1. 1 Estimating On-Road Vehicle Emissions Using CONCEPT Alison K. Pollack Ralph Morris ENVIRON International Corporation.

2. 2 Overview WRAP On-Road Emissions Estimates for Regional Haze Modeling CONCEPT Overview CONCEPT MV Inputs Required CONCEPT MV Emissions Approach & Steps Summary of Key Differences Between SMOKE and CONCEPT MV Denver Example Application and comparison of SMOKE vs CONCEPT approaches

3. 3 On-Road Mobile Emissions Estimates - General Approach Local Travel Data Vehicle Miles Traveled (VMT) and Speed by Roadway Type VMT Growth Rates Local Parameters Temperature Fuel Specifications Control Programs Vehicle Registration MOBILE6.2 Generates g/mi Emission Factors X = Emissions

4. 4 WRAP On-Road Emissions Estimates (2002, 2008, 2013, 2018) Regionally consistent estimation methods Surveyed State/Local Agencies for Most Up-to-Date Detailed Inputs – 2002 Base year inputs and VMT – Future year inputs, VMT growth – Unpaved road VMT – Responses were received from all state agencies and key local agencies Estimated County-level Emissions Emissions Were Processed Using SMOKE – Spatial allocation using Census TIGER roadway mileage – Temporal allocation of emissions using temporal profiles for activity data

5. 5 CONCEPT vs. SMOKE for On-Road Mobile SMOKE – Emissions are estimated using county-level activity data – Can use day/hour-specific county average – Useful for areas where detailed link data are not available CONCEPT MV – Estimates emissions based on detailed link-based activity data (VMT and speed) from a Traffic Demand Model – Uses day-specific, grid cell-specific meteorological data – Highly resolved temporal and spatial variation of VMT and fleet mix

6. 6 CONCEPT Overview Consolidated Community Emissions Processing Tool – Open Source – Freely Available – Database Management System – Community Support & Enhancements – Balances Transparency/QA and Performance/Integration with GIS/Spatial Tools – Quality Assurance – Intermediate Tables Output Includes all Emissions Source Categories – Point, Area, On/Off-Road Mobile, Biogenics Performs same emissions processing as SMOKE – Temporal and Spatial Allocation, CEM Emissions Processing, Speciation, Growth & Control Motor Vehicle (MV) Module Estimates Highly Detailed On- Road Vehicle Emissions

7. 7 CONCEPT MV Uses output from transportation demand models (TDM) Generates gridded, hourly link-level emissions by vehicle class Uses 8 MOBILE5 vehicle classes Detailed temporal resolution of traffic volume, speeds, and VMT mix Uses day/hour-specific gridded met data Stores many intermediate emissions tables for QA and data review

8. 8 TDM Link Inputs to CONCEPT MV Endpoint coordinates and projection definition Volume and capacity – Typically provided for multi-hour periods for typical weekday, e.g. am peak, midday, pm peak, overnight Speeds – Generally free flow speeds are provided – Congested speed is calculated from free flow speed and volume/capacity ratio for each hour Roadway type – MOBILE6 has different emission factors by roadway type

9. 9 CONCEPT Temporal Allocation CONCEPT disaggregates link data (volume, capacities) for multi-hour periods into hourly volumes and capacities Requires hourly total volume profiles by roadway type, month, day of week Total volume temporal profiles are developed from analysis of local area automated traffic recorder (ATR) data (available from State DOTs)

10. 10 Denver Example Total Volume Temporal Profiles

11. 11 CONCEPT Vehicle Mix Disaggregation TDM link volumes are generally total across all vehicle classes CONCEPT disaggregates total volume into eight MOBILE5 vehicle classes Requires hourly VMT mix profiles by roadway type, month, day of week, hour of day VMT mix temporal profiles are developed from analysis of local area vehicle classification recorder data

12. 12 Denver Example Vehicle Mix Temporal Profile HDDV VMT fraction is lower during am and pm peak traffic hours

13. 13 CONCEPT MV Steps in Estimation of On-Road Motor Vehicle Emissions Temporally allocate VMT to hours Adjust free flow speeds for congestion Spatially allocate links to grid cells Allocate total VMT to 8 MOBILE5 vehicle classes Run MOBILE6 with grid-specific meteorology Apply MOBILE6 emission factors using county inputs for fuel parameters and control programs Speciate emissions for air quality modeling

14. 14 Summary of CONCEPT vs. SMOKE Methods for On-Road Mobile Emissions SMOKECONCEPT MV VMTCounty-level, by roadway type, by seasonLink-level, hour-specific, from Transportation Demand Model (TDM), adjusted using hour/day/month temporal profiles from traffic counter data SpeedsEPA national defaults by roadway type and vehicle class if data not submitted by States Link-specific speeds from TDM, adjusted for congestion using hourly volume/capacity ratio VMT mixUniform across all hours, days. Uses EPA national defaults where data not submitted by States Detailed temporal profiles by hour/day/month from traffic classification monitoring data Temperature and humidity Calculates hourly average across all grid cells in each county Hour- and grid cell-specific temperatures Temporal allocationEmissions are temporally allocated using EPA default national-level profiles based on activity data Link-level hourly VMT estimated from link- specific inputs and hour/day/month temporal profiles from local traffic counter data Spatial allocationCounty-level emissions are allocated to grid cells using Census TIGER files, smaller roads not included Link-based emissions are allocated to grid cells based on link coordinates

15. 15 Spatial Distribution of VMT for Grid Modeling is Very Different When Link VMT is Used Link VMT gridded to 1km VMT distributed across more roads – more detailed spatial coverage Max VMT = 301 Max VMT = 89 County-level VMT gridded to km Allocation based on Census TIGER roadway surrogates; creates VMT “hot spots”

16. 16 Denver O 3 SIP CONCEPT MV Application

17. 17 Improved Denver Ozone Performance using CONCEPT MV Link-Based Emissions Denver 8-hour ozone Early Action Compact (EAC) SIP (2002-2003) allocated county-level mobile emissions using Census TIGER roadways surrogates – Ozone suppressed in Denver metropolitan area every day of the episode leading to underprediction bias New Denver 8-hour ozone SIP modeling using CONCEPT MV to generate spatially and temporally highly resolved inventory does not exhibit such a large ozone hole – Better distinguishes weekend effect when high ozone occurs in metropolitan Denver

18. 18 Example 8-h ozone model performance from Denver 8- h ozone EAC SIP modeling Ozone “hole” occurs every day Modeled ozone “hole” in metropolitan Denver of 55-60 ppb ozone where some observed values > 70 ppb

19. 19 Example 8-h ozone model performance for new Denver SIP modeling using CONCEPT MV (Saturday) Model now able to reproduce high ozone in metropolitan Denver

20. 20 CONCEPT MV Regional Application: 22 State and Local Networks in LADCO Domain