Overview of Appendix W Changes ERM Webinar January 10, 2017 Insert then choose Picture select your picture. Right click your picture and Send to back. Copyright 2015 by ERM Worldwide Limited and/or its affiliates ( ERM ). All Rights Reserved. No part of this work may be reproduced or transmitted in any form or by any means, without prior written permission of ERM.
Webinar Overview Introductions The purpose of air dispersion modeling What is Appendix W and why is it important? Appendix W Updates: Timeline of Appendix W and other related regulations Key elements of the Appendix W revisions Into the weeds Experience and surprises with AERMOD updates Summary; Q&A 2
Why Air Dispersion Modeling? The Clean Air Act (CAA) is the most complex US Environmental Regulation Requires The EPA to continuously update National Ambient Air Quality Standards (NAAQS) Impacts every industrial sector The Air Permit is the Permit to Construct for most large capital projects NGO s focus on using the CAA to block new development and shut down existing plants Air Dispersion Modeling is the cornerstone for demonstrations showing compliance with the CAA Local Scale (50km) AERMOD Major Source Permitting Federal Designation Programs State-Specific initiatives and programs Regional Scale (50km+) CALPUFF, CAMx, CMAQ, SCICHEM Class I Analyses Regional Haze / BART Single Source Ozone and Secondary PM2.5 Impacts 3
What is Appendix W? The Guideline on Air Quality Models, (40 CFR Part 51, Appendix W) establishes approved models and modeling techniques that may be used for regulatory modeling Originally published in April, 1978 Last Updated in November, 2005 Proposed Revisions appeared in Federal Register on July 29 th, 2015 After long delays at OMB, EPA Administrator Gina McCarthy signed the rule on December 20 th, 2016 Effective 30 days after publication in Federal Register, though most agencies are adopting the changes now 1-year Transition Period mostly All related supporting files are located at: https://www3.epa.gov/ttn/scram/appendix_w- 2016.htm 4
Timeline of Appendix W and Related Guidance Jul. 29, 2015 Proposed Revisions to Appendix W appear in Federal Register. Dec. 28, 2015 EPA strengthens the 8-hour ozone NAAQS to 70 ppb. Aug. 1, 2016 EPA Publishes Draft Guidance on SILs for Secondary PM2.5 and Ozone. Dec. 2, 2016 EPA Publishes Guidance on the Development of Modeled Emission Rates for Precursors (MERPs) for Secondary PM2.5 and Ozone. Dec. 20, 2016 EPA Publishes signed Final Appendix W Revisions. Publication in Federal Register Pending. 5
Key Changes in NAAQS Permit Modeling Cumulative modeling against the NAAQS and PSD Increments is the most significant pain point to a modeling demonstration in terms of additional cost, complexity, and time to completion. Appendix W revises two key aspects of cumulative modeling methodology: Use of the Significant Impact Level (SIL): The wording has softened : changed from requirement to recommendation. Approval for use of SIL s is pushed out to state agencies and regional EPA offices. Language makes the use of SIL s easier to challenge. Simplification of NAAQS cumulative modeling domain: Modeling domain is now defined as the farthest distance where modeling predicts a SIL is exceeded, or 50 km, whichever is less. EPA emphasizes that the number of non-project sources to be explicitly included in cumulative modeling should typically be small. Clarification of the process for selecting and excluding ambient background monitors to prevent double counting impacts. 6
Key Appendix W Developments 1. Removal of CALPUFF as preferred model 2. Model Clearinghouse Procedure 3. Revisions to the AERMOD Modeling System New version of AERMET & AERMOD including revised options for NO x -to- NO 2 conversion and fixes dealing with known over-predictions in low-wind events. A new option for generating prognostic meteorological data. Incorporation of the BLP and CALINE3 models to cover buoyant line sources (e.g. aluminum pot lines, roof monitors) and highway emissions. Replacement of SCREEN3 with AERSCREEN. 4. Requirement to account for secondary pollutant (Ozone and PM 2.5 ) formation for single sources. 7
Into the Weeds... 8
Status of CALPUFF Removed as a preferred model for long-range transport (LRT). Can be used as screening approach beyond 50 km. Class I PSD Increment cumulative analysis will need EPA approval on a case-by-case basis. The Federal Land Managers (FLM) have confirmed that CALPUFF will still remain the preferred model for Class I Air Quality Related Value (AQRV) analyses (which address visibility and deposition). 9
Model Clearinghouse Process Applicant submits request to State State reviews and submits to Region Region reviews and submits to MCH 10
Conversion of NO X to NO 2 Now all Tiers are default options Tier 1: 100% conversion of NO X to NO 2 Tier 2: New Ambient Ratio Method (ARM2) Tier 3: Ozone Limiting Method (OLM) or Plume Volume Molar Ratio Method (PVMRM) Requires project specific information such as the in-stack ratio of NO 2 to NO X and ambient ozone data 11
Ambient Ratio Method 2 (ARM2) Revised NO X -NO 2 transformation based on better science than the original ARM and hourly ambient monitor data The ARM2 scaling is based on a curve fit of NO 2 /NO X ratios based on ambient NO X concentrations ARM2, as included in Appendix W, has a minimum default NO 2 /NO X ratio of 0.5 approval required for lower minimum ARM2 Increases the compliance range compared to ARM from 235 to 376 µg/m 3 At NO X concentrations below 149 µg/m 3, ARM2 provides LESS refinement than ARM did. Could cause challenges in SIL demonstrations 12
Low Wind Options AERMOD consistently over-predicts in low-wind conditions. The new default adjusted u-star (ADJU) option not only provides improved model performance but will reduce time and costs in mitigating compliance issues. ADJ_U* is now a default option Alters the calculation of hourly meteorological parameters which govern the dispersion of a plume under predominantly nighttime conditions. LOWWIND3 (LW3) will remain beta (i.e. still requiring EPA Model Clearinghouse justification). Both low-wind options diminish downwash-related impacts near any controlling buildings 13
Default vs. Low Wind Meteorology 14
Meteorological Data for Dispersion Modeling Meteorological data representativeness is key to any regulatory air quality modeling analysis A lack of available representative meteorological data is often seen as a fatal flaw when selecting a site for development Onsite meteorological data collection can be costly, and technically impractical in certain situations One full year of data is required Must be subject to a quality assurance and quality control program, established with an approved meteorological monitoring protocol Time to acquire and configure equipment and construction of site The recent update to Appendix W introduces another option 15
Prognostic Meteorological Data Data from three dimensional atmospheric models are now able to be used in a regulatory setting with AERMOD 8.4.5.1 (a) of Appendix W: For some modeling applications, there may not be a representative NWS or comparable meteorological station available (e.g., complex terrain), and it may be cost prohibitive or infeasible to collect adequately representative site-specific data. For these cases, it may be appropriate to use prognostic meteorological data, if deemed adequately representative, in a regulatory modeling application. The prognostic meteorological model is executed for the area of interest (model domain) to produce wind speeds & directions for different parts of the domain that reflect the effects of local topography and land cover Data from prognostic meteorological models like WRF can be processed through EPA s Mesoscale Meteorological InterFace (MMIF) program to develop data that are adequate for AERMET and subsequently AERMOD 16
Resolution Key to Capturing Influence of Terrain Terrain as seen by WRF-4km Domain with 4km Terrain Resolution Source: WRF/WPS Terrain as seen by WRF-1km Domain with 900m Terrain Resolution Source: WRF/WPS Terrain as seen by WRF-450m Domain with 30m Terrain Resolution. Source: ASTER Global DEM 17
Prognostic Meteorological Data Final Thoughts In the future, EPA and/or States may make standardized prognostic meteorological data available Could offer time savings in establishing a protocol for the application of the prognostic meteorological modeling Likely most applicable to situations where representative data are not available simply due to distance, land cover discrepancies Complex terrain situations would likely require customized, case-specific application of prognostic meteorological models Either way, this approach represents a considerable potential for time and cost savings The availability of representative meteorological data may not be the fatal flaw it once was for project site selection 18
Changes to Ozone and PM 2.5 Assessments Appendix W now places more emphasis on using chemical transport models (CTMs) to quantify secondary impacts on ozone and PM2.5; two tier approach is recommended. Tier 1 Tier 2 Rely on technically credible, existing relationships between precursor pollutants and secondary impacts Use a Chemical Transport Model to quantify secondary impacts New guidance December 2016: Addresses both Tier 1 and Tier 2 approaches for Ozone and PM 2.5. Not a draft; replaces May 2014 PM 2.5 guidance for some aspects; less prescriptive. EPA has proposed new Significant Impact Levels (SILs) Ozone: 1 ppb (8-hr) PM 2.5 : 0.2 µg/m 3 (annual, reduced from 0.3 µg/m 3 ), 1.2 µg/m 3 (24-hr) 19
Ozone and PM 2.5 Assessments: The Tiers 20 Tier 1: Not quite a get out of jail free card. Protocols need to show similarity of chemical environment and source characteristics new source and existing relationships. Uncertainties regarding primary PM2.5 impacts remain. The line between Tier 1 and Tier 2 is somewhat blurred: guidance states that CTMs can be used to..inform development of an appropriate screening technique as part of a Tier 1 demonstration or directly estimate the single-source impacts as part of a Tier 2 demonstration Door appears to be open to reduced form models such as SCICHEM to address secondary impacts. Tier 2: No specific recommended model but CMAQ and CAMx identified as appropriate photochemical models.
Model Emissions Rates for Precursors (MERPS) The MERP represent a level of emissions of precursors that is not expected to contribute significantly to concentrations of ozone or secondarily-formed PM 2.5. In December, 2016, EPA proposed guidance on developing regional specific MERP s for each averaging period of ozone and PM 2.5. MERPs are based on NO X and VOC precursors for ozone and NO X and SO 2 precursors for PM 2.5. Each precursor gets a separate MERP; however, cumulative assessment is needed. Tier-1 analysis vs. a Tier-2 analysis; impacts in excess of the MERP s result in a Tier-2 analysis involving time and cost intensive Photochemical Grid Modeling (PGM). 21
Range of MERP Values in EPA guidance 24-hr PM 2.5 NO X : 1,075 tpy to 100,000 tpy SO 2 : 210 tpy to 27,000 tpy Annual PM 2.5 NO X : 3,184 tpy to 779,000 tpy SO 2 : 1,795 tpy to 75,500 tpy 8-hr Ozone NO X : 107 tpy to 5,573 tpy VOC: 814 tpy to 145,000 tpy These MERP values are based off of the full range of hypothetical sources modeled by EPA, at various locations across the US The values are heavily dependent on where the source is located local atmospheric chemistry Lower range MERP values for SO 2 for 24-hr PM2.5 and NO X for ozone point toward need for Tier 2 analyses for large projects Comments on the MERPs are due by February 2, 2017 22
Insights: AERMOD modeling system v16126 Comparison of predicted concentrations using the default ADJ_U* low wind option in v16216 versus the beta ADJ_U* in v15181. AERMOD v16216 erroneous calculations associated with circular area sources (e.g. exposed tanks). Preparation of prognostic meteorological data using AERMET v16216. LOWWIND3 comparisons of v15181, v16216 and field data. 23
Summary of Appendix W Improvements Improvements Improvements to model formulation (v.16216) to better handle low wind situations and the conversion of NO X to NO 2. An option for generating prognostic meteorological data with WRF/MMIF. Incorporation into AERMOD of the BLP and CALINE3 models to cover buoyant line sources (e.g. aluminum pot lines, roof monitors) and highway emissions. Replacement of the obsolete SCREEN3 model with AERSCREEN. Simplification and clarification of the methodology for developing cumulative modeling inventories. Not Appendix W but EPA introduces MERP s as a possible Tier-1 screening tool for ozone and secondary PM 2.5 analysis. 24
Summary of Appendix W Challenges Challenges Requirement for accounting for secondary pollutants (Ozone and PM 2.5 ) formation for single sources could potentially add time, effort, and cost if photochemical modeling is required. With CALPUFF removed as a preferred model, Class I analyses are now handled on a case by case basis, potentially adding time and uncertainty to the project and exposure to challenges by 3 rd parties. Weakening of SIL language exposes their use to challenges. Unexpected and unannounced change to ADJ_U* handling, reducing its effect in many cases, has left numerous projects scrambling for new answers. Model Clearinghouse Procedure 25
Contact Information Today s Speakers: Mark Garrison Philadelphia, PA +1 484 9130 369 mark.garrison@erm.com Tom Wickstrom Philadelphia, PA +1 484 913 0453 tom.wickstrom@erm.com Richard Hamel Boston, MA +1 617 646 7826 richard.hamel@erm.com Carlos Szembek New Orleans, LA +1 504 846 9236 carlos.szembek@erm.com Beth Barfield Raleigh, NC +1 919 233 4501 beth.barfield@erm.com Toby Hanna Ewing, NJ +1 609 403 7518 toby.hanna@erm.com Daniel Guido Indianapolis, IN +1 317 249 4737 daniel.guido@erm.com 26