Case Studies of Truck Activity Data for Emissions Modeling (2019)

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.

Guide to Truck Activity Data for Emissions Modeling E-1 Appendix E. Case Study #5: Speed Distributions from the National Performance Management Research Data Set E.1 Emissions Model Inputs Supported • Average speed distribution. E.2 Level of Effort for Local Application This method requires a moderate level of effort for local application. E.3 Overview This case study compares the emissions effects of using vehicle speed distributions from the National Performance Management Research Data Set with MOVES defaults for Virginia. A sensitivity analysis was performed by running MOVES with truck speed distributions as observed in the NPMRDS and comparing the results with MOVES default speed distributions. NCHRP Web-Only Document 210 notes that while speed distributions specific to source type are rarely used in practice due to data availability, evidence from California suggests that truck speeds on freeways can be significantly lower than average speeds for all traffic. In MOVES2014 documentation, U.S. EPA references research performed by the University of California at Riverside showing a ratio of truck to car speeds of approximately 0.92.8 Variations by source type can have a modest to moderate effect on emissions. A sensitivity analysis prepared for NCHRP Web-Only Document 210 found that reducing the average speed of trucks by 5 mph compared to the MOVES default distribution (for all road types) increased overall VOC emissions by 1.2 percent, NOx by 2.4 percent, and PM10 by 5.7 percent. This case study further explores the potential impact of source-specific speed distributions on emissions using empirical NPMRDS speed data from Virginia. The NPMRDS is a nationwide dataset of roadway speeds by road segment and time period for the National Highway System. FHWA funds the NPMRDS and makes it available to states and metropolitan planning organizations (MPOs) to support performance measurement. The terms of the NPMRDS licensing agreement do not permit use of the dataset by other entities, including for NCHRP research projects. However, in 2016 the Virginia Department of Transportation separately funded a study to evaluate the use of the NPMRDS to develop speed distributions suitable for use in the MOVES model. The results of the study were documented in a memorandum published by VDOT (Cambridge Systematics, Inc., 2017). The speed distributions presented in that memorandum were used in this study as inputs to the MOVES model to compare emissions outputs. The NPMRDS methodology, documented in the VDOT memorandum, is also summarized here. Any State or MPO with access to the NPMRDS data could process their own data using similar methods to develop speed distributions. 8 U.S. EPA (2016), p. 72. The EPA document states that embedded or “default” truck speed distributions in MOVES were adjusted based on this ratio for rural and urban restricted access road types.

Guide to Truck Activity Data for Emissions Modeling E-2 E.3.1 MOVES Approach to Speed Distributions The average speed distribution for county-scale MOVES analysis is a set of 16 fractions that sum to 1, which represents the distribution of vehicle-hours traveled among 16 speed bins. MOVES requires this information for every combination of 13 source (vehicle) types, four road types, 24 hours of the day, and two types of days (weekdays/weekends). For project-scale analysis, MOVES accepts inputs of average speed by source (vehicle) type on the project link. E.3.2 Case Study Approach The following data evaluations and comparisons were made in this case study: • The VDOT memorandum reported speed distributions for trucks as observed in the NPMRDS for the four types of roads identified in MOVES: rural restricted access, rural unrestricted access, urban restricted access, and urban unrestricted access. Note that the NPMRDS only includes data from the NHS, so mainly higher-volume, higher functional class roads will be represented. • The NPMRDS speed distributions for Virginia were compared with the default speed distributions for single-unit (52 and 53) and combination (61 and 62) trucks in MOVES. • The NPMRDS speed distributions were provided to MOVES along with other Virginia-specific input data at the county scale of analysis. The resulting emissions were compared with the emissions obtained using the same Virginia-specific inputs along with the default (embedded) speed distributions in MOVES. E.4 Data Sources Data Source Source Agency/Organization Availability and Cost National Performance Management Research Data Set (Virginia sample) Federal Highway Administration Free to states and MPOs The passenger car and truck speed data in the NPMRDS are provided to FHWA by Inrix, and are from 2015. Truck data are collected from trucks in fleets with GPS data loggers. E.5 Data Processing and Analysis E.5.1 Speed Data The key issues in processing NPMRDS data to obtain speed distributions suitable for MOVES inputs include: • Aggregating NPMRDS travel time data from the 5-minute level to the hourly level. • Associating NPMRDS road types with MOVES road types. • Matching NPMRDS travel time data with traffic volume data and segment lengths so that travel times by segment can be weighted appropriately. This is the most challenging issue and requires use of segment

Guide to Truck Activity Data for Emissions Modeling E-3 information from the Highway Performance Monitoring System and volume information from the Traffic Management Analysis System, both of which are maintained by FHWA. The approach to processing NPMRDS data, as outlined in the VDOT memo (which also gives screenshots of data tables), is summarized as follows: 1. Join the Highway Performance Monitoring System (HPMS) database with truck volume data from TMAS based on functional classification ID and urban/rural code. 2. Aggregate NPMRDS travel time data from the 5-minute level to the hourly level. 3. Join the NPMRDS dataset with the HPMS database according to Traffic Message Channel link ID and roadway name, begin mile post and end mile post. 4. Calculate hourly volumes based on truck annual average daily traffic (AADT) and functional class from HPMS, hourly volume distribution from TMAS, and hours from NPMRDS. 5. Aggregate hourly volumes from HPMS segments to TMC segments. 6. Aggregate NPMRDS speeds from TMC segments to routes as provided in the TMC description table (this step may not be necessary, instead retaining segment-level data). Space mean speeds are calculated based on aggregated route lengths and travel times, and TMC volumes are aggregated to route volumes using distance as weight. 7. Calculate vehicle-hours traveled (VHT) on each route by dividing the length of the route (miles) by speed (miles/hour) and multiplying by AADT. 8. Tag each route with a MOVES speed bin based on the route speed. 9. Sum VHT by speed bin for each hour and weekday/weekend. Note that the segments are defined differently in the NPMRDS, which uses TMC segments, and TMAS, which reports volumes for HPMS segments, so the volume-weighting of speeds by segment will not be precise. The time periods for the speed and volume observations also may not match exactly; in this case, speed data were from 2015 and volume data from 2013 – 2014. Some issues with roadway classifications were noted in the Virginia HPMS file; in particular, some Interstate highways were classified as unrestricted in the HPMS file and were manually reclassified. E.5.2 Other MOVES Inputs and Assumptions Other MOVES input data were obtained from VDOT or using MOVES defaults. Key data sources and assumptions included: • Mode: Inventory. • Scale: County—Prince William County, Virginia (51153). • Timeframe: 2014, April, Weekdays, 10 – 11 a.m.

Guide to Truck Activity Data for Emissions Modeling E-4 • Pollutants: PM2.5, NOX and associated processes. • Source types: 52, 53, 61, 62 (the VDOT analysis also examined source type 21, passe). • Inputs: – Average Speed Distribution: NPMRDS for one run and MOVES defaults for the other run. MOVES defaults from several different Virginia counties were compared and confirmed to be consistent across these counties. – Fuel inputs: MOVES defaults. – VMT and associated inputs: VDOT.9 – Source Type Population: VDOT. – Road Type Distribution: VDOT. – Age Distribution: VDOT. – Meteorology: VDOT. E.6 Findings from Sample Data E.6.1 Speed Distributions Single-Unit Short-Haul and Long Haul Trucks (52, 53) share the same MOVES default speed distributions. Combination short-haul and long-haul trucks (61, 62) share the same MOVES default speed distributions. The NPMRDS does not distinguish between these source types, but only provides a single speed distribution for all truck source types. Figure E.1 to Figure E.4 compare the MOVES and NPMRDS truck speed distributions for each road type. 9 Inputs from VDOT were obtained through the Outside VDOT database in March 2017.

Guide to Truck Activity Data for Emissions Modeling E-5 Figure E.1 Rural Restricted Access Speed Distribution Virginia Figure E.2 Rural Unrestricted Access Speed Distribution Virginia 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 2.5 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 Fraction Speed Bin (mph) NPMRDS Truck Default Single-Unit Truck Default Combination Truck 0 0.05 0.1 0.15 0.2 0.25 0.3 2.5 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 Fraction Speed Bin (mph) NPMRDS Truck Default Single-Unit Truck Default Combination Truck

Guide to Truck Activity Data for Emissions Modeling E-6 Figure E.3 Urban Restricted Access Speed Distribution Virginia Figure E.4 Urban Unrestricted Access Speed Distribution Virginia 0 0.1 0.2 0.3 0.4 0.5 0.6 2.5 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 Fraction Speed Bin (mph) NPMRDS Truck Default Single-Unit Truck Default Combination Truck 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 2.5 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 Fraction Speed Bin (mph) NPMRDS Truck Default Single-Unit Truck Default Combination Truck

Guide to Truck Activity Data for Emissions Modeling E-7 Observations include: • For both rural restricted and unrestricted access roads, the NPMRDS truck and auto speed distributions peak at lower speeds than the MOVES defaults. • For urban unrestricted access roads, the default distributions are flatter with higher fractions of VMT occurring in the lowest speed bins compared to the NPMRDS distributions. A possible explanation for this discrepancy is that compared to defaults, NPMRDS data might be collected from relatively major roads that tend to facilitate higher vehicle speeds. • The default distributions tended to be smoother and have lower peaks than the NPMRDS speed distributions, presumably due to limited sample sizes in NPMRDS. Note that the since the NPMRDS only contains NHS roads, it may not be representative of all roads in the State. Most restricted access roads will be part of the NHS, but there will be many unrestricted access roads that are not part of the NHS. The VDOT memo also compares truck and car speed distributions from the NPMRDS (included at the end of this case study as Figure E.9). The memo notes that when comparing average truck speeds to average car speeds on the same roadway route, trucks speeds appear to be about 7 percent lower than car speeds on average across all roadway types. However, when comparing the ratio of weighted average speeds across segments, truck speeds appear closer to auto speeds. This appears to be due to the higher amount of truck travel on higher-speed roads. This information is included to show some of the considerations when using a fleetwide average speed distribution, which is primarily weighted to autos, to represent truck speeds. E.6.2 Emissions Results Figure E.5 to Figure E.8 show the effect on emissions of the different speed distributions. Figure E.5 and Figure E.6 show NOx emissions for combination and single-unit trucks, respectively, while Figure E.7 and Figure E.8 show PM2.5 emissions.

Guide to Truck Activity Data for Emissions Modeling E-8 Figure E.5 Combination Truck NOx Emissions (g/mi) Figure E.6 Single-Unit Truck NOx Emissions (g/mi) 0 2 4 6 8 10 12 14 16 Rural Restricted Rural Unrestricted Urban Restricted Urban Unrestricted Grams/Mile NPMRDS Combo Short-Haul Default Combo Short-Haul NPMRDS Combo Long-Haul Default Combo Long-Haul 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 Rural Restricted Rural Unrestricted Urban Restricted Urban Unrestricted Grams/Mile NPMRDS Single-Unit Short-Haul Default Single-Unit Short-Haul NPMRDS Single-Unit Long-Haul Default Single-Unit Long-Haul

Guide to Truck Activity Data for Emissions Modeling E-9 Figure E.7 Combination Truck PM2.5 Emissions (g/mi) Figure E.8 Single-Unit Truck PM2.5 Emissions (g/mi) 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 Rural Restricted Rural Unrestricted Urban Restricted Urban Unrestricted Grams/Mile NPMRDS Combo Short-Haul Default Combo Short-Haul NPMRDS Combo Long-Haul Default Combo Long-Haul 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 Rural Restricted Rural Unrestricted Urban Restricted Urban Unrestricted Grams/Mile NPMRDS Single-Unit Short-Haul Default Single-Unit Short-Haul NPMRDS Single-Unit Long-Haul Default Single-Unit Long-Haul

Guide to Truck Activity Data for Emissions Modeling E-10 Observations include: • The largest differences in grams per mile emission rates between the NPMRDS and defaults occurred on the urban unrestricted access roads. Truck NOx emissions were 14 – 22 percent lower for NPMRDS speed distributions than for default speed distributions. Truck PM2.5 emissions were 20 – 28 percent lower. • For rural restricted access and unrestricted access roads and for urban restricted access roads, differences in NOx truck and auto emissions between the NPMRDS and default speed distributions were all relatively small (within +/-8 percent). • On rural restricted access roads, PM2.5 emissions for single-unit trucks were 17-18 percent higher for NPMRDS than for default speed distributions. • On rural unrestricted access roads, PM2.5 emissions for both single-unit and combination trucks were 11 to 16 percent higher for NPMRDS than for default speed distributions. • For other combinations of road types and source types, differences in PM2.5 emissions between NPMRDS and default speed distributions were relatively small (within +/-7 percent). These included combination trucks on rural restricted access roads and all trucks on urban restricted access roads. E.7 Transferability The procedure described in this case study could be applied in any State or metropolitan area with access to NPMRDS data. As noted, the NPMRDS data are most appropriate for representing higher functional class roads and may not be representative of the entire road system. The relative impact on emissions may differ depending upon the specific speed distributions observed within a given State. The VDOT memo noted the potential for further data-mining and evaluation, such as: • Developing speed distributions using disaggregated roadway segments rather than the route-level aggregation applied here, and explore the effect of level of aggregation on the accuracy of speed distributions. • Comparing speeds on NPMRDS versus non-NPMRDS roads (as gathered from a more comprehensive source, such as HERE) to evaluate how representative the speeds in the NPMRDS are of speeds on all roads. • Evaluating differences in speed distributions by hour of the day and robustness of hourly data for use in developing MOVES inputs. It should be noted that FHWA is undertaking research to explore the use of the NPMRDS for development of emissions estimates on a nationwide basis.

G uide to Truck A ctivity D ata for E m issions M odeling E-11 Figure E.9 Auto and Truck Speed Distributions from VDOT NPMRDS Data 0.0 0.1 0.2 0.3 0.4 0.5 0.6 Speed range (mph) Speed Distribution, Rural Restricted auto_vht% truck_vht% 0.0 0.1 0.2 0.3 0.4 0.5 0.6 Speed range (mph) Speed Distribution, Rural Unrestricted auto_vht% truck_vht% 0.0 0.1 0.2 0.3 0.4 0.5 0.6 Speed range (mph) Speed Distribution, Urban Restricted auto_vht% truck_vht% 0.0 0.1 0.2 0.3 0.4 0.5 0.6 Speed range (mph) Speed Distribution, Urban Unrestricted auto_vht% truck_vht% Source: Cambridge Systematics, Inc. (2017).