versus rural trips using micro and macro submodels (Wilbur Smith Associates, 2004). The micro model focuses on mainly urban trips that are shorter distances and the macro model estimates long-distance intercity trips on rural highways. Accord- ing to the 1995 ATS, 45 percent of long-distance auto trips in Louisiana were intrastate and the rest were interstate. Maryland The Maryland Statewide Travel Model (STM) was developed with a three-tier geographical approach: regional, statewide, and urban (Parsons Brinckerhoff, 2009). The regional model net- work covers the full United States, Canada, and Mexico. The statewide model includes Maryland and its border states, and the urban submodel focuses on the major urban areas within the state and is comprised of the Baltimore and metro D.C. models. The regional submodel includes a vis- itor model and long-distance person module to simulate long-distance trips. The long-distance person submodel handles trips that originate in Maryland. The long-distance trip model is dif- ferent from traditional models in that it uses microsimulation. Initially, long-distance travelers are individually generated and then trip information is applied to get the final long-distance trip estimated. The trips are based on data from the 2005 NHTS and use origin, destination, mode, party size, income, and time of day to simulate trips. Destinations are provided by state, not zone; therefore, if the destination is within one of the surrounding states in the statewide model, the model randomly assigns the resulting trips to a zone based on population and employment factors. Train, bus, and air travel modes are simulated for long-distance travel. The model appor- tions non-auto trips to either a station or airport, but does not include the egress travel to the final destination. Departure times are assigned a time of day (a.m. or midday, peak or off-peak). No nighttime long-distance trips are simulated due to the small percentage of these trips. The visitor travel submodel simulates trips that have an origin outside of Maryland and a destination within Maryland. The visitor submodel works similarly to the person long-distance submodel, but in the reverse travel direction. In 2000 there were 49,355 daily long-distance travelers exiting the state and 46,733 daily visitors entering the state as estimated in the Maryland statewide model.

F-4 Long-distance and Rural Travel Transferable parameters for Statewide Travel Forecasting Models Mississippi The Mississippi statewide model operates in a similar manner as the Louisiana statewide model (Wilbur Smith Associates, 2002). It was created to complement the urban models that already existed and focus on interurban travel. The main sources of data were the 1995 ATS and 2001 NHTS. The long-distance trips were considered to be 100 miles and greater. Friction factors were calibrated to observed trip lengths in the NHTS. Ohio The Ohio statewide model uses a tour-based approach with separate short-distance and long- distance models that are later combined (Parsons Brinckerhoff, date unknown). The long-distance model is based on a long-distance travel survey that was administered by the state and only looked at trips greater than 50 miles. The Ohio model is a tour-based model and since most long-distance trips occur over multiple days, the Ohio model uses a 2-week window instead of a single day, meaning the trips either started and/or ended within this window of time. The beginning tours were given a departure time, ending tours received an arrival time, and complete tours had both departure and arrival times. Although the model does indeed predict the incidence of travel over a 2-week period (in its first step), later, the amount of that travel occurring on the “model day” is predicted as either starting, ending, traveling, or “not on this date,” with “not on this date” occurring close to 80 percent of the time (i.e., the person did a long-distance trip at some time in the 2-week window, but not on the “model day”), hence comparable daily travel statistics can be generated. Utah The Utah Statewide Travel Model (USTM) does not include a separate model for long-distance trips but does estimate LD trips by two purposes: work and other (Wilbur Smith Associates, 2009). The data source used for these trips was the 2001 NHTS. Daily long-distance trip genera- tion rates were estimated at 0.0310 per household for LD home-based work trips and 0.0532 per household for LD home-based other trips. When calibrating LD trips, an average trip length of 120 minutes was used for long work trips and 80 percent of the 120 minutes was used for nonwork long trips. The 120 minutes was used because it is assumed that any urban area in the state can be reached within 120 minutes. The validated LD trip lengths were 89.54 minutes for work and 81.73 minutes for other long-purpose trips. The average auto occupancy rates were 1.33 and 2.06 for work and other, respectively. Virginia The Virginia model was developed similarly to the Louisiana and Mississippi models (Wilbur Smith Associates, 2004). The 1995 ATS was the primary data source for long-distance trip tables. Long-distance trips were defined as those trips longer than 100 minutes, which is assumed equiva- lent to 75 miles, and differs from the ATS 100-mile definition. The average trip lengths reported were 303.73 minutes for interstate trips and 126.13 minutes for intrastate trips. The long-distance auto occupancy rate was 1.82 for business trips and 2.69 for tourist and other long-distance purposes. Wisconsin The Wisconsin statewide model focuses on long-distance trips in and through the state that are 50 miles and greater (Cambridge Systematics, Inc. and HNTB, 2006). The purpose was to get a handle on intercity travel for different travel modes. The model focuses on trips that originate or end in Wisconsin or its surrounding states. Trip rates were estimated from the 2001 NHTS

Review of Statewide Models F-5 Daily Trip File and the 2001 NHTS Long-Distance Trip File, and transit travel characteristics were gathered from various transit authorities. When modeling the trips, the initial focus was on understanding the difference between recurring trips and nonrecurring trips. These trips included all internal-internal and internal-external trips. The external-external trips were more difficult to model due to a lack of available data. The state was divided into five regions with different trip generation rates to reflect each region. A destination choice model was used to distribute intercity flows and, from this, the mode choice model was applied. The modes included auto, bus, and rail. Overall, 97.69 percent of the long-distance trips estimated were auto, with 1.95 percent and 0.35 percent for bus and rail, respectively. F.2 Age, Frequency, and Purpose of Statewide Model Use and Assumptions for Rural and Long-Distance Travel Table F.1 depicts information obtained by e-mail request from several state DOTs regarding their statewide models. This information is intended to supplement what is described in avail- able technical reports and includes the age of models and assumptions, purposes for which statewide models are applied, and frequency of statewide model application and model updates. If the nine responses provided in Table F.1 are considered as representative of the full range of statewide travel demand models, there is considerable range in the age of models and support- ing data, model applications, and frequency of model use and updating. In terms of common findings from this dialogue, it does appear that most state DOTs plan on continuing the process of updating their statewide models every 5 to 10 years. Oldest data assumptions include reliance on the 1995 ATS and Census 2000. This informal survey found a wide range of uses for statewide models and indicates that state DOTs are making regular use of their statewide models. Not surprisingly, the most common uses of statewide models include statewide transportation plans, corridor studies, air quality conformity, freight planning, and providing performance statistics for other reporting purposes. Perhaps the most significant conclusion from this analysis is that statewide models will con- tinue to be updated and will clearly need new and updated data sources that help quantify trip generation and distribution parameters. States still largely rely on 1995 ATS and 2001 NHTS data to identify characteristics of long-distance trips, although several states have conducted their own surveys of long-distance and rural trip-makers or participated in the 2009 NHTS Add-On. The types of studies being conducted through use of statewide models should also help focus on the most important components and related parameter needs and accuracy levels. F.3 Conclusions A review of statewide models, discussions with state DOT staff, and available BTS analysis of long-distance trips from the 2001 NHTS provide a few common findings with respect to long- distance travel, as follows: • Long-distance travel thresholds vary considerably by model, ranging from 50 to 100 miles and sometimes using minutes of travel as the breakpoint instead of miles traveled; however, it is fair to say that most statewide models distinguish long-distance trips from short-distance trip purposes found in urban area models. • Transferable parameters and statistics for long-distance travel were sparsely documented based on this review of statewide model technical reports.

F-6 Long-distance and Rural Travel Transferable parameters for Statewide Travel Forecasting Models State Age of Statewide Model and Assumptions Statewide Model Applications Frequency of Statewide Model Applications and Updates Arizona • 1st generation (2009) • 2nd generation (2011) • 3rd generation (just started) • Oldest assumption – 1995 ATS • Statewide Framework Visioning Studies • Phoenix-Tucson Intercity Rail Study (3rd generation) • Central Arizona Association of Governments (CAAG) Long- Range Transportation Plan (3rd generation) • Model still in its infancy but will identify more uses when done • Anticipate annual routine updates, as new data available Florida • 1988 base year model (1989-1993) • 1990/2020 (1995-1998) • 2000/2030 (2003-2007) • 2005/2035 (2007-2011) • Oldest assumption – 2000 era tourist data • Corridor studies • Corridor master plans • New corridor initiative • Freight flows • Subarea extraction • Input to decision tool • Strategic Intermodal System • Model is applied for internal purposes on a regular basis • Routine annual updates and major updates every 5 years Iowa • 2005/2035 (2009), socioeconomic (SE) data available in 5-year increments • Oldest assumption – 2005 population, employment, and traffic counts • Bypass analyses • Passenger rail demand • Airport drive time • Detour analyses • Statewide regional corridor scenario analysis • Supplemental traffic forecasts • MPO external forecasts • Rest area analysis • Snow plow optimization • River crossing closures • Rural Interstate Interchange Justification • Model applications: one significant project every 3 months • Minor projects several times per week • Model updates planned for every 5 years Michigan • 1990/2020 (1996) • 2000/2030 (undocumented) • 2005/2035 (2010) • 2008/2035, 2040 (2012) • 2010/2040, 2045 (2014) • Above model SE data typically available in 5-year increments between base and horizon years • Statewide VMT Forecast • Corridor Studies • Air quality conformity • Work Zone Safety and Mobility Analysis • Detour Evaluation • Small Area Models • Environmental Justice Analysis • Economic Impact Analysis • Project Alternatives and Project Selection • Select Link Analysis • Deficiency Analysis • Congested speeds and travel times • Economic analysis of 5-year plan – annually • Air quality conformity – a few times a year • Work zone analysis – 6 to 10 requests per year • Growth rates – 5 to 10 requests per year • Proximity analysis – annually • Growth rates – < 10 per year • Detour evaluation – approximately 1 per year • Select link analysis – approximately 2 per year • Small area models – approximately 1 every 3 years Table F.1. State department of transportations’ use of statewide models.

Review of Statewide Models F-7 State Age of Statewide Model and Assumptions Statewide Model Applications Frequency of Statewide Model Applications and Updates Ohio • Final V1 “model” 2003 • Updated to V1.1 2007 • Final V2 model 2009 • V3 anticipated by 2013 • Corridor/bypass studies • Tolling studies • Project design traffic • Project ranking process • Economic impact analysis • Long-range planning • Strategic planning • Freight planning • Grant applications • Air quality conformity • Model is applied on a regular basis • Major updates every 3 to 5 years Oregon • First: 1990 base year model (1999) • 1998 recalibration • Newest: 2000 (2009) • Underway: 2009 base • Oregon Transportation Plan • Oregon Freight Plan • Oregon Bridge Study • Willamette Valley Alternative Futures • Newberg-Dundee Bypass • New freeway in central Oregon • Land use – transportation • Model is applied on a regular basis • Routine annual updates and major updates every 5 years Tennessee • 2003/2030 (2005) • Planned update (2011-2012) • State Long-Range Transportation Plan • I-40/I-81 corridor study • I-75 corridor study • Rural Planning Organizations • Subarea analysis • Only updated as needed to conduct model applications (e.g., network corrections, etc.) Utah • 2007 base year (July 2009) • 2010 base year (2013-2014) • UDOT Long-Range Transportation Plan • External model refinement for urban models/studies • 2008 Baseline data for Envision Utah Study • Model is applied on an as-needed basis • Plan to update in 2013-2014 timeframe using statewide household survey just started Virginia • Base year 2000 model (2005) • Forecast year of 2025 • Model update initiated in 2008 but effort shelved due to staff reductions • I-81 corridor study • Dependent on funding and staffing availability Table F.1. (Continued). • Statewide model updates are being planned by most state DOTs, showing the need for more updated data sources for calculation of model parameters, especially with continued reliance on 1995 ATS and 2001 NHTS for assumptions on long-distance trips. • The vast majority of long-distance trips are less than 500 miles on average. • Long-distance mode of travel is impacted by type/purpose of trip, income, and geography, with business trips, higher incomes, and urban geographies showing a greater likelihood of air travel than their counterparts. Relatively little information was available from these same sources on rural trips, perhaps in part due to rural trips not being an isolated set of trip purposes (as with long-distance trips) and the treatment of rural trips being largely the same as urban trips and purposes in most statewide models.

Abbreviations and acronyms used without definitions in TRB publications: AAAE American Association of Airport Executives AASHO American Association of State Highway Officials AASHTO American Association of State Highway and Transportation Officials ACI–NA Airports Council International–North America ACRP Airport Cooperative Research Program ADA Americans with Disabilities Act APTA American Public Transportation Association ASCE American Society of Civil Engineers ASME American Society of Mechanical Engineers ASTM American Society for Testing and Materials ATA American Trucking Associations CTAA Community Transportation Association of America CTBSSP Commercial Truck and Bus Safety Synthesis Program DHS Department of Homeland Security DOE Department of Energy EPA Environmental Protection Agency FAA Federal Aviation Administration FHWA Federal Highway Administration FMCSA Federal Motor Carrier Safety Administration FRA Federal Railroad Administration FTA Federal Transit Administration HMCRP Hazardous Materials Cooperative Research Program IEEE Institute of Electrical and Electronics Engineers ISTEA Intermodal Surface Transportation Efficiency Act of 1991 ITE Institute of Transportation Engineers NASA National Aeronautics and Space Administration NASAO National Association of State Aviation Officials NCFRP National Cooperative Freight Research Program NCHRP National Cooperative Highway Research Program NHTSA National Highway Traffic Safety Administration NTSB National Transportation Safety Board PHMSA Pipeline and Hazardous Materials Safety Administration RITA Research and Innovative Technology Administration SAE Society of Automotive Engineers SAFETEA-LU Safe, Accountable, Flexible, Efficient Transportation Equity Act: A Legacy for Users (2005) TCRP Transit Cooperative Research Program TEA-21 Transportation Equity Act for the 21st Century (1998) TRB Transportation Research Board TSA Transportation Security Administration U.S.DOT United States Department of Transportation