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.
E-1 The documentation for NCHRP Project 8-36-B Task 91, Validation and Sensitivity Consider- ations for Statewide Models (Cambridge Systematics, Inc., 2010d) noted that a primary issue for statewide modeling of trucks is the consideration of freight. The Task 91 Final Report suggested using the Freight Analysis Framework (FAF), the national policy tool for freight analysis main- tained by FHWA, as the most comprehensive tool for considering freight issues. Version 3 of the FAF was released in July of 2010 and includes flows for a base year of 2007. The FAF3 highway network is available for download from the FHWA Website (FAF3 Net- work Database and Flow Assignment: 2007 and 2040, http://www.ops.fhwa.dot.gov/freight/ freight_analysis/faf/faf3/netwkdbflow/index.htm, accessed on July 20, 2011) as both a TransCAD network (FAF3 Network ESRI Format: faf3_1_1_esri.zip, http://www.ops.fhwa.dot.gov/freight/ freight_analysis/faf/faf3/netwkdbflow/network/esri/faf3_1_1_esri.zip, accessed on July 20, 2011) and as an ESRI shapefile (FAF3 Network TransCAD Format:faf3_1_1_TransCAD.zip, http://www. ops.fhwa.dot.gov/freight/freight_analysis/faf/faf3/netwkdbflow/network/TransCAD/faf3_1_1_ TransCAD.zip, accessed on July 20, 2011), with the actual loaded network volumes, congested speeds, and performances stored as a DBF file (FAF3 network Output:faf3_1_1_data.dbf, http:// www.ops.fhwa.dot.gov/freight/freight_analysis/faf/faf3/netwkdbflow/database/faf3_1_1_data.dbf, accessed on July 20, 2011), which can be joined to the network in either software format. The FAF3 network was examined to determine patterns of truck usage that could be reported for this project. E.1 Freight Analysis Framework Version 3.0 The FAF3 network includes all of the major highways in the United States. As shown in Table E.1, the FAF3 highway network includes virtually all of the center line miles in the three highest functional classifications for rural and urban areas, shown in the shaded rows. While differences do exist, characterizations by the authors of this report will be made at a system level and these minor differences between the FAF3 and official mileages are not considered to be important. It is inappropriate to load the FAF3 origin and destination flows of trucks with only the 123 domestic U.S. zones reported in FAF3. Doing so without disaggregating to smaller zones would incorrectly load only the principal highway routes connecting the centers of these regions. The FAF3 highway assignment, which was prepared for FHWA, first converted annual flows by truck into daily truck flows, using relationships between tonnage and truck and body type by commodity. Then, the origin-destination matrix was disaggregated from the 123 domestic zones to 4,609 network-specific zones consisting of county centroids, border crossings, major ports, intermodal terminals, and other significant truck generators. FAF3 used impedances on highway links that include preloaded auto and nonfreight truck volumes, developed in cooperation with A p p e n d i x e Urban Versus Rural Truck Trips
E-2 Long-distance and Rural Travel Transferable parameters for Statewide Travel Forecasting Models state DOTs, and assigned FAF truck flows using these impedances in a stochastic user equilib- rium traffic assignment. The resulting network flow, which includes calibration and validation against known truck volumes, is shown in Figure E.1. The FAF3 network consists of more than 170,994 links. Of these links, 150,344 had complete Functional Classification System and state identification information that could be used to sort and characterize the network volumes. The reported daily link volumes for total vehicles, FAF3 trucks, and non-FAF3 trucks were weighted by link length, aggregated by Census region, and reported separately for Interstates and other expressways and other arterials and local roads, for urban and rural areas. (FAF3 trucks report major commodity movements by trucks. It does not report movements of trucks that are the empty movement of trucks in support of freight, local delivery of freight, service trucks, construction trucks, utility trucks, etc. These trucks do contribute to road congestion and must be considered in any assignment of FAF3 trucks. They collectively are called non-FAF trucks in the FAF3 assignment.) The results of that analysis are shown in Table E.2. This analysis suggests that, on average, non-FAF truck usage is similar for urban and rural areas across all functional systems. The usage is slightly lower on Urban Interstates (5.5 percent) and is even lower on Rural Interstates (3.7 percent) but this is to be expected since these roads have widely spaced interchanges, particularly in rural areas, and would be less suitable for serving local truck trips. If the percentage of non-FAF truck flows is expressed as a percentage of total flows excluding FAF trucks, the similarity becomes much more evident. Excluding FAF trucks, non-FAF truck flows in rural areas, on average, on the Interstates and expressways would be 4.9 percent of the revised total in rural areas and 5.8 percent in urban areas; and on other road- way types would be 7.4 percent in rural areas and 6.2 percent in urban areas. It is also noted that non-FAF trucks on rural highways constitute approximately one-third of the volumes of those average volumes in urban areas both for Interstates (900 rural versus 4,000 urban) and for other roadways (400 rural versus 1,200 urban). This likely reflects the Center Line Miles Functional Classification FAF3 2007 Highway Statistics Rural Interstate 1 32,892 30,360 Other Principal Arterial 2 100,385 94,766 Minor Arterial 6 142,884 135,296 Major Collector 7 754 419,437 Minor Collector 8 34 262,899 Local 9 10 2,045,000 Urban Interstate 11 13,537 16,312 Other Freeways and Expressways 12 9,428 10,913 Other Principal Arterial 14 57,598 63,282 Minor Arterial 16 2,270 104,033 Collector 17 232 109,555 Local 19 56 740,273 Table E.1. Highway mileagesâFAF3 versus highway statistics.
Urban Versus Rural Truck Trips E-3 Weighted AADT Coefficient of Variation Geography Miles Total Trucks FAF Trucks Percent FAF Trucks Non-FAF Trucks Percent Non-FAF Trucks Total FAF Trucks Non-FAF Trucks Rural Interstates (FC=01) 32,892 24,500 7,000 6,100 24.9 900 3.7 4.77 5.44 8.45 Urban Interstates and Freeways (FC=11 and =12) 22,965 73,000 7,800 3,800 5.2 4,000 5.5 1.23 1.64 1.74 Other Rural Roads (FC 02 through 09) 244,067 5,700 700 300 5.3 400 7.0 8.34 15.79 10.86 Other Urban Roads (FC 14 through 19) 60,156 19,700 1,400 200 1.0 1,200 6.1 1.60 4.01 1.95 Table E.2. Average truck usage by functional system. Source: Battelle Memorial Institute, Network Assignment of Highway Truck Traffic in FAF3, PowerPoint presentation, FHWA Talking Freight Seminar, Freight Analysis Framework, Version 3, October 20, 2010. Figure E.1. Average daily long-haul volume on the National Highway System, 2007.
E-4 Long-distance and Rural Travel Transferable parameters for Statewide Travel Forecasting Models combination of two conflicting effects, where the activities that generate truck traffic are more numerous in urban areas, driving down the ratio, while the density of the road system is lower in rural areas, driving up the ratio. Table E.2 shows that, on average, FAF3 truck volumes on arterials and other local roadways are approximately equal: 200 trucks per day on these roadways in urban areas and 300 trucks per day on these roads in rural areas. While the volumes are considerably higher on Interstates and expressways, these numbers are also more divergent, on average, at 3,800 trucks per day on these roads in urban areas and 6,100 trucks per day on these roads in rural areas. It is suggested that the order of magnitude is similar if it is recognized that there are typically more paths uti- lizing Interstate highways between FAF regions through urban areas than through rural areas. The FAF volumes in Table E.2 suggest that there are approximately twice (6,100 versus 3,800) as many effective paths through urban areas, than through rural areas, using Interstates and other freeways. This seems like a reasonable relationship. Table E.2 also shows the coefficient of variation for the reported values. The coefficient of variation is the ratio of the standard deviation of the records, divided by the mean value. As shown, the variation as expected is higher in rural areas than in urban areas, but the values for the variation are similar for the three reported quantities (total volume, FAF trucks, and non-FAF trucks). This suggests that the reported averages, while informative, should only be considered generally applicable. Specific forecasts for individual roads would be preferable to averages. Table E.2 highlights that it would be preferable for any travel-demand forecasting model to include separate matrices of freight and nonfreight trucks. It is suggested that a commodity flow database at a disaggregated level be used to develop the freight truck table, or be used as a cali- bration database to develop the parameters for freight truck trip tables. This is consistent with the approach outlined in Chapter 4 of NCFRP Report 8 (Cambridge Systematics, Inc., 2010b). Table E.2 also suggests that the relationship between total vehicles and nonfreight trucks is similar in urban and rural areas. On average, the nonfreight truck volume is 6.1 percent of total AADT in urban areas and it is 7.0 percent of total AADT in rural areas. It is therefore reasonable to expect that whatever relationship exists between auto trips and activity generation in urban and rural areas will also exist for nonfreight truck generation in urban and rural areas. Thus if auto trips are found to be generated by households, and the auto trips per household in rural areas are 70 percent of the value of auto trips per household in urban areas, it is reasonable to expect that however the activity is generated for nonfreight truck trips in urban areas, that same indicator of activity should be used in rural areas, but the rate of truck trips per activity in rural areas should be 70 percent of the value found in urban areas. Table E.3 shows the same values as in Table E.2 reported by U.S. Census divisions, which are aggregations of states as shown in Figure E.2. The behavior among those Census divisions along both U.S. coasts as well as metropolitan areas along the Great Lakes, as shaded in light grey, would be expected to be similar and this is the case. The remaining Census divisions, which have substantially more rural area, are shaded in dark grey, and their behavior is similar. It is suggested that the differences in percentage of FAF traffic versus total traffic is largely a function of the larger values of non-FAF trucks and passenger cars in these denser, more metropolitan divisions. However, the variation between these Census divisions and the U.S. averages is not substantial. One notable exception is the average value for FAF trucks on rural Interstates in the New England Division. It is suggested that this is a function of geography where the travel between FAF regions does not have to pass through significantly rural sections of New England. Canadian traffic is assigned to New England counties on the border. Although the results of these intermediate steps are not shown, the assumption has to be that the low New England
Urban Versus Rural Truck Trips E-5 Table E.3. Average truck usage by functional system by Census divisions. (continued on next page) Geography Miles Weighted AADT Coefficient of Variation Total Trucks FAF Trucks Percent FAF Trucks Non- FAF Trucks Percent Non-FAF Trucks Total FAF Trucks Non- FAF Trucks Rural Interstates (FC=01) New England Division 993 25,400 2,900 1,800 7.1% 1,100 4.3% 3.30 3.50 3.87 Middle Atlantic Division 2,260 28,600 7,000 6,000 21.0% 1,000 3.5% 3.48 4.21 5.28 East North Central Division 4,524 26,900 8,000 6,900 25.7% 1,100 4.1% 3.56 3.98 6.50 West North Central Division 4,395 16,700 6,200 6,100 36.5% 100 0.6% 5.00 5.44 14.72 South Atlantic Division 4,430 39,000 8,200 6,400 16.4% 1,800 4.6% 3.22 3.60 4.62 East South Central Division 2,406 31,700 9,800 8,600 27.1% 1,200 3.8% 3.63 4.17 8.47 West South Central Division 4,003 24,500 8,500 7,600 31.0% 900 3.7% 5.03 5.41 9.13 Mountain Division 6,404 13,700 4,900 4,500 32.8% 400 2.9% 7.97 9.16 15.66 Pacific Division 3,477 24,300 5,900 5,000 20.6% 900 3.7% 7.69 8.37 12.36 Total 32,892 24,500 7,000 6,100 24.9% 900 3.7% 4.77 5.44 8.45 Urban Interstates and Freeways (FC=11 and =12) New England Division 1,512 65,500 5,500 2,100 3.2% 3,400 5.2% 1.22 1.39 1.95 Middle Atlantic Division 3,185 62,900 7,000 2,900 4.6% 4,100 6.5% 1.31 2.11 2.15 East North Central Division 3,463 62,700 8,400 4,700 7.5% 3,700 5.9% 1.13 1.55 1.55 West North Central Division 1,706 55,400 5,800 3,300 6.0% 2,500 4.5% 1.36 1.64 1.83 South Atlantic Division 3,814 75,800 8,000 3,700 4.9% 4,300 5.7% 1.40 1.67 1.87 East South Central Division 1,306 58,300 9,200 6,400 11.0% 2,800 4.8% 1.43 1.79 1.77 West South Central Division 3,190 69,700 8,000 3,900 5.6% 4,100 5.9% 1.19 1.36 1.65 Mountain Division 1,345 67,500 7,600 3,600 5.3% 4,000 5.9% 1.35 2.20 1.59 Pacific Division 3,443 112,500 8,700 3,700 3.3% 5,000 4.4% 0.94 1.19 1.31 Total 22,965 73,000 7,800 3,800 5.2% 4,000 5.5% 1.23 1.64 1.74 Other Rural Roads (FC 02 through 09) New England Division 5,698 7,700 500 100 1.3% 400 5.2% 5.93 35.26 12.20 Middle Atlantic Division 16,253 7,100 700 200 2.8% 500 7.0% 5.12 23.90 10.39 East North Central Division 32,213 6,300 800 300 4.8% 500 7.9% 6.07 15.62 9.04 West North Central Division 50,264 3,200 500 300 9.4% 200 6.3% 10.03 15.44 26.39 South Atlantic Division 35,955 8,300 900 300 3.6% 600 7.2% 6.22 16.16 8.11
E-6 Long-distance and Rural Travel Transferable parameters for Statewide Travel Forecasting Models Source: http://www.census.gov/geo/www/us_regdiv.pdf Figure E.2. U.S. Census divisions. Table E.3. (Continued). Geography Miles Weighted AADT Coefficient of Variation Total Trucks FAF Trucks Percent FAF Trucks Non- FAF Trucks Percent Non-FAF Trucks Total FAF Trucks Non- FAF Trucks Other Rural Roads (FC 02 through 09) (continued) East South Central Division 21,684 6,300 800 300 4.8% 500 7.9% 6.88 13.85 9.90 West South Central Division 30,149 6,000 1,000 400 6.7% 600 10.0% 8.28 10.76 8.00 Mountain Division 29,094 3,500 500 200 5.7% 300 8.6% 16.69 25.92 14.83 Pacific Division 22,756 6,700 800 300 4.5% 500 7.5% 11.03 14.43 8.71 Total 244,067 5,700 700 300 5.3% 400 7.0% 8.34 15.79 10.86 Other Urban Roads (FC=14 through 19) New England Division 3,288 16,800 800 100 0.6% 700 4.2% 1.41 3.34 1.68 Middle Atlantic Division 7,610 19,000 1,200 200 1.1% 1,000 5.3% 1.55 3.32 1.88 East North Central Division 10,375 18,600 1,400 200 1.1% 1,200 6.5% 1.56 3.62 1.88 West North Central Division 4,006 14,600 1,000 300 2.1% 700 4.8% 1.64 2.90 1.90 South Atlantic Division 10,077 23,900 1,600 300 1.3% 1,300 5.4% 1.84 4.68 1.88 East South Central Division 4,352 17,100 1,400 300 1.8% 1,100 6.4% 2.14 4.14 2.70 West South Central Division 8,156 18,200 1,500 300 1.6% 1,200 6.6% 1.67 4.19 2.33
Urban Versus Rural Truck Trips E-7 volumes are primarily a reflection of geography based on the following statement found in FAF3 documentation: For international O-D pairs, the process is static where an adjacent network ânodeâ of each border crossing or port geo-location is a virtual O-D zone. The virtual O-D zone for international movement was further divided into cross-border movements (U.S.-Canada and U.S.-Mexico) and port movements. Cross-border movements were defined as O-D pairs originating from FAF zone adjacent to Canada or Mexico and destined to other FAF zone and vice versa. Similarly, for ports, the O-D pairs originated from or were headed toward a FAF zone containing one or more ports or gateways (http://faf.ornl.gov/fafweb/Data/ FAF_3_network_assignment_executive_summary.pdf). E.2 Conclusions It is recommended that any travel-demand forecasting model, particularly statewide models, include separate matrices of freight and nonfreight trucks. It is suggested that a commodity flow database at a disaggregated level be used to develop the freight truck table, or be used as a calibration database to develop the parameters for freight truck trip tables. It is recommended that the relationship between the coefficients and parameters for passen- ger trips in urban and rural areas should guide the development of coefficients and parameters for nonfreight truck generation and distribution in urban and rural portions of travel demand models. The assumption is that if the shopping passenger trips per retail employee are 50 per- cent higher in urban models, compared to rural models, that the truck trips per retail employee would also show the same relationship. It is recommended that further research be undertaken to determine these parameters, making the effort to distinguish freight (FAF) truck trip generation and distribution from nonfreight truck generation and distribution for the customary treatment of total truck generation and distribution. It is suggested that the differences in percentage of FAF total traffic are largely a function of the larger values of non-FAF trucks and passenger cars in these denser, more metropolitan U.S. Census divisions and that the values found in Table E.3 be used as a reasonableness check. One notable exception is the average value for FAF trucks on rural Interstates in the New England Division. It is suggested that this is a function of geography where the travel between FAF regions does not have to pass through significantly rural sections of New England and that lower values for FAF trucks in New England should be expected. It is suggested that there is no value in developing transferable freight rates, from public com- modity sources such as the Commodity Flow Survey (CFS) (Research and Innovative Technology Administration, Bureau of Transportation Statistics, Commodity Flow Survey, http://www.bts. gov/publications/commodity_flow_survey/index.html) or FAF3, because freight traffic to, from, and through jurisdictions differ geographically, as shown in Table E.3. Hence, this appendix of the report includes no transferable parameters or rates for that reason. Even though freight truck traffic is different enough not to be transferable, nonfreight truck traffic does show consistent behavior. Although models have not yet been developed to make such distinctions, it is hoped that over time the process shown in Figure E.2 of the aforementioned NCFRP Report 8 will become more widely developed, making distinctions between freight and nonfreight trucks. At that time, the rates for nonfreight trucks should be tested for transferability. At this time, almost all truck models combine freight and nonfreight trucks, which means since freight truck rates are nontransferable, the combined rates are nontransferable.
