Not logged in - Login
< back

Nonroad Mobile Workgroup

Collaborative Wiki Main Page

Inventory Collaborative Google Drive Link

Members

Co-leads: Sarah Roberts (EPA OTAQ), Rebecca Simpson (Colorado Department of Public Health and Environment)

Alexandra Catena (District of Columbia), Alison Eyth (EPA OAQPS), Andy Bollman (North Carolina), Brian Sullins (Alabama), Brian Timin (EPA OAQPS), Brian Trowbridge (Pennsylvania), Chris Bovee (Wisconsin), Chris Kite (Texas), Chris Rochester (New York), Collin Smythe (Vermont), Dale Wells (Colorado), Gil Grodzinsky (Georgia), James Smith (Tennessee), Jim Koroniades (New Jersey), Ken Santlal (Massachusetts), Marc Bennett (Massachusetts), Mark Janssen (LADCO), Mike Maleski (Ohio), Peter Verschoor (Utah), Sonya Lewis-Cheatham (Virginia), Susanne Cotty (Pima County, AZ), Sylvia Vanderspek (California), Tim Wallace (Maryland), Tom Moore (WESTAR)

Work Group Meetings

November 29, 2018

Attendees

Co-lead: Sarah Roberts (EPA OTAQ)

Alison Eyth (EPA OAQPS), Andy Bollman (North Carolina), Brian Sullins (Alabama), Brian Timin (EPA OAQPS), Chris Bovee (Wisconsin), Collin Smythe (Vermont), Dale Wells (Colorado), Gil Grodzinsky (Georgia), James Smith (Tennessee), Mark Janssen (LADCO), Mike Maleski (Ohio), Sonya Lewis-Cheatham (Virginia), Debbie Wilson (MARAMA), Kusandra King (North Carolina), William Kenney (Ohio)

Agenda

  • Nonroad beta inventory
    • Posted materials
    • County-level equipment populations
  • Nonroad v1 inventory
    • v1 timeline
    • Geographic Allocation Updates for v1
  • Discussion
  • Other business

Minutes/Notes

Nonroad beta inventory: EPA has compiled the 2016, 2023, and 2028 beta nonroad inventories. Summary files are posted on EPA's ftp: ftp://newftp.epa.gov/air/emismod/2016/beta/reports/nonroad/

Nonroad v1 inventory: EPA is aiming to start the v1 nonroad runs in mid-February, so all updates, including data supplied by States and the workgroup's geographic allocation updates, should be provided to EPA by January 15th. State submissions should be sent to Sarah Roberts and Alison Eyth.

November 1, 2018

Attendees

Co-leads: Sarah Roberts (EPA OTAQ), Rebecca Simpson (Colorado Department of Public Health and Environment)

Alison Eyth (EPA OAQPS), Andy Bollman (North Carolina), Brian Timin (EPA OAQPS), Chris Bovee (Wisconsin), Chris Kite (Texas), Dale Wells (Colorado), Gil Grodzinsky (Georgia), Jim Koroniades (New Jersey), Ken Santlal (Massachusetts), Mark Janssen (LADCO), Sonya Lewis-Cheatham (Virginia), Debbie Wilson (MARAMA)

Agenda

Minutes/Notes

Nonroad beta inventory: EPA has compiled the 2016 beta inventory; 2023 and 2028 beta runs have been initiated. The beta runs use MOVES2014b model defaults (no state-supplied data included). Sarah shared maps by equipment sector showing county-level difference in NOx emissions from the alpha and beta inventories. (Alpha inventory compiled with MOVES2014a and state-supplied data.). Across all nonroad SCCs, national NOx emissions in the beta inventory are 10.3% lower than in the alpha inventory, primarily due to equipment population changes stemming from the engine population growth update in MOVES2014b. Percent increase/decrease in national NOx inventory from alpha and beta by equipment sector vary:

  • Agriculture: -6.17%
  • Construction: -10.42%
  • Lawn & Garden: -2.21%
  • Recreational: -42.01%
  • Industrial: -4.26%
  • Commercial: -22.38%
  • Pleasure Craft: -15.49%
  • Logging: -6.78%
  • Underground Mining: +1.07%
  • Railway Maintenance: -10.18%

Mark Janssen raised the idea to compare nonroad fuel consumption output from the beta run to other data sources (e.g., USDA), to assess if there is any systematic bias in the results. Could then possibly make input adjustments for v1. EPA to investigate extracting fuel consumption output. Also related to comparing model inputs/outputs to external data sources, Rebecca Simpson noted that she recently found an off-road registration database maintained by the Colorado Department of Revenue. The equipment populations in the CO database appear to compare well to MOVES-Nonroad, but she will also try to compare the equipment age distributions between the two sources. Brian Timin suggested comparing EPA's results from California to various California equipment registries. Alison Eyth confirmed that EPA runs MOVES-Nonroad for California, although results are not used in the platform or included in the output files posted for review.

The Nonroad Workgroup's Google Drive folder contains a full set of maps by pollutant. For each pollutant and equipment sector, four maps are supplied: (1) output from the 2016 alpha inventory (map suffix “_2016fe”), (2) output from the 2016 beta inventory (map suffix “_2016ff”), (3) difference in tons between the beta and alpha inventories (map suffix “_2016ff_minus_2016fe”), (4) percent difference between the beta and alpha inventories (map suffix “_2016ff_minus_2016fe_percent”). A summary output file comparing alpha and beta runs by state-SCC and county is also saved to Google Drive. EPA will also post for review population and activity output, as well as output from the 2023 and 2028 beta nonroad runs will be uploaded to Google Drive by mid-November. UPDATE: suite of nonroad beta inventory files, including maps and summary files, are available for review on EPA's ftp (copy and paste link: ftp://newftp.epa.gov/air/emismod/2016/beta/reports/nonroad/).

Nonroad v1 inventory: EPA is aiming to start the v1 nonroad runs in late-February,mid-February, so all updates, including data supplied by States and the workgroup's geographic allocation updates, should be provided to EPA by January 15th. State submissions should be sent to Sarah Roberts and Alison Eyth.

Geographic Allocation Updates for v1:

  • Agricultural Equipment: The USDA will not provide EPA with an advanced copy of the 2017 Census of Agriculture, so the allocation update will need to use the 2012 Census. NC DEQ will start compiling updated allocations based off of the proposal presented to the workgroup in September.
  • Construction Equipment: Sarah working on improving allocations discussed last month (using County Business Patterns dataset of number of business establishments corresponding to NAICS 237) by incorporating state-level data from the 2012 Economic Census (cost of off-highway use of fuel for NAICS 237). This will require the development of state-level Area Modification Factors. Sarah also developing county-level allocations based on 2014 NEI fugitive dust emissions (measure of acres disturbed by residential, non-residential, and road construction).
  • Lawn & Garden Equipment: no update
  • Recreational Equipment: Rebecca continuing to research data sources.

Mark Janssen, LADCO: Importance of Offroad to Ozone Formation in Midwest: Mark shared results from a source apportionment study that looked at the contribution of onroad and nonroad sources to modeled ozone concentrations. First noted that nonroad sources are 2% of the VOC inventory in the Midwest and 19% of the NOx inventory. Source apportionment modeling, however, shows disproportionately large contributions from nonroad sources (~ 38 ppb, vs 24 ppb from onroad sources), with largest values over the lakes (especially Lake Michigan). Found that the primary contribution was from recreational marine sources (confirmed that other nonroad equipment is not incorrectly allocated to out over lake, so was able to isolate the issue to recreational marine). It follows that elevated emissions over the lake could be due to too many recreational marine vessels allocated to the area and/or transport and chemistry happening over lake. Further analysis suggests that usage over Great Lakes does not seem incorrectly modeled based on the model's allocations. Mark did, however, note that the fraction of activity between urban and rural counties in Wisconsin seems off. Will need to look into other reasons for why nonroad is creating such a large impact over Lake Michigan. Additional information available in LADCO's Interstate Transport Modeling for the 2015 NAAQS Technical Support Document.

Action Items

  • All: review Construction allocation comparison spreadsheet; send any feedback to Sarah
  • Sarah: develop allocation factors based on 2014 NEI fugitive dust allocations
  • Sarah: refinine Construction allocation factors based on CBP data by incorporating 2012 Economic Census (including AMFs)
  • Andy: compile updated allocations for Agriculture, based on methodology presented to the workgroup in September
  • EPA: extract fuel consumption output from beta run, for possible comparison to external data sources
  • Rebecca: continue researching data sources for allocating recreational equipment
  • Sarah: reschedule December 27th workgroup meeting to January 3rd

Next call: Thursday, November 29 3:30pm EST

September 27, 2018

Attendees

Co-leads: Sarah Roberts (EPA OTAQ), Rebecca Simpson (Colorado Department of Public Health and Environment)

Alison Eyth (EPA OAQPS), Kusondra King (North Carolina), Brian Timin (EPA OAQPS), Chris Bovee (Wisconsin), Chris Kite (Texas), Chris Rochester (New York), Collin Smythe (Vermont), Dale Wells (Colorado), Gil Grodzinsky (Georgia), James Smith (Tennessee), Jim Koroniades (New Jersey), Mark Janssen (LADCO), Mike Maleski (Ohio), Sonya Lewis-Cheatham (Virginia), Debbie Wilson (MARAMA)

Agenda

  • EPA Update
    • Beta inventory with MOVES2014b
  • Geographic Allocation Updates for v1
  • Discussion
  • Other business

Minutes/Notes

EPA Update: EPA has initiated 2016, 2023, and 2028 nonroad beta inventory runs using MOVES2014b. As part of testing running MOVES2014b on Linux (cloud platform), Linux output from four counties (Cook, IL, Salt Lake, UT, Fulton, GA, and Lancaster, PA) were compared with output generated by MOVES2014b run on Windows. The Linux and Windows results matched, and Sarah shared plots of the four-county MOVES2014b results with results from the same four counties run with the previous model version (MOVES2014a), for 2016, 2023, and 2028 (NOx, PM2.5, VOC; Agricultural, Construction, Recreational, and Lawn & Garden categories). Results vary by county, but in most cases MOVES2014b produced lower inventories. The nonroad beta inventory will be posted on EPA's ftp site in October.

v1 Geographic Allocation Update: No update from the Construction, Recreational, and Lawn & Garden Sectors. Kusondra King (North Carolina) presented a proposal to update agricultural equipment allocations, using data from USDA's Farm Production Expenditures Summary and USDA's Census of Agriculture. The proposed allocation steps are:

  1. Develop agricultural equipment allocation data for Alaska and Hawaii based on the 2012 Census of Agriculture’s “Gasoline, Fuels, and Oils Purchased” dollar values
  2. Allocation data for all other states are developed at the regional-level, for the 5 regions for which the 2016 Farm Production Expenditures Summary develops data (i.e., Atlantic, South, Midwest, Plains, and West).  These data will be based on “Petroleum Fuel and Oil” expenditures in each region.
  3. Allocate regional proportions further down to the state-level for the 15 states for which “Petroleum Fuel and Oil” expenditures data are reported in the 2016 Farm Production Expenditures Summary report.
  4. Use the 2012 Census of Agriculture’s “Gasoline, Fuels, and Oils Purchased” dollar values to allocate the state/regional data compiled above down to the county-level. 

Kusondra pointed out that while the latest version of the Census of Agriculture is from 2012, USDA is set to release the next (2017) version in February 2019. Sarah will look into if EPA can access the 2017 Census data early. In the meantime, we can continue to implement the above method using the 2012 Census.

Discussion: Mark Janssen is interested in updating the geographic allocation of recreational marine equipment, as MOVES places too many vessels on Lake Michigan. Sarah will provide Mark with the MOVES2014 tables related to updating allocations (nrstatesurrogate, nrsurrogate, nrequipmenttype) and Mark will aim to present a proposal to the work group at the October meeting.

Mark also mentioned that beta inventory will likely be used by eastern and Midwestern states for ozone SIPs

Gil Grodzinsky shared with the group some observations from running MOVES2014b inventories for 13 counties around Atlanta. Gil noted no changes to onroad inventories and a decline in MOVES2014b nonroad inventories. Gil also confirmed updates to the county table to reflect actual county updates, and that nonroad fuels now go out to 2060 (NB: EPA extended fuels to 2060 in advance of a future model update when the entire MOVES model will include modeling capabilities to 2060).

Action Items

  • All: review Construction allocation comparison spreadsheet; send any feedback to Sarah
  • Sarah: send Mark Janssen copies of the MOVES2014b tables relevant to geographic allocation update (nrstatesurrogate, nrequipmenttype, nrsurrogate)
  • Sarah: analyze fugitive dust allocation factors for potential application to the Construction sector
  • Susanne: start to formulate how annual rainfall data could be used to scale allocations of residential L&G equipment
  • Rebecca: continue researching data sources for allocating recreational equipment
  • Sarah: compile 2016 CBP data (number of employees in landscaping sector) for NAICS 561730; compare to values currently used in Nonroad
  • Sarah: look into EPA accessing 2017 Census of Agriculture data in advance of USDA's February release date

Next call: Thursday, October 25 3:30pm EDT

August 23, 2018

Attendees

Co-leads: Sarah Roberts (EPA OTAQ), Rebecca Simpson (Colorado Department of Public Health and Environment)

Alison Eyth (EPA OAQPS), Andy Bollman (North Carolina), Kusondra King (North Carolina), Brian Sullins (Alabama), Brian Timin (EPA OAQPS), Chris Bovee (Wisconsin), Dale Wells (Colorado), Gil Grodzinsky (Georgia), Marc Bennett (Massachusetts), Mark Janssen (LADCO), Mike Maleski (Ohio), Susanne Cotty (Pima County, AZ), Tim Wallace (Maryland), Debbie Wilson (MARAMA)

Agenda

  • EPA Update
  • Progress in Updating Geographic Allocation of Construction, Agriculture, Lawn & Garden, and Recreational Equipment
  • Discussion
  • Other business

Minutes/Notes

EPA Update: Sarah confirmed that EPA will be releasing MOVES2014b during the last week of August or first week of September. EPA is getting MOVES2014b set up to run in the cloud, and this latest version of the model will be used to generate 2016 beta, 2016 v1, and 2017 NEI nonroad inventories. EPA plans to kick-of the 2016 beta (including 2023 and 2028) runs in the next few weeks.

Updating Geographic Allocation Factors for v1:

Agriculture: Andy Bollman discussed NC DEQ's work to identify and evaluate candidate data sources for the Agriculture sector. First considered some of the data sources listed in the Excel file shared with the group last month, and concluded that some sources were inferior; for example, equipment count information was limited to only certain types of equipment and didn't indicate fuel type used. Also, County Business Pattern's (CBP) number of employees didn't seem like a good choice. Decided to focus efforts on fuel-related data from USDA and EIA (FOKS). Compiled:

  • 2012 (latest year data) of harvest crop acreage from Census of Agriculture. The Nonroad model currently uses this 2002 dataset as the allocation factor for agriculture.
  • USDA 2016 Farm Production Expenditures (diesel, gasoline, LPG, other). Annual survey that is available for 5 regions (Atlantic, South, Midwest, Plains, West). Sub-regional data for 10 sub-regions only available in 2003. Farm production expenditures available only for 48 contiguous states. "Petroleum fuel and oil" expenditures for 2016 available for 5 regions and 15 core states.
  • USDA 2012 Census of Agriculture (gasoline, fuels, oils purchased). Doesn't contain fuel-specific information
  • EIA 2016 FOKS (fuel delivered to farm consumers)

Calculated percentage of total national activity by region/state for each data source. Allocation percentages using harvested crop acreage generally don't align with the fuel-based data sources. For example, in CA, using harvest crop acreage to allocation national activity would allocate 2.5% to CA; using fuel data would allocate 9.4-9.9% to CA. This could indicate that the crops grown in CA might be more fuel-intensive than crops in other parts of the country. The opposite signature is seen for Midwestern states. NC generally has more confidence in USDA fuel data over EIA data (while EIA surveys fuel suppliers, USDA surveys end users and is therefore a more bottom-up estimate) and suspects that a fuel surrogate would be an improvement over the harvest acres metric.

Mark Janssen emphasized the utility of crop calendars to understand agricultural activity, and Brian Timin asked if any of the data sources included fuel sales by month. Andy indicated that monthly data are not available.

Alison Eyth asked how we might use the various data sources to re-allocate equipment to the county level. Andy thought that perhaps using a combination of Farm Production Expenditures to allocate to the core states/regions (2016) and Census of Agriculture's Gasoline, Fuels, and Oils Purchased at the county-level (2012) might be appropriate. NC will look into how to best utilize the USDA data.

Gil Grodzinsky suggested, should time and resources allow, updating the model's month allocation defaults.

Construction:

Sarah Roberts walked through a spreadsheet that compares the Nonroad model's current allocations for the construction sector with allocations based on 2016 County Business Patterns data. As defined in the nrsurrogate and nrstatesurrogate tables, the model currently uses dollar value of construction (adjusted for relative geographic cost of construction material) from 2003 to allocate construction equipment to the county level. These 2003 allocations are compared with those based on 2016 CBP data of the number of establishments corresponding to NAICS 237 (heavy and civil engineering construction) at the state and county levels. NAICS 237 was chosen following a separate EPA analysis that looked at the correlation between state-level economic variable and FOKS (No. 2 diesel adjusted sales for off-highway construction) and found that in the CBP dataset, NAICS 237 had the highest correlation with FOKS (at the state level).

Work Group members are encouraged to review the comparisons for their state/region, and provide feedback to Sarah.

Dale Wells emphasized the need to break down construction by type (highway, commercial, industrial), noting that this breakdown is already available in the county allocations used for fugitive dust. Sarah will look into using the fugitive dust dataset for construction allocations.

Recreational:

Rebecca Simpson has identified a handful of promising studies and datasets, including a Colorado-specific study from 2014-2015 looking at the economic impact of recreational vehicles, as well as presentations by California Air Resources Board (CARB) related to CARB's Off-Highway Recreational Vehicles Red Sticker Program. CARB observes in an April 2018 presentation that annual sales of Off-Highway Motorcycles correlate well with nationwide housing starts. Rebecca will continue researching possible data sources.

Lawn and Garden:

Susanne Cotty has been looking at this sector primarily from the lens of coming from an arid western state, and is advocating for the use of annual rainfall as a scaling factor for residential L&G equipment. For Commercial L&G equipment, the Nonroad model's current approach of allocating equipment based on the number of employees in the landscaping sector (NAICS 561730) is likely sound, although this dataset could be updated with 2016 CBP values.

For residential L&G equipment, Susanne has considered looking at sales of lawn and garden equipment (e.g., Home Depot, Lowe's), but sales data are cost-prohibitive and rarely available at the sub-national scale. Also of potential interest is NASA's work identifying lawns from satellite imagery. Additionally, Rebecca has come across some related studies, including a Milesi et al. study relating turf grass area to fractional impervious surface area.

Finally, for the September 27 work group meeting, Mark Janssen offered to discuss some of LADCO's recent work in the Recreational Marine sector. Massachusetts has also expressed an interest in undertaking improvements in this sector.

Action Items

  • All: review Construction allocation comparison spreadsheet; send any feedback to Sarah
  • Sarah: analyze fugitive dust allocation factors for potential application to the Construction sector
  • Andy: start to formulate how USDA data could be best used to allocate Agricultural equipment
  • Susanne: start to formulate how annual rainfall data could be used to scale allocations of residential L&G equipment
  • Rebecca: continue researching data sources for allocating recreational equipment
  • Sarah: compile 2016 CBP data (number of employees in landscaping sector) for NAICS 561730; compare to values currently used in Nonroad

Next call: Thursday, September 27 3:30pm EDT

July 26, 2018

Attendees

Co-leads: Sarah Roberts (EPA OTAQ), Rebecca Simpson (Colorado Department of Public Health and Environment)

Andy Bollman (North Carolina), Brian Sullins (Alabama), Brian Timin (EPA OAQPS), Chris Bovee (Wisconsin), Chris Kite (Texas), Dale Wells (Colorado), Gil Grodzinsky (Georgia), Ken Santlal (Massachusetts), Marc Bennett (Massachusetts), Mike Maleski (Ohio), Sonya Lewis-Cheatham (Virginia), Susanne Cotty (Pima County, AZ), Tim Wallace (Maryland)

Agenda

Minutes/Notes

EPA Update: Sarah reiterated that for the 2016 beta nonroad inventory, EPA will run MOVES2014b using model defaults. State-supplied data and any additional updates will be included in the v1 run to be conducted later this year/early2019. OTAQ is on track to release MOVES2014b in late August/early September. The model release will be announced via the EPA-MOBILENEWS Listeserv.

Updating geographic allocation of Construction, Agriculture, Lawn & Garden, and Recreational equipment: Sarah shared some recent work on the topic performed by an EPA contractor tasked with identifying and evaluating data sources that could be used as surrogates to allocate national equipment populations (see above hyperlink). Each data source was evaluated on its contents (geographic coverage, alignment with equipment place of residence vs. where equipment is used) and quality/reliability (sampling methodology, QA procedures, availability, documentation, limitations). Lawn and Garden and Agriculture sector fuel consumption data from FHWA's Highway Statistics are deemed unusable, as FHWA uses the Nonroad Model in its estimates. Datasets from the U.S. Census Bureau (e.g., County Business Patterns) and the U.S. Department of Agriculture (Ag Census) are generally available at the county level and are well documented. Industry groups like the Motorcycle Industry Council and the International Snowmobile Manufacturers Association are also good sources of state-level population data.

Susanne Cotty raised the concern of allocating lawn & garden equipment to arid counties with little rainfall and greenspace (e.g., Pima County) and suggested that the model use annual rainfall values to define a threshold below which only scant equipment are allocated. Susanne also noted that recent surveys have indicated that 40% of homeowners employ commercial landscaping services, raising the possibility that some L&G equipment might be double-counted by the model. Andy Bollman clarified that the Power Systems Research population data used in the model splits out the commercial and residential equipment. Rebecca Simpson also noted that an uptick in Homeowners Associations employing commercial landscaping services might also lead to an over-allocation of residential L&G equipment to these residential areas, if single family housing is used as the surrogate to allocate residential L&G equipment. Rebecca also suggested we consider employing GIS methods to calculate greenspace/lawn coverage in the L&G sector.

Gil Grodzinsky discussed Georgia's work to update nonroad geographic allocations, including a walk-through of the Nonroad tables (e.g., nrequipmenttype, nrsurrogate, nrstatesurrogate) and data fields (e.g., sectorID, surrogateID, surrogatequant) relevant to the task of updating geographic allocations (see hyperlink above). Gil described the data sources Georgia used in their work, including the U.S. Census Bureau's American Fact Finder, the U.S. Forest Service's Forest Inventory and Analysis National Program, the U.S. Department of Agriculture's Census of Agriculture, and financial reports available from the State of Georgia (note that an Area Modification Factor, available from Craftsman Book Company, must be applied to all economic/financial data, in order to adjust for the local value of the dollar). Georgia's updated allocations were compared with Nonroad's default allocations by running MOVES and comparing results for the 13-county Atlanta area. Georgia's updated values resulted in an 8% increase in NOx emissions (driven primarily by the Construction sector) and a 12% decrease in VOC emissions (driven by the Lawn & Garden sector).

Action Items

  • Rebecca, Susanne, Gil, Ken, Andy, and Sarah: research data sources; begin to develop recommendations for the work group
  • Sarah: schedule short check-in meeting with volunteers working on compiling and evaluating geographic surrogates

Next call: Thursday, August 23 3:30pm EDT

June 28, 2018

Attendees

Co-leads: Sarah Roberts (EPA OTAQ), Rebecca Simpson (Colorado Department of Public Health and Environment)

Andy Bollman (North Carolina), Brian Sullins (Alabama), Brian Timin (EPA OAQPS), Chris Bovee (Wisconsin), Chris Kite (Texas), Dale Wells (Colorado), Gil Grodzinsky (Georgia), James Smith (Tennessee), Ken Santlal (Massachusetts), Mike Maleski (Ohio), Sonya Lewis-Cheatham (Virginia), Susanne Cotty (Pima County, AZ)

Agenda

  • New work group co-chair!
  • EPA Update: MOVES2014b
  • Running MOVES2014b for beta inventory: model defaults vs. incorporating 2014 submittals
  • Updating Nonroad model inputs for the 2016 inventory
  • Other business

Minutes/Notes

MOVES2014b: Sarah provided results from EPA's analysis comparing MOVES2014a with MOVES2014b (updated growth indices, Tier 4 emission rates and populations, and diesel sulfur levels), by fuel type and equipment sector. Overall, emissions decrease for almost all pollutants in almost all equipment sectors (exceptions: Oilfield and Industrial sectors; CNG). OTAQ is planning to release the model later in the summer.

Preparing for 2016 beta version: discussion of if the beta run later this summer using MOVES2014b should include state-supplied data or be run with just model defaults, so as to provide States the opportunity to review in better detail the inventory impacts of the model updates. Because updated growth rates are state-specific, there was general consensus to use model defaults for the beta inventory. Massachusetts has recreational marine populations (based on registrations) for inclusion in beta version; Georgia has updated fuel (RVP) information that must be incorporated.

Updating high-emitting nonroad categories: work group previously identified wanting to improve the equipment populations of highest-emitting categories (construction, agriculture, lawn & garden, recreational) for the 2016 inventory. As EPA has already refined equipment population growth rates, the work group will embark on an effort to refine spatial allocations (national-state; state-county). There's a paucity of county-level data to use as surrogates, but County Business Patterns and data from MPOs are a good starting point. Andy Bollman, Rebecca Simpson, Gil Grodzinsky, Ken Santlal, Susanne Cotty, and Sarah will investigate and compile possible data sources. Sarah to circulate some work that EPA has recently done in this area.

Action Items

  • Sarah: circulate EPA's draft list of spatial surrogate data sources
  • Rebecca, Susanne, Gil, Ken, Andy: review EPA's draft list of data sources; search for additional datasets
  • Ken: provide EPA (Sarah and Alison) with recreational marine data for inclusion in beta run
  • Sarah: confirm Georgia's updated RVP values are in 2016 fuel tables

Next call: Thursday, July 26 3:30pm EDT

May 24, 2018

Attendees

Co-leads: Sarah Roberts (EPA OTAQ) and TBD

Alison Eyth (EPA OAQPS), Andy Bollman (North Carolina), Brian Timin (EPA OAQPS), Chris Bovee (Wisconsin), Chris Kite (Texas), Collin Smythe (Vermont), Dale Wells (Colorado), Gil Grodzinsky (Georgia), Ken Santlal (Massachusetts), Kevin Briggs (Colorado), Marc Bennett (Massachusetts), Mark Janssen (LADCO), Mike Maleski (Ohio), Rebecca Simpson (Colorado), Sonya Lewis-Cheatham (Virginia), Susanne Cotty (Pima County, AZ), Tim Wallace (Maryland)

Agenda

  • EPA Update: MOVES2014b
  • Proposal to aggregate Nonroad SCCs for reporting nonroad emissions
  • Updating Nonroad model inputs for the 2016 inventory
    • Available resources from the work group
    • Targeting equipment sectors and model inputs
    • Data sources
  • Other Business
    • Preparing for beta version
    • Call for volunteers to serve as work group co-chair
    • Questions and discussion

Minutes/Notes

EPA update (MOVES2014b): Sarah reiterated that EPA will be showing the results of their MOVES2014a vs MOVES2014b comparisons at the next MOVES Review Work Group meeting, on June 13th. Andy Bollman asked if EPA would make state-level comparison results available for review. Sarah responded that EPA currently has comparison results at the national level and for four select counties, but will speak with the MOVES team about spinning up state-level results.

Proposal to aggregate nonroad SCCs: Sarah walked through an EPA proposal to aggregate nonroad SCCs by fuel type and equipment sector, in an effort to get the nonroad emissions file (currently at 1.0 - 1.6 GB compressed) to a more manageable size. EPA is seeking to understand if users need the more detailed emissions that are currently available, or if aggregated emissions would suffice. Brian Timin stated that he uses detailed information to better pin-point emissions. Alison Eyth clarified that EPA would continue to produce output at the detailed SCC level (and those results would still be available), but that the more aggregated form would be put into SMOKE. Work group members are asked to review the proposal (see hyperlink above) and provide any feedback to Sarah.

Updating model inputs for 2016 inventory: workgroup members expressed a clear interest in developing updated inputs for the 2016 inventory. On a scale of 1 to 5 (1 = no interest; 5 = high interest), over 90% of those in attendance indicated a high level of interest (4-5 on the scale) in developing inputs. Furthermore, workgroup members indicated that they or their organizations would be willing to commit resources to the effort, on the order of 1-8 hours per week (average = 2.9 hours per week). For the most part, workgroup members felt that they would be able to best contribute to the effort by QAing results/data and developing documentation. There was also some interest in compiling data and sources. With respect to which model inputs the workgroup should focus its efforts, equipment populations and state-to-county allocations generated the most interest. Dale Wells pointed out that with respect to surrogate data for allocating state-level equipment populations, there could be some overlap with the NEI nonpoint sector, particularly in the agriculture and construction sectors. The equipment sectors which garnered the most interest were construction, agriculture, lawn & garden, and recreational. Some workgroup members indicated an awareness of or familiarity with some available data sources such as pleasure craft and recreational (ATVs, snowmobiles) registrations and some surveys in the lawn & garden and construction sectors (in addition to LADCO's work in the agriculture sector). Mark Janssen suggested an approach of focusing discussion on the highest-emitting categories and talk through as a group how we want to tackle data improvements. This approach will form the basis of the discussion for the June 28th call, which the group broadly agreed should run for 90 minutes instead of one hour.

Action Items

  • All: if you're planning to submit data for inclusion in the beta inventory, please contact Sarah Roberts and Alison Eyth to indicate what inputs you intend to provide
  • All: review proposal to aggregate nonroad SCCs and provide feedback to Sarah
  • Sarah: check with MOVES team about providing state-level MOVES2014a vs. MOVES2014b comparisons
  • Sarah: send updated meeting notice for extended meeting on June 28th.

Next call: Thursday, June 28 3:30pm EDT

April 26, 2018

Attendees

Co-leads: Sarah Roberts (EPA OTAQ), Joseph Jakuta (Ozone Transport Commission)

Alison Eyth (EPA OAQPS), Andy Bollman (North Carolina), Chris Bovee (Wisconsin), Chris Kite (Texas), Chris Rochester (New York), Collin Smythe (Vermont), Dale Wells (Colorado), Gil Grodzinsky (Georgia), James Smith (Tennessee), Ken Santlal (Massachusetts), Mark Janssen (LADCO), Mike Maleski (Ohio), Rebecca Simpson (Colorado), Sonya Lewis-Cheatham (Virginia), Tim Wallace (Maryland)

Agenda

Minutes/Notes

EPA still targeting a summer release of MOVES2014b, with the goal of using the updated model for the 2016 beta inventory in late summer 2018. EPA will present results of their MOVES2014b testing, including nonroad inventory comparisons with MOVES2014a, at the June 13th meeting of the MOVES Review Work Group (to be added to the MOVES Review Work Group distribution list, please email Sarah). Draft MOVES2014b nonroad technical reports and related peer-review materials are posted on EPA’s Science Inventory page (search “MOVES201X”).

Reminder that the 2016 alpha release, along with 2014 and 2015 platform materials are available for download from EPA’s FTP site.

Chris Kite (Texas Commission on Environmental Quality) gave a presentation on TexN, the ERG-developed tool that TCEQ uses to facilitate the inclusion of regional updates (e.g., activity, population data collected in specific counties) to estimate nonroad emissions for Texas. This integrated calculation and data management system drastically reduces the effort associated with NONROAD file preparation, model execution, and output file aggregation. The tool automates the running of NONROAD for all 254 Texas counties (a single run usually completes within 24 hours) and contains 25 distinct sub-sectors with distinct equipment population and activity profiles (diesel construction equipment comprises most of these sub-sectors; miscellaneous equipment less than 25 hp and all non-diesel construction equipment uses EPA’s NONROAD defaults). Chris reviewed TexN output in a series of plots that highlighted trends in equipment populations and emissions by Texas geographic area, fuel/engine type, and certification standard.

In addition, Chris identified some questions for the work group to consider for 2016 modeling platform needs:

1. Are monthly nonroad inventory runs required or are season runs acceptable? The temporal allocation of nonroad activity is uniform for months within the same season.

2. For each season/month modeled, are day type inventories needed for both weekend and weekday? TCEQ develops weekday inventories only, and then applies separate temporal profiles during emissions processing to obtain weekend inventories.

Mark Janssen (LADCO) asked about TexN handling of age distributions and scrappage curves. Except for skid-steer loaders, NONROAD’s scrappage curves were maintained.

Subsequent discussion centered around the need to transition the work group’s efforts to improving estimates from select equipment categories (e.g., expanding LADCO’s work utilizing agricultural data to refine activity profiles). The work group call on May 24 will be dedicated primarily to coming up with a plan for how we might apply some “lessons learned” from state/local work to this national inventory. We’ll focus on defining the target equipment sector (e.g., construction, agriculture) and 2016 input data (e.g., activity, population, allocation), identifying data sources (direct and/or surrogate), and determining level of effort and available resources.

Reminder that States and Locals are invited to provide data they’d like to have incorporated into the 2016 beta inventory. Data can be provided in spreadsheets and supplied to Sarah Roberts and Alison Eyth.

Joseph Jakuta will soon be leaving his position at OTC, and unfortunately must step down as the co-chair of this work group. The work group thanks Joseph for his contributions to the work group, and hopes he can find a way to continue to participate in the effort. Work group members who are interested in volunteering to serve as the next co-chair should contact Sarah.

Action Items

  • All: If you'd like to volunteer to serve as the workgroup's co-chair, please contact Sarah
  • All: if you're planning to submit data for inclusion in the beta inventory, please contact Sarah Roberts and Alison Eyth to indicate what inputs you intend to provide
  • All: if you would like to present on local efforts to improve nonroad emissions estimates, please let Sarah know by COB May18th

Next call: Thursday, May 24 3:30pm EDT

March 22, 2018

Attendees

Co-leads: Sarah Roberts (EPA OTAQ), Joseph Jakuta (Ozone Transport Commission)

Alison Eyth (EPA OAQPS), Andy Bollman (North Carolina), Brian Sullins (Alabama), Chris Bovee (Wisconsin), Chris Kite (Texas), Chris Rochester (New York), Dale Wells (Colorado), Gil Grodzinsky (Georgia), James Smith (Tennessee), Ken Santlal (Massachusetts), Mark Janssen (LADCO), Mike Maleski (Ohio), Peter Verschoor (Utah), Rebecca Simpson (Colorado), Sonya Lewis-Cheatham (Virginia), Susanne Cotty (Pima County, AZ), Tim Wallace (Maryland)

Agenda

Minutes/Notes

EPA is targeting a summer release of MOVES2016b for use in 2016v1 (and perhaps the beta version), with a primary focus on improving nonroad inventories. The improvements involve recalculations of engine populations from refining the engine population growth rates, a bug fix in the nonroad fuel supply, and improvements in Tier 4 and after-treatment data. These are primarily database updates, but some minor code changes were necessary. MOVES2016b will not be considered a new model for SIP and conformity purposes. The release will also resolve an issue involving scratch files that creates problems when running virus scan. The new growth rates are comparable to what was submitted by a handful of states as part of the recent NODA process. EPA will share results of MOVES2014b testing with the workgroup as they become available.

2016 alpha release is now available for download from EPA’s FTP site. Joseph reviewed a summary of NOx, PM2.5 and SO2. Agriculture is the largest category and has the most regional variation. Construction equipment is the next more important category. Pleasure craft, commercial, industrial and lawn and garden are around the same magnitude as the third most important.

Mark Janssen reviewed slides on some of LADCO’s efforts to rethink NONROAD. Focused on “following the money” to determine the best source for population and activity data. Agricultural equipment was a prime category, and Department of Agriculture data estimating the amount of fuel required for agricultural operations was used as a starting point. LADCO generated a new monthly activity profile with spring and fall peaks (corresponding to planting and harvesting activities), in contrast to EPA’s default monthly activity levels that show a peak during the summer ozone season. Another area of concern is the spatial allocation problem shown by fitting data into counties and the distribution of pleasure craft in the Great Lakes.

Andy Bollman (North Carolina) reviewed slides on efforts to improve growth of agriculture and construction equipment. Began by looking at 2011 nonroad NOx emissions and finding the levels high. Developed surrogate data using sales data from EIA. Construction equipment sales decreased and agriculture equipment decreased then increased. Used new values for years up to 2015 and let NONROAD calculate growth thereafter. Construction equipment in particular varied considerably from NONROAD. Did only look at the years in the model rather than every year which may have resulted in spikes or extreme dips being glossed over.

Dale Wells (Colorado) asked if for the NEI, there were any efforts to harmonize fugitive dust emissions with nonroad activity (see EPA's nonpoint construction dust workbooks here: ftp://ftp.epa.gov/EmisInventory/2014/doc/nonpoint/2014_Construction_Dust_v3.0_18feb2016.zip). Alison Eyth indicated that she was not aware of such efforts, but that that might be worth digging into. Gil Grodzinsky (Georgia) reviewed his work in which he ran 15 counties in GA looking at the impact of updating growth indices on emissions, compared to EPA defaults and a no-growth scenario. Overall NOx reductions occur with no-growth and the updated surrogate, but VOCs only changed for the no growth scenario. A review was conducted of various sectors.

Mark also showed LADCO data illustrating the growth of snowmobiles may be overestimated as well. Peter Verschoor (Utah) also commented that snowmobile populations are declining in Utah. Sarah concurred that NONROAD’s current growth rates for the recreational equipment sector appear to overestimate populations, and that the updated growth rates in MOVES2014b will bring that curve down substantially.

Action Items

  • All: review draft workgroup charge; provide comments to Sarah and Joseph by April 1st, 2018
  • All: if you're planning to submit data for inclusion in the beta inventory run, please contact Sarah, Joseph, and Alison Eyth to indicate what inputs you intend to provide
  • All: if you would like to present on efforts to improve emissions estimates from the Lawn & Garden and Recreational Marine sectors at the April 26th workgroup meeting, please send a note to Sarah and Joseph by COB Monday, April 23rd

Next call: Thursday, April 26 3:30pm EDT

February 8, 2018

Attendees

Co-leads: Sarah Roberts (EPA OTAQ), Joseph Jakuta (Ozone Transport Commission)

Alexandra Catena (District of Columbia), Alison Eyth (EPA OAQPS), Brian Sullins (Alabama), Brian Timin (EPA OAQPS), Brian Trowbridge (Pennsylvania), Chris Bovee (Wisconsin), Chris Kite (Texas), Chris Rochester (New York), Dale Wells (Colorado), Gil Grodzinsky (Georgia), James Smith (Tennessee), Ken Santlal (Massachusetts), Kusandra King (North Carolina), Marc Bennett (Massachusetts), Mark Janssen (LADCO), Mike Maleski (Ohio), Rebecca Simpson (Colorado), Sonya Lewis-Cheatham (Virginia), Stephanie Huber (California), Susanne Cotty (Pima County, AZ), Vanessa Crandell-Beck (Alaska)

Agenda

  • Welcome and Introductions
  • Overview of Inventory Collaborative
    • Organizational Structure
    • Platform Schedule
    • Resources
  • Nonroad Work Group Tasks
  • Planned Updates to NONROAD Model
  • Next Steps

Minutes/Notes

EPA heard from states that they wanted to be more involved in providing inputs in development of inventories, as well as the methods used, in particular when projecting growth. Inventory will be in 3 phases. Alpha will be based on 2014 NEI v2, which has been posted on the ftp site. The nonroad files are rather large and might need to be resized. Beta will include projections to 2023 and 2028 and improvements to 2014. 2016v1 will be completed by winter 2019.

NONROAD standalone model was release about 20 years ago and besides folding it into the MOVES GUI and updating a few emissions rates there have been no improvements. EPA is planning on updating the growth rates that are growing populations or activity from 1996, 1998, or 2000 depending on class in the next update to MOVES. It currently uses a national growth rate, which is based on a linear extrapolation. Planning on using a variety of mostly public data sets including projection surrogates such as US Census population as well as historic data such as fuel and kerosene sales and boat registrations. This new data saw an overall decrease in engine populations in most sectors and reductions in emissions as well. For this effort EPA is considering whether they will provide an updated NONROAD input database or develop adjustment factors to apply after MOVES2014a NONROAD runs are complete.

Gil Grodzinsky conducted a similar test of improving growth rates for some sectors and found similar results for the sectors they analyzed.

It should be noted that aircraft ground support equipment and oil and gas equipment from NONROAD is not used and will not fall under the purview of this group.

Mark Janssen brought up the issue of whether some of this work could be taken outside of the NONROAD model and would like to block off time to hear from others on what they have done in individual states.

In subsequent calls we would like to plan on having members of the workgroup present some of their findings and methodology. Sarah Roberts proposed looking at sectors in order of emissions importance. We will start with agriculture and construction next month.

Datasets should be sent to Sarah Roberts (Roberts.Sarah at epa.gov) and Alison Eyth (Eyth.Alison at epa.gov) at EPA as well as Joseph Jakuta at OTC (jjakuta at otcair.org)

Action Items

  • If you would like to receive information and you are not on the distribution list please contact Sarah Roberts.
  • Everyone should look at the 2016 nonroad inventory and the NEI documentation to understand how base year data is developed.
  • We need to identify areas where data can be improved, populations or activity, including state data, and whether we should target particular sectors.
  • Everyone should become familiar with the updates EPA plans on making to NONROAD in MOVES.
  • Joseph Jakuta and Sarah Roberts will work on a workgroup charge prior to the next call; will circulate to the group
  • Sarah Roberts will send out an email to inform everyone of the next meeting and the sectors to focus on (agriculture and construction).
  • If you would like to share work that’s been done at the local, state, and regional level to improve emissions estimates from the Construction and Agriculture nonroad equipment sectors, please send a note to Sarah and Joseph by COB Monday, March 19th.
  • If anyone has any preferences for dividing the nonroad inventory into smaller sizes for review send them to Sarah and Joseph.

Next Call: Thursday, March 22 3:30 pm EDT