Industrial Point Projection Task Force
Proposed approach for calculating 2022 emissions for the non-EGU point sector.
- Use all data submitted for 2022 by SLTs
- Gapfill with data submitted for 2021 by SLTs, minus closures
- Gapfill with data submitted for 2020 NEI by STLs, minus closures
- Do not apply projection or control information to the 2020 or 2021 data, use as is
- Apply closures to base year data
- Add projection information from EIS
- Add federal control programs
- Add state/local control programs
- See section 184.108.40.206 in the 2016v3 TSD for non-EGU point projection methods
- See section 4.2.4 for control methods
- For the 2016v3 platform, 2023 non-EGU point was set to 2019 EIS and state submitted point source data minus plant closures that were known to occur between 2019 and 2023
- 2026 emissions were then projected from the 2023 emissions
- Start with a base year (e.g., 2020 or 2021)
- What will this be for the 2022 EMP?
- Remove known closures from the inventory (based on EIS and state-submitted data)
- Q: Where can we find the most recent CLOSURE packet?
- Apply projection factors from AEO; MARAMA provides their own projections
- For the 2016v3 inventory projections are based on 2022 regional industrial source energy consumption (e.g., AEO, 2022)
- See Table 4-14 in the 2016v3 TSD for the list of AEO tables of projection information
- Projection factors (2026 energy/2023 energy) capped at 1.3; no floor for contraction of emissions
- Q: What is the mapping between the AEO energy use tables and SCCs or NAICs?
- Apply controls for national rules: RICE NSPS, natural gas turbine NSPS, process heaters NSPS, boiler MACT, etc.
- Check in on the 2022 non-EGU point inventory approach and messaging about this approach to the wider community
- Brainstorm ideas for projection methods/data review
- Identify items that we want to report out from this task force on the February 7 national report out call
Industrial Point Inventory Approach for 2022:
- Judy drafted a 2022 inventory methodology for this sector based on our work group conversation at the Dec. 18th call, and EPA provided comments. This was sent to MJO leads to share with their regions. Zac noted he sent it to the inventory contacts in the LADCO region. Mark Jannsen is helping those agencies prioritize review. Of note, it's up to the agencies to ask EPA for a report of permanent and temporary shutdowns, should they need it for their review. We can pass this message during our workgroup update on the National Report Out.
- EPA is looking for feedback on the document by Feb. 1st.
- At the last meeting, Alison showed the facility comparison sheet that identified sources where the facility emissions were the same in 2020 and 2021 (to help identify unsubmitted sources that we may see from 2021 to 2022). Alison will add 2022 once we have the draft 2022 inventory out of EIS. Then, we'll have a picture of what years data are being pulled from.
- Question about CEMS - how are CEMS files pulled into the inventory?
- Alison - We only use the CEMS data for EGUs in the modeling platform. For non-EGU sources with CEMS, we rely on the fact that facilities likely report CEMS data to the agencies, and then we receive this as part of the EIS reporting. We don't have a process to pull in CEMS data for non-EGU, industrial point sources like we have for EGUs.
- Zac - do you have a sense of the extent of CEMS in industrial? Are we missing out on data here?
- Alison - there are some matching difficulties in matching CEMS data to NEI, so perhaps.
- Susan - we have those units in the ERTAC tool, so we can pull those out and provide them
- To confirm understanding with the group - EPA will take the 2022 inventory as it is, what's been submitted to EIS, once it's reviewed, and not adjust anything, except railyards.
- Andy - is aircraft part of the data category?
- Alison - yes, it is in the point data category. We use the Aviation Environmental Design Tool (AEDT) to estimate emissions for the top 50ish airports and apply adjustment factors for the other airports. We'll be reviewing this with the MAR group. This time, we plan to load the adjusted emissions into EIS once they're reviewed. Once we get to the end of 2022v1, we'll load these.
- Zac - airport and rails need to get from 2020 to 2022 - so it is good to know that EPA is working on this, and we'll get a chance to look at the data. For LADCO, we're comfortable with this.
- Alison - once we have the flat file in February, we can start looking at this immediately.
Ideas for Projection Methods:
- EPA will start the work on the analytic years in April during the data review period.
- Going back to spreadsheets Zac shared earlier: 2016v3 projections from 2023-2026, using AEO factors.
- Alison - We took a conservative approach; we set 2023 =2019, then we projected to 2026.
- Zac - that jump from 2023 to 2026 gives us an idea of what EPA will do this time. Just activity factors, we can see what AEO data sources and tables EPA used to go from 2023 to 2026. This is the logical starting point for us to review the methodology.
- Alison - Normally, we would use AEO, we haven't talked about it this time. Earlier in my career, we held them constant. If the decision is made to grow the sources, then we'd grow using AEO (from last year, there is no new AEO this year).
- Zac - MARAMA had its own methodology. NJ's are slightly different. MARAMA used employment and AEO. NJ used a mix of employment and AEO but only used AEO for storage terminals.
- Judy - For non-MARAMA states, EPA used various versions of AEO to develop growth factors to project point sources from 2016 actuals to 2023.
- Zac - we should review the different approaches - MARAMA, NJ, EPA. When, where, why we want to flatline sources. We also need to talk about the offsets.
- Andy - Perhaps include an analysis of what the historical trends have been?
- Susan - look at controls, too.
- Zac - Good point, Susan - we will have a good idea for OTB, but the OTW are trickier given where courts are on things
- Judy - I apply a limited growth to the industrial source, so the offset bank is built into the factor. It's not completely flatlined, but it's not overestimated either. If I get comments on the SIP, I can say I am including growth. I don't want to overestimate the point source because the permit sets the defined limit.
- Alison - add a small note - when I was researching the GN rule comments on our approaches, we did look at some permit data, and we found that our projections were within the bounds of the permit (sources aren't always operating at their permit limits).
- Judy - let's look back. What are the actual emissions from the last years and see how that measures up to what we grew.
- Zac - what about getting accurate information on controls on non-EGUs? In this region, we were looking at a NOx RACT modeling run for one of our states. We saw differences in what was actually on the sources vs what was in the inventory.
- Andy - I assume this information differs from state to state based on how good their control reporting to inventories are.
- Judy - we’re talking about two different ideas of controls here. First, we have the control as is in the base inventory, presumably reported via the agency, then, in a projection inventory, we may want to apply a control factor based on a new rule or based on an assumption on what the future will be.
- Andy - Yes, relative future control vs current control. General procedures is to take off the current control and add on the new control. The base control is the first step, getting that accurate. It's going to be state-by-state. It's such an important component of the projection that people don't focus on.
- Zac - is there anything we can do to communicate this need? EPA already asks for this, but can we do anything else to communicate to regions and states?
- Andy - identify priority source categories. Based on total emissions in the base year, could be based on future controls that want to be applied incrementally. Where is it the most important?
- Zac - yes, what can we target - "control potential"
- Andy - EPA will be looking at federal controls, knowledge of those source categories would be the priority.
- Zac - I like what you did on the NP task force - break the sector up into subsectors. I tried to take a similar approach (US_state_SCC-ptnonipm_2016v3_26jan2023.xlsx, Column F - projection factor associated with Sector Fuel categories). One way we could categorize the inventory is by using the AEO projection category. Another way is to take subcategories of SCCs because of the different tiers within each SCC.
- Judy - you may overestimate growth with AEO to the SCC code. I applied to the NAIC. Suggest we look at the projections vs actuals; e.g., we projected 2023 from 2016, but now we have actually 2023. Let's look at the projected 2023 compared to our actual 2023.
- Zac we don't have an actual 2023 yet. We have a 2022 actual
- Judy - compare 2022 actual to 2016v1 projection?
- Zac - trends look at 2016 and 2017, 2020, and 2022. Gather all the actuals we can get. Then, look at the 2016 projections to see if the method was valid. Base this on empirical information. I hear your point Judy on AEO categories, based on combustion info some of the non-combustion factors get dropped, Do we break down by SCC instead of AEO then?
- Andy - for NC we do a combo of NAICS and SCC-based application of growth factors.
- ACTION ITEM - try to summarize this sector by SCC code with combustion and non-combustion sources? Then, be able to sort by different sources by pollutant
- Andy - yes - that would be helpful. At whatever geographic level you can do it at will be helpful.
- Judy - It would be helpful to look at 2016 v1 and double-check its AEO to get to 2023. If that's what they did, then 2016v1 vs 2022 actual emissions. That would give us an indication of whether we estimated or overestimated.
- Zac - I like this. We just don't know what is going to happen in the future. We need a 2038 projection for haze! Judy's suggestion makes sense for the near term 2026, maybe even 2032. But for a longer projection?
- Susan - MARAMA factors were submitted to 2016.
- ACTION ITEM - try to compare 2016v1 and 2023 projections to 'actuals' from in-between years. Create a line with points along the way that show how actuals vs 2023 projections appear in relation to each other.
Identify items from this task force to report on during the Feb. 7 National Report Out call:
- Quickly - thoughts on what we should report out on national call? The obvious is the 2022 document. We're starting to review the projections for analytic years.
- Susan - summarize any inputs EPA has received to date, if possible. March 15th is the deadline, as outlined in the point projection approach document, to submit shutdowns. And, if states need reports on what is marked permanently or temporarily shutdown, ask EPA. And a heads up on controls - we need good control information for analytic year work, and if EPA can share info on national rules, that would be helpful.
- Review the analysis tools/products on the taskforce wiki (see below).
- Discuss current default approach with U.S. EPA to calculate 2022 emissions from 2020, 2021, and 2022 data
- Identify alternative approaches for 2022 and how to recommend those approaches to U.S. EPA
- Spreadsheet of 2016v3 nonEGU Point SCCs (US_State_SCC_ptnonipm_2016v3_26jan2023.xlsx): This workbook is intended to help identify subsector projection approaches and priorities based on 2016v3 (2019) data. See the "README" tab in the workbook for further explanation. Note that AEO is used for analytic year projections, not how to get from 2020 to 2022. However, we can use these subcategories to prioritize things and identify COVID impacts. The workbook contains three sheets:
- Projection Priorities: based on AEO Projection Category (sector-fuel) and how EPA applies projection factors to the subsectors of the inventory. EPA mapped AEO categories (economic indicators) to groups of SCCs to project activity.
- 2016v3 (2019) Emissions: these are the best emissions we have projected. Organized by region.
- Projections by Region - SCC: Can organize to see the projection numbers (2023 to 2026) associated with an AEO category.
- 2020-2021 point source shutdown report
- Pulled out of EIS. By region, state, FIPs code, EIS ID, all the way to the facility, what planned, perm, and temporary shutdowns there are, and when those shutdowns will happen.
- Annual total emissions of 2016-2021 point inventories by state-SCC (2016-2021 (pre-2021b) point state-SCC comparison 29nov2023.xlsx)
- Used to help with COVID impacts
- Zac created a ratio from one year to the next to help identify a signal. We need to think about if there are any COVID impacts.
*Question - this gives you the economic impacts to apply to future projections, but in NSR, we need to account for offsets. Can we consider this, or assume they will find the credits when they need to build?
- States are tracking this issue, but EPA does not account for it (each state is different). States can add emissions to the inventory or account for offsets in the growth factors. Legally, states must account for them in the modeling inventory for analytic years. However, this is an SIP requirement and not something done in previous collaborations - 2016 did not account for offsets.
- NOTE - we need to discuss this when we start to work on the analytic years. Possible suggestions or approaches for how we can consider this? It's not been done systematically on other platforms.
Discuss current default approach with U.S. EPA to calculate 2022 emissions from 2020, 2021, and 2022 data
- Alison created the Facility summary spreadsheet (Annual total emissions of 2016-2021 point inventories by state-SCC). This spreadsheet identifies facility emissions that were the same in 2020 and 2021. This is to help with unsubmitted sources to get from 2021 to 2022 (we are using 2021 as a surrogate for 2022).
- The spreadsheet contains Facility, sector, and key pollutants for 2019, 2020, 2021, and 2022. Added a pivot table - can organize it. This will help understand what emissions were already pulled forward.
- Shantha - If a state submits a full inventory, don't do a pull forward??
- Alison - Ron Ryan will help with this, yes…
- Judy - it probably will make sense to wait until the 2022 report comes out before we contact Ron. Is there a way we can add 2022 to this spreadsheet once we get it?
- Alison - yes, cy2022, for any source where it is not 2022, you can find it. This is a post-processing product.
- Judy - okay, going to EGU convo - I remember that CEMS takes priority over state submittals. This is an issue if the CEMS goes down, so I remember having issues I had to fix. Talk about this on the EGU call.
- Zac - this report fills a hole that states need to review.
Identify alternative approaches for 2022 and how to recommend those approaches to U.S. EPA
What do states need to do to modify the default inventory approach?
- Use all data submitted for 2022 by SLTs
- Gapfill data submitted for 2021 by SLTs, minus closures
- Gapfill with data submitted for 2020 NEI by SLTs, minus closures
- Do not apply projection or control information to the 2020 or 2021 data, use as is.
- Zac - What about data submitted for 2022 that doesn't have a source? States don't want to pull forward from 2021.
- Alison - Sources will be listed as closed (PS).
- Zac - instead of shutdown,perhaps we can use a different term/code - temporarily suspended.
- Susan - is this spreadsheet showing the facility closed down or the unit shutdown and reflected that at the facility?
- Alison - Good question - Janice - see if we can get one at the unit.
What's the process for states that want to propose an alternative to the loop above?
- Alison - Ron can help facilitate the process...let Ron know that your submission is complete, and then he can help facilitate sources not to pull forward. Ron will identify the sources that shouldn't be pulled forward and temporarily shut them down. 2022v1 EIS should match the modeling platform once we finish the whole process!!!!!
- Zac - document the process. We need to put a note or something we say this is carrying forward. Even if a state submits 2022 data, they still have to work with the EPA to not pull emissions from 2021 if they don't want them pulled forward to 2022. They may not be essentially closed down, but they shouldn't be shutdown either.
- Judy - I can craft a few sub-bullets that describe this process.
- Zac - we should send it to the MJOs, the NEI newsletter, and the collaborative steering committee. Get this out to the inventory folks for them to understand. What about timing, Alison? It would be most helpful to catch this in the latter half of January.
- Alison and Zac can get with Ron to see if the process we're going to take works. Will wait until get Judy's summary.
- Mark, there is one other category of sites that don't meet the reporting requirements for that year. Are there several smaller ones that don't meet the PTE and fall out?
- Alison, we need to understand how this works.
ACTION ITEM - Judy sends a sub-bullet list describing the process to Zac to review. Zac takes to Alison. Alison and Zac take it to Ron for his review. Then, broadcast this message!
MEET AGAIN - later in January? 15th or 22nd? Let's do 22nd.
- 2022 inventory
- Discuss methods to calculate 2022 emissions for this sector
- Projection methods and data
- Review US EPA methods for projection of non-EGU point sources
- Orient to where to access data used in projecting this sector
- Next call in December
- Compare MARAMA and US EPA methods
- Review data for the 2022 EMP: closures, projections (which version of AEO?), controls
- EPA's plans for the 2022 inventory
- EIS automatically generates 2022 inventories based on coding and data available in the system
- Ptnonipm sector includes point oil and gas, airports, and railyards, and there is a process to separate these categories out from the industrial, non-EGU point sources because they use different projection methods or actual year data
- EPA wouldn't apply control projection info for 2020->2022 because states submit the controlled emissions information as part of the EIS submission.
- Questions about what EPA does when a source has data in one year but not the next
- If EPA did not receive a closure request through EIS they will use data for the sources from the most recent available triennial inventory; presume the data is still operating unless they get closure information from the state
- Unless marked closed, then EPA presumes the source is still operating. For example, if a source is operating in 2020 but not in 2022, carry it into 2022 unless the agency confirms that it is closed in 2022. This means that this information is carried over between NEI years. When it's an NEI year, it's fresh - e.g., in 2020; Alison doesn't believe closures were carried forward from 2019. It will likely be the same between 2022 and 2023 (the next NEI year). For a triennial year, no emissions are carried forward.
- EPA starts from scratch with each triennial inventory; do not look back for sources in previous years
- State requests
- This task force will have a challenge with the data that is different from each year (certain states submit every year, some do not, etc.)
- Farren (WA) asked if there is an option to use past inventory years to estimate 2022 rather than 2020 to address concerns about COVID; Can they use 2017 instead of 2020 to estimate 2022 emissions?
- Alison noted that this isn't the way that EIS is set up, and acknowledged that they see a lot of changes between 2016v3 and 2020 in this sector
- Andy (NC) will submit a comprehensive 2022. If that's the case, can we tell EPA not to use 2020 in any way? Judy (NJ) wants the same thing - do not bring forward 2020 at all.
- ACTION ITEM - Alison will double-check how this process will work for states that do not want emissions carried forward - the 'bring forward' processes are all coded in the EIS software, so we have to make sure states' 2022 submissions aren't overwritten by previous year
- Can we see a closure list for recent years?
- It's in EIS and not part of a typical report; EPA will check to see if a report can be generated
- Question - What about temporary vs permanent shutdowns? Temporary would not be submitted that year but may be submitted the next. There must be a different code for temporary. Only the permanent shutdowns would be kept in the modeling platform, correct? Will the temp code get into the flat file?
- ACTION ITEM - Alison will check if the temp closures are included in the flat file - however, if they are not, likely won't be in the analytic years.
- Are shutdowns marked in EIS only for permanent or temporary, or both?
- Alison thinks they're for permanent shutdowns only
- Some sources could be below the reporting threshold, still operating in but not reporting to EIS; the algorithms will still carry forward the most recent reported emissions for these types of sources unless they are told otherwise by the state/source
- Alison suggests looking at the percentage of sources carried forward in 2021 relative to 2020 for an indication of the sources that likely won't report in 2022
- Someone noted that 2022 is special because of the explicit call for more information this year (e.g., Type B sources)
- How to address COVID impacts in 2020
- Look at how much each SCC changed from past (2016, 2017, 2019, 2020, 2021) inventories to see if there is a big difference in 2020; see this spreadsheet with point (EGU, non-EGU, oil & gas) emissions for 2016 through 2021. In the SCC total tab columns I through M you'll see the differences from year to year; column J highlights the 2019 to 2020 difference
- Link to folder with spreadsheets comparing unit and facility level emissions for recent base years to 2020
- Qualitatively consider the impacts on different sub-sectors of the nonEGU point sector (e.g., glass or ICI boilers)
- Suggestion to identify certain ptnonipm sectors that were impacted by the pandemic, similar to what the nonpoint projections task force has done.
- Good idea - look at 2019 vs 2020, and see if we can see differences between SCCs.
- Judy noted the situation in NJ with offset banks and the need always to grow the data into the future; also noted permit conditions limit that point source growth; she could consider keeping the data flat from 2020 to 2022, and even into the future analytics years for parts of this sector
- Byeong notes that we should communicate to the larger group of states during report-out calls the issue of reporting closures to ensure that sources won't be carried forward; it's not enough to just not report emissions for a source
- Alison shared a slide presented at the last quarterly report-out about the need to report closures; we will emphasize this on the next call again
- Discussion wrap up
- Industrial point projections co-chairs (Zac and Rhonda) will work together to create an analysis between 2019 and 2020 SCC changes, and see if there are outliers for 2020. Will bring back to group for discussion.
- Need a list of closures, from the 2020 - 2021 - 2022 years pulled from EIS. We will circulate for QA. Janice and/or Alison will pull the data/create the report.
- Doodle poll will be sent for a few dates in December/January - we'll cover the items we didn't get to today.
- Next call
- We will talk in December but not before the projections WG call; Rhonda to send out a Doodle Poll
- Review 2020 vs other recent years to see if we can identify COVID impacts
- Review list of closures
- Talk about plans for analytic years
- Alison: double-check how it will work for states that do not want emissions carried forward from 2020 to 2022
- Alison: get a list of 2020, 2021 shutdowns from EIS
- See the 2020-2021 point source shutdown report; note that PS = permanent shutdown and TS = temporary shutdown
- Alison: get a list of recent base year inventories to compare for identifying COVID impacts. Zac and Rhonda will use list to create an analysis between 2019 and 2020 SCC changes, and see if there are outliers for 2020.
- Alison will check if the temp closures are included in the flat file
- Rhonda: send out a Doodle poll to talk before the winter break
- Done, scheduled for 12/18 @ 1:00 Eastern
- Read the US EPA 2016v3 TSD sections on the non-EGU point sector (note that nonegu point and ptnonipm are used interchangeably)
- Section 4.2/4.2.1: general background on projection approach and CoST
- Section 4.2.2: closure methods (SMOKE CLOSURE packet)
- Section 4.2.3 intro: summary of projection methods (SMOKE PROJECTION packet)
- Section 220.127.116.11: non-EGU point methods/details
- Section 4.2.4: summary and descriptions of controls applied to non-EGU point (SMOKE CONTROLS packet)