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.
11 3.1 Overview Age distribution is characterized in MOVES as the fraction of total vehicle population with age = x, where x spans from 0 to 30 years old, and the sum of age distribution is one. For county-scale analysis, age distributions can be provided for each of the 13 source types. For project-scale analysis, an age distribution is provided for the source type associated with the analysis link. 3.2 MOVES Embedded Data Embedded data in MOVES represent national average age distributions. Truck age distribu- tions in MOVES2014 are for a 2011 base year and are estimated from 2011 registration data compiled by R. L. Polk (now part of IHS Automotive) and data from the 2002 VIUS to develop detailed categories (short-haul versus long-haul trucks). Age distribution is grown dynamically year to year based on sales and scrappage inputs when output is requested for calendar years after the base year (U.S. EPA, 2016). EPA guidance for SIP and conformity modeling suggests local (state) vehicle registration data as the primary source of local data for age distribution input. However, the EPA suggests MOVES default (embedded) data as a fallback for those source types lacking local data, particu- larly for combination heavy-duty trucks, which are often not registered in the state in which they are being driven. 3.3 Sensitivity/Importance Age distributions can have a moderate to substantial impact on emissions. For example, Noel and Wayson (2012) tested sensitivity to moving 10 percent of vehicles among age groups (0 to 10 years, 11 to 20 years) and 5 percent into the oldest age group (21 to 30 years). Emissions varied by up to 32 percent for single-unit trucks and up to 10 percent for combination trucks. An analysis for the Coordinating Research Council (CRC) (Eastern Research Group, Inc., 2013) comparing the 10th and 90th percentile average ages of state National Emissions Inventory (NEI) submissions, found variances of between 19 and 119 percent, depending upon the truck type and pollutant. These are based on large age shifts, and the typical effect of age distribution would likely be smaller than is shown in the CRC analysis. A comparison of trucks observed at inspection stations versus MOVES defaults, documented in Case Study #2 (See Appendix B, available in NCHRP Web-Only Document 261), found differences ranging from 10 to 72 percent, with the on-road fleet being older than the defaults assumed in MOVES. Results from various sensitivity analyses are shown in Table 3.1. S E C T I O N 3 Age Distributions
12 Guide to Truck Activity Data for Emissions Modeling Age distributions can vary across regions due to economic and climate factors, as well as the uses of regionally registered trucks. For example, older combination trucks are often retired from long-haul service and placed into use as drayage trucks serving ports. Age dis- tributions also can change over time in response to economic cycles and policies that affect vehicle costs and technology. Figure 3.1 shows age distributions of trucks at inspection sta- tions in Virginia over a 3-year period (2013 to 2015). A low volume of trucks in the model year timeframe of 2008 to 2010 (age 3 to 7 years) is observed; this corresponded not only with an economic downturn, but also the introduction of new emission standards that increased the cost of trucks. State registration data may not represent the on-road truck fleet well, especially in smaller states and for combination trucks. An analysis of truck age distributions at inspection stations in four sample statesâCalifornia, Michigan, Texas, and Virginiaâfound that the out-of-state population varied by state, but was significant in all cases. California had the lowest proportion of out-of-state trucks being inspected (6 to 13 percent of straight trucks, about 30 percent of truck tractors), while Virginia had the highest proportion of out-of- state trucks being inspected (46 to 50 percent of straight trucks, about 70 percent of truck tractors). As expected, straight trucks (which are typically used for shorter hauls) consis- tently show lower percentages of out-of-state inspections than truck tractors. Case Study #2 provides more detail on this analysis (See Appendix B, available in NCHRP Web-Only Document 261 at website URL). Age distributions for trucks in long-haul service might be expected to be more consistent across the country because these trucks are often registered in one state and used in many other states. However, inspection station data from MCMIS still showed differences among states; From Average Age (yrs) To Average Age (yrs) Delta Age (yrs) Delta Volatile Organic Compound (VOC) Delta NOx Delta PM2.5 Urban Restricted at 60 mph, Calendar Year 2015a Single-Unit Long-Haul Truck 8.7 11.5 2.8 32% 30% 32% Combination Long-Haul Truck 9.3 10.5 1.2 8% 8% 10% State NEI SubmissionsâCountywide Inventory, Calendar Year 2013b Light Commercial Truck 8.5 12.2 3.7 60% 41% 55% Refuse Truck 9.0 14.5 5.5 57% 55% 84% Single-Unit Short-Haul Truck 9.9 15.9 6.0 65% 42% 74% Single-Unit Long-Haul Truck 7.3 15.3 8.0 104% 66% 119% Combination Short-Haul Truck 9.9 14.4 4.5 33% 49% 50% Combination Long-Haul Truck 8.5 13.6 5.1 18% 38% 49% Observed at Inspection Stations in Texas and Virginia versus MOVES Defaults, 2013â2015c Single-Unit Truck 11.6â11.7 15.0â16.9 3.4â5.2 18â39% 20â41% 23â53% Combination Truck 11.2â11.7 14.0â15.9 2.8â4.2 10â37% 20â57% 28â71% aNoel and Wayson, 2012. bEastern Research Group, Inc., 2013. cAnalysis of MCMIS data by Volpe National Transportation Systems Center, Eastern Research Group, Inc., and Cambridge Systematics. Table 3.1. Age distribution effects on emissions.
Age Distributions 13 the average age of combination trucks in Texas was 1.2 to 2.3 years older than the average age of combination trucks in Virginia for the same time period. The average age of combination trucks also varied by up to 3.4 years across three counties evaluated in Virginia and by up to 4.1 years across three counties evaluated in Texas (see Case Study #2). 3.4 Generating Local Data The following options may be available for generating local truck age distribution data: â¢ State vehicle registration data. Volume 1, Section 4, of NCHRP Web-Only Document 210 (Porter et al., 2014a) discusses this data source in detail and that discussion is not repeated here. State registration data are more likely to be appropriate for trucks in short-haul use than for trucks in long-haul use. â¢ National safety inspection station data. Inspection station records can be used to identify vehicles at a specific location, and these records can be mapped to trucks registered in the MCMIS to identify the ages of trucks. An example of the use of these data is provided below. â¢ License plate survey data combined with MCMIS and/or state registration data. If there are no inspection stations representative of the project or area of interest, license plate surveys can be conducted with records matched with MCMIS inspection records and/ or state registration data to identify local age distributions. A sample license plate survey approach is described below. Note that age distributions based on field observations will need to be corrected for dif- ferences in mileage accumulation rate (MAR) by vehicle age to make them comparable with Source: Analysis by Volpe National Transportation Systems Center of MCMIS data. 0 0.1 0.2 0.3 0.4 0.5 0.6 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 Age Fraction Age (yrs) 2013 2014 2015 Figure 3.1. Age distributions for combination trucks observed at Virginia inspection stations (2013â2015).
14 Guide to Truck Activity Data for Emissions Modeling state or regional, population-based age distributions input to MOVES for county-scale analysis. Otherwise, newer vehicles will be oversampled because they travel more. This correc- tion does not need to be made for project-scale analysis, where the MOVES age distribution input represents the distribution of vehicles traveling on the facility. The correction procedure is described below. 3.4.1 Inspection Station Data from MCMIS Database Source, Availability, and Cost Data from the FMCSAâs MCMIS can be used to assist in developing vehicle age distributions. MCMIS is FMCSAâs source for inspection, crash, compliance review, safety audit, and registra- tion data.2 Each commercial motor vehicle (CMV) operating in the United States is required to undergo a safety inspection on at least an annual basis to determine its roadworthiness. The data collected from each safety inspection are stored in MCMIS. MCMIS essentially serves as a national truck database. MCMIS data can be made available by making a DART request through FMCSA. DART is a system to record and track requests for FMCSA data. The Volpe Center manages the MCMIS database, as well as all DART requests on behalf of FMCSA. The NCHRP Project 08-101 team was able to obtain MCMIS data at no cost. The DART system requires different forms of online registration depending on the user type: state or federal agencies, motor carriers, or the general public. State or federal agencies must complete and submit the FMCSA Information Technology Account Request Form.3 Motor carriers can view additional MCMIS data by entering a valid personal identification number (PIN) provided by FMCSA.4 The general public must register through the DART online portal.5 Once registered, users are requested to fill out an MCMIS file extract order form. Inspection file extracts cost $70 per calendar year for general public requests and are typically delivered on CD-ROM.6 Costs may vary depending on the type of user and number of calendar years requested. File extracts also may be made for motor carrier census and/or crash data. Level of Data-Processing Effort This analysis requires moderate effort, primarily for statistical processing of acquired data. Data-Processing Steps 1. Make an official DART request for MCMIS data, specifying the geography and year(s) of interest. Table 3.2 includes the data fields that were obtained by the research team. The total file size for 3 years of data (2013 through 2015) for the entire United States was approximately 2.3 gigabytes. 2. Identify inspection station(s) corresponding with the study area geography. Safety inspec- tion locations are commonly paired with weigh station facilities. CMVs will sometimes be flagged after completing the weighing process to proceed to the inspection area for a safety 2 U.S. DOT, FMCSA, MCMIS Catalog and Documentation, https://ask.fmcsa.dot.gov/app/mcmiscatalog/mcmishome. 3 U.S. DOT, FMCSA Information Technology Account Application, https://infosys.fmcsa.dot.gov/PublicDocument/MCMIS/ MCMISapplication.pdf. 4 U.S. DOT, FMCSA Registration, U.S. DOT Number/Docket Number PIN Request, http://mcmis.volpe.dot.gov/mcs150t/ Pkg_PUBLIC_MCS150.prc_public_dot_pin_request. 5 U.S. DOT, FMCSA DART Registration, https://ai.fmcsa.dot.gov/DART/register.asp. 6 U.S. DOT, FMCSA MCMIS File Extract Order Form, https://ask.fmcsa.dot.gov/euf/assets/mcmiscatalog/PDF/MCMIS extractorderform20140227_Senture.pdf.
Age Distributions 15 inspection. It is important to note that only a small number of CMVs are inspected relative to all the vehicles that pass through a weigh station facility. 3. Choose an appropriate data analysis platform. The project team used a relational database (Microsoft SQL Server), statistical analysis package (R scripts), and spreadsheet (Microsoft Excel) to conduct its analysis. 4. Load MCMIS data from the DART request into a relational database to facilitate querying and analysis. Filter the data by inspection year and state. 5. Analyze truck ages separately for two different vehicle types: (a) tractor-trailers and (b) straight trucks. The MCMIS tractor-trailer type description (INSP_UNIT_TYPE_DESC) is analogous to MCMIS Fields Field Description INSPECTION_ID Unique identifying number assigned to an inspection record. INSP_DATE Date the inspection was conducted (YYYYMMDD). INSP_START_TIME Time the inspection started. Format: HHMM (military time) where HH is the hour and MM is the minutes. Midnight is recorded as 2400. 9999 = Time unknown. INSP_END_TIME Time the inspection ended. Format: HHMM (military time) where HH is the hour and MM is the minutes. Midnight is recorded as 2400. 9999 = Time unknown. COUNTY_CODE_STATE State/district/province of the county where the inspection took place. COUNTY_CODE Federal Information Processing Standards (FIPS) code identifying the county in which the inspection occurred. COUNTY_NAME County name. LOCATION State-assigned code representing the physical location of the inspection. LOCATION_DESC Physical location where the inspection took place. INSP_UNIT_ID Unique identifying number assigned to an inspected unit. INSP_UNIT_TYPE_ID Unique identifying number assigned to an inspection record. INSP_UNIT_TYPE_DESC A code identifying the type of vehicle unit: 1. BUâBus 2. DCâDolly Converter 3. FTâFull Trailer 4. LMâLimousine 5. MCâMotor Carrier 6. OTâOther 7. PTâPole Trailer 8. SBâSchool Bus 9. STâSemi Trailer 10. TRâStraight Truck 11. TTâTruck Tractor 12. VNâVan 13. ZZâUnknown 14. Intermodal Chassis INSP_UNIT_LICENSE_STATE State/district/province issuing the license tags on the vehicle unit. INSP_UNIT_LICENSE License tag number on the vehicle unit. INSP_UNIT_VEHICLE_ID_ NUMBER Vehicle identification number of the unit. INSP_UNIT_YEAR Model year of the vehicle unit. INSP_UNIT_GVWR Gross vehicle weight rating (GVWR) of the vehicle unit. Table 3.2. MCMIS data fields.
16 Guide to Truck Activity Data for Emissions Modeling the MOVES combination truck class, and the MCMIS straight truck type description is analo- gous to the MOVES single-unit truck class. Note that MCMIS does not permit distinguishing between short-haul and long-haul trucks due to limited information on average daily mileage of the trucks inspected. Instead, compare MCMIS tractor-trailer ages against the aggregate ages of MOVES short-haul combination trucks (sourceTypeID 61) and long-haul combina- tion trucks (sourceTypeID 62). Likewise, compare MCMIS straight truck ages against the aggregate ages of MOVES short-haul, single-unit trucks (sourceTypeID 52) and long-haul, single-unit trucks (sourceTypeID 53). 6. After being appropriately formatted, import the MCMIS data from the database into a statis- tical analysis package for analysis and plotting. For each inspection record in the MCMIS data analyzed, query the calendar year from the inspection date field (INSP_DATE). Similarly, for each MCMIS record, take the truck model year directly from the year of the inspection unit field (INSP_UNIT_YEAR). Determine the truck age by subtracting the inspection unit year (model year) of the vehicle from the inspection year (calendar year). Place any trucks more than 30 years old in the 30-year age bin, and place any early sale trucks with a model year newer than calendar year in the zero-year age bin. 7. Once all the trucks are assigned ages, calculate the number of trucks inspected in each county (COUNTY_NAME) for each calendar year. To maintain sufficient inspection sample sizes, it may be necessary to limit the analysis to the counties with the most tractor-trailer inspections by year. To determine the county distributions of truck age, divide the number of trucks of each age by the total number of trucks for each county, each calendar year, and each vehicle type. State-level distributions can also be reported (COUNTY_CODE_STATE) using all truck inspections for each calendar year and each vehicle type. 8. Optionally, compare the MCMIS and MOVES age distributions by plotting the age dis- tributions and comparing average truck ages. To directly compare the VMT-weighted age distributions derived from MCMIS with the population-weighted age distributions typi- cally input to MOVES, adjust the MOVES national default truck age distribution by the national default MAR, as discussed further below. (This adjustment should be made if the MCMIS distribution is to be used at a project or link level. For county-level analysis, the MCMIS distribution should be adjusted by the MAR to obtain an equivalent population- weighted distribution.) Adjustments for MAR When conducting a national-scale or county- or regional-scale analysis using MOVES, the age distribution entered into the input database is based upon vehicle populations. When MOVES is executed, an adjustment is applied to the age distribution to account for vehicle activity. That adjustment is the MAR. Essentially, the MAR will apply adjustments to the age fraction due to newer vehicles having more activity than older vehicles. The MCMIS age distributions observed at a specific facility represent the age distribution of the vehicles traveling on the facility (i.e., VMT age distributions). Newer vehicles travel more miles per year and will therefore be disproportionately represented in observed vehicle travel compared to older vehicles. If observed MCMIS age distributions are to be compared with MOVES age distributions, a MAR adjustment must be applied to either the MCMIS distribu- tions (to âageâ the fleet) or to the MOVES registration-based distribution (to make the fleet âyoungerâ). An example of how the MAR adjustment can be applied is shown in Table 3.3. In this table, columns (B) and (C) are MOVES-output VMT and vehicle counts for source type 62 (long-haul combination trucks) for the State of Virginia for 2014. Column (D) is the computed VMT per vehicle, which is column (B) divided by column (C). Column (E) is the relative MAR, or the
Age Distributions 17 (A) MOVES Age ID (B) MOVES VMT (C) MOVES Vehicle Count (D) MOVES VMT/ Vehicle (E) MOVES Relative MAR (F) VMT Age Fraction (Road Observation) (G) Age Fraction/ MAR (H) Population Age Fraction (Normalized) 0 370,547,260 3,614 102,524 1.0000 0.0992 0.0992 0.0472 1 328,909,600 3,210 102,460 0.9994 0.0763 0.0763 0.0363 2 319,174,154 3,128 102,052 0.9954 0.0691 0.0694 0.0330 3 281,995,558 2,483 113,587 1.1079 0.0330 0.0298 0.0142 4 209,731,756 1,930 108,672 1.0600 0.0232 0.0219 0.0104 5 261,317,216 2,529 103,310 1.0077 0.0294 0.0292 0.0139 6 196,817,110 2,051 95,946 0.9358 0.0188 0.0201 0.0096 7 634,587,774 6,962 91,150 0.8891 0.1162 0.1307 0.0621 8 439,604,766 5,133 85,637 0.8353 0.0830 0.0994 0.0473 9 389,547,616 4,887 79,704 0.7774 0.0730 0.0939 0.0446 10 217,752,611 2,960 73,575 0.7176 0.0338 0.0471 0.0224 11 192,678,202 2,795 68,930 0.6723 0.0373 0.0555 0.0264 12 129,868,595 2,072 62,666 0.6112 0.0280 0.0458 0.0218 13 148,876,619 2,884 51,618 0.5035 0.0452 0.0898 0.0427 14 205,834,533 4,305 47,814 0.4664 0.0629 0.1350 0.0642 15 139,551,618 3,446 40,499 0.3950 0.0428 0.1084 0.0516 16 87,372,355 2,816 31,023 0.3026 0.0299 0.0988 0.0470 17 51,112,860 1,971 25,930 0.2529 0.0187 0.0740 0.0352 18 55,890,108 2,501 22,349 0.2180 0.0187 0.0859 0.0408 19 50,860,993 2,637 19,287 0.1881 0.0167 0.0888 0.0422 20 32,636,851 2,017 16,184 0.1579 0.0112 0.0708 0.0337 21 21,019,328 1,440 14,596 0.1424 0.0075 0.0529 0.0252 22 13,114,383 1,051 12,472 0.1217 0.0039 0.0320 0.0152 23 10,180,850 945 10,773 0.1051 0.0024 0.0227 0.0108 24 9,746,189 1,113 8,759 0.0854 0.0026 0.0309 0.0147 25 8,679,282 1,122 7,736 0.0755 0.0034 0.0450 0.0214 26 6,335,502 1,002 6,320 0.0616 0.0029 0.0469 0.0223 27 5,164,676 841 6,141 0.0599 0.0030 0.0503 0.0239 28 2,334,920 538 4,338 0.0423 0.0019 0.0445 0.0212 29 1,866,927 545 3,424 0.0334 0.0021 0.0639 0.0304 30 1,967,575 788 2,497 0.0244 0.0035 0.1444 0.0687 Sum: 1.0000 2.1037 1.0000 Average Age: 8.5 14.8 Source: MOVES VMT and vehicle counts for source type62 vehicles in Virginia in 2014, from MOVES2014 analysis by Cambridge Systematics, Inc., using inputs supplied by Virginia Department of Environmental Quality. Age fraction (road observation) estimated from Virginia inspection station data from MCMIS by Volpe National Transportation Systems Center. Table 3.3. Sample age distribution adjustment for MAR.
18 Guide to Truck Activity Data for Emissions Modeling ratio of VMT per vehicle of age x to VMT per vehicle of age 0. Column (F) is the observed VMT age fraction from inspection station data or data from another location sampling on-road vehicles. Column (G) is the VMT age fraction divided by MAR for each age of vehicles. Column (H) is the normalized age distribution equivalent to a registration-based distribution, which is column (G) value divided by the sum of the column (G) values (2.1037). In Table 3.3, columns (B), (C), and (F) are user-input values. Other fields are calculated. Users could start with column (D) or column (E) if they have VMT/vehicle by age or a relative MAR from a source other than MOVES outputs. The average age can be computed as the sumproduct (in MS Excel) of column (A) and column (F) or (H). In the example shown in Table 3.3, the average age of vehicles in the registered fleet is substantially older than the average age of vehicles observed at a random point on the road (14.8 versus 8.5 years in this example). This process can be reversed to convert a population-based age distribution to a road-based age distributionâif, for example, the user desires to input source type age fraction for project- scale MOVES analysis when only a registration-based age distribution is available. In this pro- cedure, column (H) is the user input; it is then multiplied by the MAR [column (E)], and then normalized to obtain column (F). Applicability and Limitations Analysis of MCMIS data shows that a large portion or majority portion of heavy-duty trucks are not registered in the same state in which they were inspected. Therefore, it is possible that age distributions on specific roadway facilities within a state differ significantly compared to the distribution presented by vehicles registered in the state. It cannot be determined with the available data whether the MCMIS-observed age distribu- tions are representative of all vehicles operating within a state. The age distributions are likely representative of traffic on Interstate and other major highways, especially in rural areas; how- ever, comparability on urban Interstates and local roads would require additional field data collection through a method such as license plate surveys. The MAR adjustment assumes that the MOVES default VMT per vehicle by age group is rep- resentative of the local truck population. Another source of VMT per vehicle by age (e.g., state registration data) could be used if available. 3.4.2 License Plate Survey Combined with MCMIS Inspection and/or State Registration Data Source, Availability, and Cost This approach is similar to the use of MCMIS data described above. However, instead of using inspection station records to sample trucks, a field survey is conducted to obtain license plate readings at a location or locations of choice. The survey can either be conducted at a single location corresponding to the project study area (e.g., a roadway link or port gates) or at multiple locations representative of the area for which emissions are to be modeled. The cost of the effort will depend upon the number of sampling locations and the volume of vehicles processed at each location. This approach is most likely to be suited for a location with a high volume of truck traffic with characteristics that may differ substantially from the average fleet characteristics shown in state or national data. If most of the trucks at the study location(s) are expected to be registered in state, license plates may be matched with state vehicle registration data obtained from the state
Age Distributions 19 department of motor vehicles. If multi-state registrations are to be considered, MCMIS can be used as described above, as long as both license number and state are col- lected as part of the survey. Level of Data-Processing Effort Collecting field survey data of truck license plates would likely require high-resolution imaging or video from well- positioned cameras, coupled with machine-learning algo- rithms, to process and decipher the license plate number and state. Alternatively, researchers could take pictures, or even simply write down license plate information, although this approach often works best when the trucks are station- ary, such as when vehicles are stopped at a rest station or an intersection. Manual license plate data collection and/or extraction is prone to error at higher speeds and/or larger volumes of trucks, so advanced sensors and techniques are preferable. Field surveys to acquire license plate data require a high level of effort, regardless of whether researchers take an advanced approach or rely on manual methods. How- ever, once the license plate data are collected and format- ted, joining license plate data from the field survey to the MCMIS data set to determine truck ages should not be par- ticularly burdensome. Data-Processing Steps General steps for estimating age distributions from license plate surveys include the following: 1. Select one or more sites for plate capture that are representative of traffic in the MOVES modeling domain. 2. Determine license plate capture technology and sampling timeframe. Pilot tests may be warranted to determine the volume of trucks that can be captured in a given time period and to match with the source registration data to determine the duration of data collection needed to obtain an adequate sample size. 3. Obtain any needed permissions to deploy LPR equipment or personnel. 4. Conduct data collection. 5. Process and clean data to develop records that include the license plate and state of registra- tion, as well as any other vehicle-related information (e.g., configuration, number of axles) that can be collected with the technology. 6. Obtain state registration data (may require multiple states) and/or MCMIS truck inspec- tion data. 7. Match license plate numbers from the field survey with registration and/or inspection data. Tag each observed license plate record with MOVES vehicle type and age informa- tion from the registration and/or MCMIS data (if information is collected from both the field survey and registration and/or inspection data, a determination that one or the other data source is more reliable may be required). An appropriate crosswalk system may need to be established between state registration vehicle classifications and MOVES source types. 8. Aggregate observed population by age (or model year). Example: Port of Houston Emissions Inventory For the Port of Houston emission inventory, a field measurement campaign was conducted using license plate recognition (LPR) cameras installed at six terminal gates for 1 to 2 weeks each. The LPR systems were installed on both the inbound and outbound lanes at each of the locations and were operational 24 hours a day during the data collection. Having entry and exit LPR allowed an estimate of time spent in a terminal by each truck. The LPR system collected both still pictures of each license plate that it detected and video each time the camera saw movement (even if no plate was detected). After extensive review of the data, including manual cross checks between the LPR and video systems, the data were analyzed to determine truck count at the terminal, average time the trucks spent in the port, and model year distribution at each terminal. These data were used to develop MOVES inputs for individual terminals. Source: Port of Houston, 2017.
20 Guide to Truck Activity Data for Emissions Modeling 9. Determine the age distribution at the observed locations (fraction of observed vehicles by model year). If multiple locations were observed and data collection timeframes differed, weighting factors may need to be applied. 10. If the age distribution is to be compared with or used to represent a population-based distribution (e.g., for a MOVES county-scale analysis), adjust the age distribu- tion by MAR, as described in the previous section. Applicability and Limitations Field surveys present their own challenges outside of any data processing after collection. The greatest concern for this type of data analysis is obtaining an adequately large sample. Without a large sample of surveyed trucks, it is nearly impossible to create a realistic age distribution. A sample of at least 500 to 1,000 trucks may be needed to develop an age distribution with reasonable confidence. 3.4.3 Other Resources Case Study #2 (See Appendix B, available in NCHRP Web-Only Document 261) presents additional detail on the analysis of MCMIS data. One of the MS Excel files that supplements this guidebook (NCHRP08-101_Data_CS2_AgeDist.xlsx) contains the MCMIS age distribution data for Texas and Virginia from Case Study #2. This file is available on the NCHRP Research Report 909 web page (http://www.trb.org/Main/Blurbs/ 178921.aspx). Example: License Plate Matching in Los Angeles and Las Vegas In 2010, FHWA conducted license plate surveys at six locations in the Los Angeles and Las Vegas metropolitan areas. License plate numbers, obtained from manual processing of video, were matched with state registration data from Nevada and California to compare age distributions of vehicles from these states. The study found that 23 percent of vehicles entering Las Vegas were registered in California and that these vehicles had very different age characteristics than the Nevada fleet. Source: Boriboonsomsin et al., 2011. The NCHRP Project 08-101 team took this same data set and attempted to match license plates with 2013â2015 MCMIS records, with the objective of examining age distributions, including trucks registered in other states. The limited sample of vehicles, combined with a limited number of matches that could be made to MCMIS, meant that age distributions could not reliably be deter- mined. To examine age distributions for interstate trucks using the license plate/MCMIS approach, a larger sample focused specifically on trucks may be needed. (See Case Study #2.)