1. University of North Carolina Institute for the Environment
Projecting Base Year Emissions to the 2028 Milestone Year for Regional Haze Planning
Zac Adelman
University of North Carolina Institute for the Environment
Tom Moore
Western States Air Resources Council and Western Regional Air Partnership
Remind about emissions being regional, not a single source or single group of sources

2. High-level trends: What have we seen?
Economic Growth with Cleaner Air (
Between 1970 and 2016, the combined emissions of the six common pollutants (PM2.5 and PM10, SO2, NOx, VOCs, CO and Pb) dropped by 73 percent. This progress occurred while the U.S. economy continued to grow, Americans drove more miles and population and energy use increased.

3. Emissions Projections: What and Why?
What are Emissions Projections?
Projections are predictions/forecasts of future emissions
Projections = activity changes (growth/contraction) + controls (reduction technologies/strategies)
Why do we need to Forecast Emissions?
SIP attainment demonstrations require future year modeling
Regional haze planning requires demonstration of reasonable progress toward natural conditions

4. How are Emissions Forecasts Done? (1)
Historic, simplest approach:
Planning forecasts estimate changes in activity (i.e., number of emitting units) and the application of controls (i.e., emissions control devices or technology turnover)
Timing, location, chemical makeup of the emissions are not forecast for planning
Based on assumptions that:
a large amount of anthropogenic emissions are available to control,
control technology costs and effectiveness are available, and
can be evaluated and effectively implemented by the regulatory agency.
Activity
Temporal
Spatial
Chemical
Emissions Factor
Controls
Components of the emissions forecast
Red boxes highlight the parts of the emissions equation that are used in the simplest forecast method.

5. How are Emissions Forecasts Done? (2)
Second level of activity complexity to consider:
What if sources within a sector are changing, beyond closings and if any new sources are being built, are they identical to existing sources?
EGUs – how is electricity being generated?
Oil & Gas production – formations accessed, techniques, etc.
For mobile, how are electricity and gas supplies changing fleets?
Begins to affect settings in blue boxes as well
Activity
Temporal
Spatial
Chemical
Emissions Factor
Controls
Components of the emissions forecast
Both red and blue boxes would need to be addressed in a second level forecast.

6. How are Emissions Forecasts Done? (3)
Third level complexity across forecast steps to consider:
What if a source sector is already well-controlled in terms of emissions rates, control technology, and compliance?
Accounting for state programs and strategies that are not in the SIP, but limit future emissions
Are economic activity forecasts reliable or representative data to estimate changes in future activity and associated emissions?
How are economic data related to the permitting and compliance programs of the agency(ies) implementing additional controls in individual programs?
How are interrelated sources to be evaluated, such as natural gas production and EGUs, or control decisions vs. closure assumptions for coal EGUs, electric cars in the western grid, et cetera?
Activity
Temporal
Spatial
Chemical
Emissions Factor
Controls
Components of the emissions forecast
Simultaneous, robust evaluation of both red and blue boxes would need to be addressed in a third level forecast.

7. How are Emissions Forecasts Done?
Models and data are / can be used as the basis to project activity (silo approach)
IPM and ERTAC for electricity sector
Travel demand for MOVES
EIA for energy production/distribution
U.S. Census for population and housing
U.S. EPA oil and gas tool
Applying emissions controls
On-the-books and On-the-way control programs
Control Strategy Tool (CoST)
Database of technology controls and costs
Natural emissions not typically forecasted

8. What Sources Can/Cannot Be Forecasted?
Realities:
Climate changing with many ancillary effects
Increases in economic activity occur without emission increases (often decreases)
Ecosystem and market responses, and costs, cause many changes, so emissions mgmt. reductions or increases are often the effect, not the driver.
Availability of Information to use in Forecast
(highly variable in completeness and detail)
Emissions Estimate Technique
None A Lot
Counted
Modeled
EGU
Non-EGU Point
On-road Mobile Ag
Biogenic
Forest Fire
Oil & Gas
Off-road Mobile

9. Facts About Emissions Forecasts
Uncertainty in a base year inventory carries forward to the forecast
Certainty of a forecast decreases with time
Tradeoff between need to forecast for planning and minimizing uncertainty in the forecast
Base and future year inventories are coupled
Forecasts always start from a base year inventory
Air quality modeling forecasts should only show sensitivity to emissions projections, not differences resulting from base year assumptions
Deterministic emissions forecasts will always be wrong

10. Forecast Optimization: CoST
Control Strategy Tool (CoST) is software built on an EPA database of control measures
Control measure: technology, costs to implement and maintain, and applicable sources/pollutants
Control Strategy Development
CoST includes algorithms for developing strategies to reduce emissions from stationary sources
Least-Cost, Maximum Emissions Reduction, Multi-Pollutant Emissions Reduction, Least Cost Curve

11. Regional Haze Planning and Emissions Forecasts
1. Define what is Anthropogenic and Controllable
SOURCE
In-State
Out-of-State
International
Anthropogenic
Controllable
not controllable
Natural
2. Quantify all the emissions contributions on a daily basis for modelling
Here is how to organize thinking before making a forecast. Planning to make a forecast

12. Examples of Emissions Forecast Plots
Stacked bar chart shows changes in total magnitude and contribution by inventory sector

13. Examples of Emissions Forecast Plots
2025/2017
Annual total ratios compare two years of emissions data. Trends in the projections are highlighted.
2017/2011
2025/2011

14. Examples of Emissions Forecast Plots
Percent difference plots show the relative changes between a base and future year.
Highlights the sectors that changed between years, with no indication of the magnitude.
Difference plots show the absolute changes between a base and future year.
Highlights the sectors that have largest impact on the total emissions change.

15. S/L/T Roles in Emissions Forecasts
Document and plan the inventory projection development process in a regional manner for haze analysis
Assure transparency, completeness, documentation
Gather information on S/L/T activities and emissions factors for the inventory projection year or general timeframe
Engage with stakeholders and researchers in the state/region to collect the best-available data, e.g.:
What and how many OOOO and JJJJ sources are in the state or basin, how fast are they turning over, etc.
Representative chemical profiles, spatial allocations, operating times
Develop an inventory database and reporting/tracking tool
Databases are needed to report/track inventories, will feed into forecasting system (probably regional); if the jurisdiction that has the database they can do planning

16. Evaluation Steps for Regional Haze Emissions Forecasting
Inventory active mandatory and voluntary control programs
Determine and implement forecasting methodology and produce data for QA/QC
Collate all, best-available inventory projection data for emissions modeling
Track QA/QC of data through emissions modeling process
Modeling studies to evaluate future year impacts
Submit data to TSS for Regional Haze Planning
Tracking for SIP referencing

17. S/L/T Roles in Emissions Forecasts
Post regional database (TSS) submittal
Review 2017 and newer state- and sector-level EIs as the Regional Haze planning process moves along toward SIP submittals in mid-2021
2017 is next triennial NEI year, significant work by S/L/Ts and EPA in 2018, and well into 2019
Review activity data used by EPA for federal sources, e.g.:
Assumptions about the fleet turnover in MOVES correct
Projection curves and methodologies used by EPA for federal sources
Review federal/state rule implementation roll out and penetration rates