Roger W. Brode

1
AERMOD Update

• Roger W. Brode
• U.S. EPA/OAQPS/AQAD
• Air Quality Modeling Group

EPA Region 10 and State Meeting Seattle,
Washington October 22, 2007
2
Outline

• AERMOD Modeling System Update
• AERSCREEN Update
• AERSURFACE Update
• Questions

3
AERMOD Modeling System Status

• AERMOD promulgation as preferred model in full
  effect as of December 9, 2006 end of 1-year
  grandfather period
• Significant updates to all three components
  released January 2007 (dated 06341)
• Limited update to AERMOD released January 26,
  2007 (dated 07026)

4
AERMOD Model ChangesRecent Enhancements

• Updated processing for PM-2.5 NAAQS and removal
  of post-1997 processing for PM-10
• Additional options to vary emissions by month,
  hour-of-day and day-of-week (MHRDOW, MHRDOW7)
• Multiple urban areas in single model run
• User-specified default in-stack NO2/NOx ratio for
  Plume Volume Molar Ratio Method (PVMRM)

5
AERMOD Model Changes Recent Enhancements (cont.)

• New BETA option added to CO MODELOPT card to
  identify and allow use of new features added to
  the model that are still in a draft BETA-test
  status
• Non-DFAULT option
• Allows for more efficient evolution of code by
  reducing need for maintaining multiple versions
• Beta options for capped and horizontal releases
• EPA Model Clearinghouse procedure not applicable
  for PRIME downwash algorithm

6
AERMOD Model ChangesRecent Enhancements (cont.)

• Beta option to account for NO2/NOx plume
  chemistry of combined plumes in the computation
  of PSD increment consumption with credits for
  PVMRM
• Maximum number of vertices for AREAPOLY source
  allocated dynamically
• More memory-efficient process for allocating
  array storage

7
AERMOD Model ChangesRecent Bug Fixes

• Several minor fixes for PRIME (mostly undefined
  variables)
• Correct variable type error for AREACIRC sources
  (minor impact), and correct potential error for
  all AREA source types (impacts may vary)
• Correct error for PVMRM when multi-level on-site
  wind data are used
• A few additional minor bugs

8
AERMOD Model ChangesMiscellaneous Items

• Added range check on VPTGZI in AERMOD to avoid
  problems with data provided from sources other
  than AERMET (e.g. gridded met models)
• Tightened range checks on optional urban
  roughness length on URBANOPT card
• Default of 1m should be used in most cases
  values greater than 5.0m not allowed warnings
  issued for values less than 0.8m or greater than
  1.5m
• More details presented later
• Miscellaneous code clean-up items

9
AERMOD Model ChangesUrban Roughness Length Issue

• CO URBANOPT card includes option for
  user-specified urban roughness length default
  value of 1.0m used if not specified
• Urban Z0 used to adjust rural u based on
  pseudo-convective urban w to urbanize sigma-z
  for surface sources
• Adjustment based on matching convective sigma-w
  profile with mechanical sigma-w profile at height
  of 7Z0-urban

10
AERMOD Model ChangesUrban Roughness Length Issue

• Considered removing option based on concern that
  unrealistic values were being used, and possible
  misinterpretation of how value is used in model
• It does not adjust for difference in roughness
  between met site and urban application site
• Performed sensitivity analysis to determine
  potential impact of removing option
• Results showed sensitivity limited to low-level
  sources, but greater sensitivity than hoped for
• Issue is being addressed in revised AIG, and next
  AERMOD update will treat non-default values as
  non-DFAULT

11
AERMOD Model ChangesCapped/Horizontal Releases

• Current Model Clearinghouse procedure not
  applicable to PRIME
• Clearinghouse procedure sets Vs low (0.001m/s)
  and adjusts Ds to maintain flow rate and buoyancy
• PRIME numerical plume rise uses input Ds to
  define radius of plume use of effective radius
  may alter results in physically unrealistic ways
• AERMOD Implementation Guide suggests using
  Vs0.001m/s with actual Ds as interim solution

12
AERMOD Model ChangesCapped/Horizontal Releases

• Draft/BETA options have been implemented for
  capped horizontal
• User inputs actual stack Vs and Ds
• Source types POINTCAP POINTHOR used to trigger
  BETA options
• Non-downwash sources use Clearinghouse procedure
• PRIME sources adjust plume radius (currently 2X)
  to account for initial spread from cap
• Partition vertical and horizontal momentum

13
(No Transcript)
14
AERMOD Model ChangesCapped/Horizontal Releases
THE SUMMARY OF HIGHEST 1-HR RESULTS

DATE GROUP ID
AVERAGE CONC (YYMMDDHH) - - - - - - - - - -
- - - - - - - - - - - - - - - - - - - - STACK1
HIGH 1ST HIGH VALUE IS 2.97269 ON
90010102 STACK1C HIGH 1ST HIGH VALUE IS
2.98035 ON 90010102 STACK1H HIGH 1ST HIGH
VALUE IS 72.16824 ON 90010102 STACK1C0 HIGH
1ST HIGH VALUE IS 19767.68555 ON
90010102 STACK1CE HIGH 1ST HIGH VALUE IS
2.97411 ON 90010102
15
AERMOD Model ChangesPVMRM PSD Credit Option

• Define
• A increment-consuming sources
• B existing (non-retired) baseline sources
• C increment-expanding (retired baseline)
  sources
• Increment consumption without credits
• (AB) B
• PSD Credit (BC) B
• CHITOTAL (AB) B (BC) B
• (AB) (BC)

16
AERMOD Model ChangesUnresolved Issues/Planned
Updates

• Updated AERMOD Modeling System package planned
  ASAP to address
• Issues with portability of code to Linux for all
  three components
• Additional improvements to AERMOD memory
  allocation
• Problems with AERMAP for Alaska DEM data
• Cross UTM zone and edge receptor bugs in AERMAP
• Format problem with some ISHD data may cause
  AERMET to crash
• FIXISHD utility program interim fix released in
  April

17
AERMOD Model ChangesUnresolved Issues/Planned
Updates

• Inconsistencies between NWS surface data formats
• TD-3280 sky cover codes
• Abbreviated ISHD (TD-3505) sky cover
• Allocatable arrays for AERMAP and modifications
  to support National Elevation Dataset (NED)
• PRIME/BPIPPRM issues
• Split building/elongated building issues
• Discontinuity at GEP stack height
• Upwind impact of building on plume
• Possible bug for out-side cavity source

18
AERMOD Model ChangesUnresolved Issues/Planned
Updates

• Upgrade Fortran compiler
• Compaq Visual Fortran no longer supported
• Recent quirk with Compaq Visual Fortran
• Using open-source G95 compiler to debug codes
• Upgraded to Intel compiler for Windows limited
  tests show about 40 improvement in runtime
• Use of compiler option for double precision
• Distribute Windows and Linux executables?
• Updating users guides convert to MS Word and
  merge with Addenda

19
AERMOD Update Questions
20
AERSCREEN Finalization Workgroup

• Jim Haywood, Chair, Michigan DEQ
• Karen Wesson, EPA
• Roger Brode, EPA (formerly with MACTEC)
• James Thurman, EPA
• Bob Paine, ENSR
• Lloyd Schulman, TRC
• Acknowledge Herman Wong, EPA Region 10

21
What is AERSCREEN?

• AERSCREEN refers to applying the AERMOD model in
  a screening mode
• SCREEN option added to AERMOD in 1995 forces
  model to calculate centerline concentration for
  each source/receptor/meteorology combination
• SCREEN option limits output to 1-hour averages
  and selects NOCHKD option to eliminate date
  sequence checking
• AERSCREEN interface developed by Jim Haywood
• AERSCREEN program provides interface to run
  AERMOD in SCREEN mode, incorporates MAKEMET,
  BPIPPRM and AERMAP
• MAKEMET program generates matrix of
  meteorological conditions based on user-specified
  surface characteristics, formatted for input to
  AERMOD (.sfc and .pfl files)

22
Description of AERSCREEN

• MAKEMET includes loops through meteorological
  parameters
• Wind speed (stable and convective)
• Cloud cover (stable and convective
• Max/min ambient temp (stable and convective)
• Solar elevation angle (stable and convective)
• Convective velocity scale (w) (convective only)
• Mechanical mixing heights (stable only)
• Uses AERMET subroutines to calculate u and L,
  also calculates convective mixing heights
• Generates AERMOD-ready surface and profile files
  with site-specific screening meteorology

23
Description of AERSCREEN

• AERSCREEN command-prompt program developed by Jim
  Haywood, Michigan DEQ
• Interactive data entry (command prompts/DOS
  Screen)
• Single point, volume, area or flare source
• Flat or complex terrain (user-specified list of
  7.5-minute or 1-degree DEM hopefully also
  15-minute files soon)
• PRIME building downwash (specify stack location
  for single tier or provide BPIPPRM input file)
• MAKEMET meteorology with site-specific surface
  characteristics
• Search routine to locate worst-case impact
  location
• Re-Use of previous AERSCREEN run files
• Includes factors for 3-hour, 8-hour, 24-hour and
  annual averages similar to SCREEN3 factors (not
  finalized yet)

24
AERSCREEN Status

• Draft AERSCREEN package submitted to EPA regions
  on 4/25/06 for internal (alpha) review and
  testing
• Comments received from several regions and
  states many comments related to problems with
  AERMAP
• Public release of draft AERSCREEN package as soon
  as possible

25
AERSCREEN Status

• ASCREEN program updated to use new AERMAP and
  AERMOD executables AREA, AREACIRC and flare
  source options added
• AERMET NR_ANG bug fix incorporated
• Surface characteristic options being modified to
  provide some linkage with AERSURFACE
• Previous testing to be updated with latest
  version additional testing of AERSCREEN vs.
  on-site met data also planned
• Additional documentation, including draft users
  guide and test results, needed before public
  release

26
AERSCREEN Tests

• Significant testing to date shows good results
  across wide range of applications
• Good defined as reasonable conservatism
  compared to AERMOD refined estimates
• Testing performed for rugged terrain applications
  with downwash minor modifications to MAKEMET

27
(No Transcript)
28
(No Transcript)
29
(No Transcript)
30
Averaging Time Factors (still under development)

• 3-hour fixed ratio of 0.95 (SCREEN3 0.90
  0.10)
• 8-hour fixed ratio of 0.70 (SCREEN3 0.70
  0.20)
• 24-hour fixed ratio of 0.40 for plume heights
  up to 100m, then linearly interpolated to a ratio
  of 0.30 for plume heights above 200m (SCREEN3
  0.40 0.20)
• Annual fixed ratio of 0.10 for plume heights up
  to 100m, then linearly interpolated to a ratio of
  0.04 for plume heights above 200m (SCREEN3
  0.08 0.02).

31
(No Transcript)
32
AERSCREENTopics for Discussion

• Results from testing
• Averaging factors Results? Fixed vs varying?
• Multiple source screening?
• MAKEMET and site-specific minimum wind speed
  and anemometer height? Defaults?

33
AERSCREEN Questions
34
What is AERSURFACE?

• AERSURFACE is a tool that produces surface
  characteristics data (albedo, Bowen ratio
  surface roughness) for use in AERMET and/or
  AERSCREEN
• Original program developed by Chris Arrington (WV
  DEP) and Larry Simmons (Energy Environmental
  Management, Inc.)
• New version being developed by EPA with
  contractor support (MACTEC)

35
AERSURFACE Description

• EPA AERSURFACE program currently uses USGS NLCD
  92 data
• 21-category land cover classification scheme
• 30 m spatial resolution
• Currently draft supports both State and
  seamless files
• AERSURFACE includes updated tables of seasonal
  variations for surface characteristics data by
  land cover category

36
NLCD Land Cover Categories
37
AERSURFACE Surface Characteristics
38
AERSURFACE Status

• Draft AERSURFACE design document submitted to
  ROs for review July 2006
• Comments received from Regions and States,
  especially regarding draft look-up tables
• Draft version of AERSURFACE being reviewed by
  AIWG
• Recent changes to method for estimating surface
  characteristics from land cover data

39
AERSURFACE Design

• Revisions to calculating area-weighted averages
  for surface characteristics
• Inverse-distance weighting to account for
  increased width/area of sector with distance
• Revisions to calculation methods
• Averaging ln(zo) for surface roughness
• Geometric mean for Bowen ratio
• Revisions to default domain/fetch distances
• 1km fetch for roughness
• 10x10km domain for Bowen ratio and albedo no
  sector dependency

40
AERSURFACE Design

• Default values/functions with several user
  choices
• Number of sectors (up to 12)
• Output monthly, seasonal, or annual data
• Wet/dry/normal conditions for Bowen ratio
• Snow vs. no snow cover
• Arid vs. Non-arid
• Airport vs. Non-airport location

41
USGS NLCD 92 Example data
42
USGS NLCD 92 Example data
43
AERMET Stage 3 Formatted Datafrom AERSURFACE

• Generated by Internal Draft version of
  AERSURFACE, dated 07184
• Center Latitude (decimal degrees)
  35.891940
• Center Longitude (decimal degrees)
  -78.781940
• Study radius (km) for surface roughness 1.0
• FREQ_SECT SEASONAL 12
• SECTOR 1 0 30
• SECTOR 2 30 60
• SECTOR 3 60 90
• .
• .
• Season Sect Alb Bo
  Zo
• SITE_CHAR 1 1 0.15 0.37
  0.089
• SITE_CHAR 2 1 0.14 0.30
  0.118
• SITE_CHAR 3 1 0.15 0.24
  0.143
• SITE_CHAR 4 1 0.15 0.37
  0.123
• SITE_CHAR 1 2 0.15 0.37
  0.011
• SITE_CHAR 2 2 0.14 0.30
  0.016
• SITE_CHAR 3 2 0.15 0.24
  0.022

44
Potential AERSURFACE Issue

• ASOS station location uncertainties
• Excel file with ASOS station locations available
  on NCDC website is unreliable for location
  information
• Additional data available for about 200 ASOS
  stations as part of tropical cyclone wind study
  appears to be (generally) reliable
• Many station locations appear to be off by
  several hundred meters (median value of about
  500m)
• Use of erroneous station locations in AERSURFACE
  could invalidate results

45
ASOS Met Station Locations Cyclone Wind Study
46
ASOS Met Station Locations
47
AERSURFACE Questions