W**"*"*" Ei'titon
Guidance on the Use of the Mesoscale Model
Interface Program (MMIF) for AERMOD
Applications
-------
-------
EPA-454/B-15-001
July 2015
Guidance on the Use of the Mesoscale Model Interface Program (MMIF) for AERMOD
Applications
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Air Quality Assessment Division
Air Quality Modeling Group
Research Triangle Park, North Carolina
-------
Preface
This document provides guidance on the use of prognostic meteorological data and the
Mesoscale Model Interface Program (MMIF) in AERMOD. Included in this document are
descriptions of the inputs to MMIF and recommendations on using MMIF output in AERMOD.
11
-------
Acknowledgements
MMIF was developed by Environ International Corporation under EPA contract number EP-D-
07-102, work assignments 2-06, 4-06, 5-08, and 10-1. The MMIF user's guide was developed
by Bart Brashers and Chris Emery of Environ. This guidance document was developed by the
Air Quality Modeling Group of the Air Quality Assessment Division of the Office of Air Quality
Planning and Standards in collaboration with representatives of EPA Regions 5, 7, and 8.
in
-------
Contents
Preface ii
Acknowledgements iii
Tables v
1. Introduction 1
2. Guidance on using prognostic meteorological data for use in AERMOD 1
2.1 Number of years to model 2
2.2 Prognostic model options 2
2.2.1 Development of meteorological fields 2
2.3 Model output quality assurance 2
2.3.1 Operational evaluation 2
3. Guidance on running MMIF for AERMOD 3
3.1MMIF Input File 3
3.2 Recommended options for selected keywords 6
3.2.1 Outputs 6
3.2.1.1 AERMET 6
3.2.1.2 AERMOD 7
3.2.2 Output layers and heights 7
3.2.3 Grid cells to process 7
3.3 Surface characteristics 7
3.4 Treatment of low winds 8
3.5 Post-processing of outputs 8
4. References 10
IV
-------
Tables
Table 1. AERMET/AERMOD keywords in MMIF input file 4
-------
1. Introduction
The guidance discussed in this document are recommendations for running the Mesoscale Model
Interface (MMIF) program1 to generate inputs for AERMET and AERMOD. In this guidance
document, when references are made to running MMIF for AERMOD, it should be inferred that
this refers to AERMET as well. For regulatory applications, MMIF should be run to generate
AERMET inputs as stated in section 8.4.2(a) and 8.4.5.1(b) of the proposed revisions to EPA's
Guideline on Air Quality Models (U.S. EPA, 2015)2. Regulatory applications that do not follow
these sections of Appendix W will need to consult with the appropriate reviewing authority and
guidelines outlined in section 3.2 of Appendix W. Given that Appendix W and specific EPA
modeling guidance are often cited in relation to other non-regulatory modeling applications, such
as air quality analysis and disclosure purposes under NEPA, the approach presented in this
guidance document for regulatory applications also has relevance to these non-regulatory
applications. While MMIF can process data for other air quality models (e.g., CALPUFF and
SCICHEM), the emphasis in this guidance is for AERMOD applications conducted for
regulatory purposes.
This guidance document will summarize some of the inputs needed for AERMET and
AERMOD MMIF processing, but will refer to the MMIF User's Guide (Environ, 2014) for more
details. MMIF users are strongly encouraged to read this user's guide to obtain specific details
on running MMIF.
2. Guidance on using prognostic meteorological data for use in AERMOD
In general, air quality modeling applications rely on the use of meteorological grid models.
These models are used to more accurately simulate atmospheric processes (e.g., temperature,
wind speed and direction, etc.) across a specific area. In retrospective simulations (i.e., modeling
past events), the blending of observed data with computed fields yields results that are bound by
ground truth.
There are several meteorological grid models that can be used to develop inputs for air quality
models. The most commonly used by EPA and the modeling community is the Weather
Research and Forecasting (WRF) model (Skamarock et al., 2008)3, which is supported across a
broad community and provides state-of-the-science parameterizations of the atmosphere.
Additionally, the Fifth Generation Penn State/NCAR Mesoscale Model (MM5) (Grell et al.,
1 http://www.epa.gov/ttn/scram/dispersion_related.htm#mmif
2 Hereafter, the Guideline will be referred to as Appendix W.
3 http://www.wrf-model.org/index.php
1
-------
1994) is capable of generating the necessary meteorological inputs to air quality models;
however, its development and maintenance is no longer supported.
2.1 Number of years to model
As discussed in Section 8.4.2(e) of the proposed Appendix W, at least three consecutive years
are required to be modeled in the prognostic model. The most recent three years are preferred
and the prognostic model domain or selected grid cells should be representative of the domain.
See Section 8.4.b of the proposed Appendix W for more details about representativeness of
meteorological data.
2.2 Prognostic model options
2.2.7 Development of meteorological fields
Section 2.6 of the Modeling Guidance for Demonstrating Attainment of Air Quality Goals for
Ozone, PM2.5, and Regional Haze (U.S.EPA, 2014) recommends approaches for developing
prognostic meteorological data. While specific recommendations for model options are not
provided, this guidance discusses the proper steps to ensure the data are representative of the air
quality model domain. The reader is referred to U.S. EPA (2014) for details on the development
of the meteorological fields.
2.3 Model output quality assurance
2.3.1 Operational evaluation
Demonstration of the adequacy of prognostic meteorological fields can be established through
appropriate diagnostic and statistical performance evaluations consistent with recommendations
provided in the appropriate EPA guidance. A quantitative, statistical, and graphical analysis of
the prognostic data should be completed, comparing the data to available NWS automated
surface observation station (ASOS) data, as well as operational profiler data (if available),
pairing both in space and time. This analysis should be completed for all years (at least three) of
prognostic meteorological data to be used in the air quality simulations. Since the spatial scope
of each variable could be different, representativeness should be judged for each variable
separately as discussed in Section 8.4.2(b) of the proposed Appendix W (U.S. EPA, 2015a). For
example, for a variable such as wind direction, the data should ideally be collected near plume
height to be adequately representative; especially for sources located in complex terrain,
whereas, for a variable such as temperature, data from a station several kilometers away from the
source may be considered to be adequately representative. The grid resolution of the prognostic
-------
meteorological data should also be considered and evaluated appropriately, particularly for
projects involving complex terrain. Several software packages are available for use in
completing this evaluation (e.g., AMET (Appel et al., 2011) and METSTAT
(http://www.camx.com/download/support-software.aspx)). The adequacy of output from the
meteorological models is contingent upon the concurrence with the appropriate reviewing
authorities as defined in section 8.4.5.2(a) of the proposed Appendix W.
3. Guidance on running MMIF for AERMOD
Much of the guidance presented here follows the MMIF user's guide (Environ, 2014). Relevant
information from the user's guide is summarized in this guidance for convenience but the user is
strongly encouraged to read the full MMIF user's guide before attempting to run MMIF. Section
3.1 below discusses the inputs to MMIF, Section 3.2 discusses the relevant options to AERMOD
and grid cells to process. Section 3.3 discusses the use of surface characteristics outside of
MMIF and Section 3.4 discusses post-processing the output from MMIF needed for input into
AERMOD.
3.1 MMIF Input File
MMIF processing is done via a control file with keywords to denote inputs, processing options,
and outputs. Table 1 lists the keywords used to run MMIF for AERMET and AERMOD input.
A sample control file that illustrates all of the keywords can be generated for MMIF by typing
"mmif-sample" at the command prompt. See Section 4.2 of the User's Guide (Environ, 2014)
for more information.
-------
Table 1. AERMET/AERMOD keywords in MMIF input file.
Keyword
Description
Syntax
Start
Date and time (Local Standard Time) to start
processing
Start YYYY MM DD HH
Or
Start YYYYMMDDHH
Or
Start YYYY-MM-
DD HH:mm:ss
Stop
Date and time (LST) to stop processing
Stop YYYY MM DD HH
Or
Stop YYYYMMDDHH
Or
Stop YYYY-MM-
DD HH:mm:ss
Timezone
The global time zone shift from Greenwich Mean
Time (GMT); Western
hemisphere time zones are denoted by negative
numbers
Timezone HH
Grid
Specifies the requested output sub-grid's lower left
(LL) and upper right (UR) corners; Grid corners can
be specified by grid cell I, j coordinates (IJ), latitude
and longitude (LL or LATLON) or MM5/WRF
projected coordinate system (KM)
GRID IJ iLL jLL iURjUR
Or
GRID LL LatLL LonLL LatUR
LatRR
Or
GRID KM xLL yLL xUR yUR
Point
Output point for AERMET, AERMOD, or
AERCOARE processing. The point can be specified
by grid cell I, j coordinates (IJ), latitude and
longitude (LL or LATLON) or MM5/WRF projected
coordinate system (KM). An optional time zone
shift can also be listed1. The point keyword can be
repeated for each point to be outputted.
Point IJ IJ [Timezone]
Or
POINT LL Lat Lon [Timezone]
Or
POINT KM X Y [Timezone]
Layers
Specify the output layer structure. Layers can be
aggregated (K), interpolated using layer tops (TOP),
or interpolated using mid layer (MID).
Layers K Layeri
Layer2...LayerN
Or
Layers TOP Topi Top2...TopN
Or
Layers MID Midi Mid2...
Origin
Over-ride the X,Y grid origin values found in the
MM5 or WRF output file. The user specifies a
latitude (LAT) and longitude (LON).
Origin LAT LON
PEL recalc
A value of FALSE (default value) causes MMIF to
pass through the PEL depth from the model with no
changes. A value of TRUE causes MMIF to re-
calculate PEL depths using a Bulk Richardson
approach with 20 times the vertical resolution of the
model data.
PBL_recalc FALSE
Or
PEL recalc TRUE
AER MIN SPEED
Specify the minimum wind speed in m/s (VALUE)
for AERMOD surface output file.
AER MIN SPEED VALUE
FSL INTERVAL
Specify the number of hours (VALUE) to write for
each day to the upper air file for input into
AERMET. The default value is 12 representing the
OOZ and 12Z soundings. A value of 6 would write
output for OOZ, 06Z, 12Z, and 18Z. A value of 1
would write output for each model hour.
FSL INTERVAL VALUE
-------
Table 1. Continued
Keyword
AER_LAYERS
OUTPUT
INPUT
METFORM
Description
Specify the lowest and highest layer indices (two
integers) to write to the AERMET input site-specific
data and AERMOD profile file (PFL file). All layers
between the two indices will be written to the file.
Specifies the outputs from MMIF for AERMET,
AERCOARE, and AERMOD (MODEL keyword) and
output files2.
Input MM5 or WRF filename. This input is repeatable
for a MMIF run.
Keyword to tell MMIF which model MM5 or WRF is
being accessed. MMIF can auto-detect the model type
so in general this is not needed.
Syntax
AER_LAYERS VALUE
OUTPUT MODEL FORMAT
FILENAME
INPUT FILENAME
METFORM MM5
Or
METFORM WRF
1. See Section 4.2 of the MMIF user's guide regarding the global time zone shift and point specific time zone
shifts.
2. More information about the output options are discussed in Section 3.2 below.
-------
3.2 Recommended options for selected keywords
While most input options will be left to the discretion of the user, some recommendations on
inputs are made in this guidance. One such option is the output option of MMIF, keyword
OUTPUT. While MMIF can process data for input into AERMET, or AERMOD, the
requirement for regulatory applications is to process the prognostic meteorological data for input
into AERMET as discussed in Sections 8.4.2(a) and 8.4.5.1(b) of the proposed Appendix W, as
AERMET is the meteorological pre-processor for AERMOD as discussed in those sections..
The data is then processed in AERMET for input into AERMOD. Processing MMIF output
through AERMET also allows the user to take advantage of some of the options in AERMET,
such as the u* adjustment option. See the AERMET user's guide and addendum for details
about options (U.S. EPA, 2004; U.S. EPA, 2015b). For non-regulatory applications, the user
may choose AERMET or AERMOD and should consult with and seek concurrence from
collaborating agencies or parties involved in such modeling applications.
3.2.7 Outputs
For any particular air quality model, the OUTPUT option is used to specify several files. While
these are discussed in the MMIF user's guide in detail, they are summarized below for AERMET
and subsequent input AERMOD and direct input into AERMOD as well (non-regulatory
applications).
3.2.1.1 AERMET
For AERMET, the first set of files is specified using the USEFUL keyword. This keyword
creates a DOS batch file or Linux shell script that is used to run all three stages of AERMET in
batch mode. It also creates the stage 1, stage 2, and stage 3 input files with the appropriate
values set for the AERMET keywords such as LOCATION, XDATES, etc. Once the stage 1, 2,
and 3 input files have been created, the user should check those files to ensure the correct GMT
offset is used. For the upper air pathway of the stage 1 file, the LOCATION keyword should
have a GMT offset corresponding to the station's location. For example, if the processed grid
cell is in the Eastern timezone of the U.S. the GMT offset on the LOCATION keyword should
be 5. For the surface data, the offset should be zero as that has been formatted for local time.
The second file that is generated is specified using the keyword ONSITE. This creates a site-
specific type meteorological file that is processed via the ONSITE keyword in AERMET (U.S.
EPA, 2014). This file contains 2-meter and 10-meter data and upper air data up to levels
specified with the keywords MIN_LAYER, MAX_LAYER, or LAYERS to control the number
of output layers.
The third keyword, FSL creates a file that mimics an upper air data file in the Forecast Systems
Laboratory (FSL) format. The keyword UPPERAIR can also be used. See Section 2.2.1 of the
MMIF user's guide for more details.
-------
The final keyword is AERSFC, which generates an AERSUKFACE type output file with surface
characteristics (albedo, Bowen ratio, and surface roughness). Note, that these are monthly
surface characteristics for the period being processed. See Section 2.2.1 of the MMIF user's
guide for more details.
3.2.L2AERMOD
Three files are generated for AERMOD. The USEFUL file is a file containing the ME pathway
information of the AERMOD input file, i.e. ME STARTING, SURFFILE, PROFFILE,
SURFDATA, UADATA, etc. information. The SFC keyword generates the AERMOD ready
surface data file and the PFL keyword generates the profile data file for input into AERMOD.
3.2.2 Output layers and heights
An important keyword for output is the LAYERS keyword. As shown in Table 1, the user can
specify different options for the output layers from MMIF. While the choice of layers is case
specific and may be dependent on the prognostic model's layer structure, a default use of MID
(interpolation using layer mid-point heights) and the specification of heights corresponding to the
AERMOD vertical grid heights should be adequate in most cases. These heights are: 25, 50, 75,
100, 125, 150, 175, 200, 250, 300, 350, 400, 450, 500, 600, 700, 800, 900, 1000, 1500, 2000,
2500, 3000, 3500, 4000, 4500, and 5000. These values have been used in past MMIF
evaluations (U.S. EPA, 2015c).
3.2.3 Grid cells to process
An AERMOD run uses surface meteorological data from one point and upper air data from one
point. While MMIF can process multiple points, i.e. grid cells, the grid cell used in the
AERMOD simulation should be representative of the modeling domain, following the
recommendations of Section 8.4.5(b) of Appendix W. Depending on the size of the modeling
domain and the grid resolution of the prognostic meteorological data, most often the
representative grid cell will be the grid cell containing the facility of interest. This will often be
the case for NSR/PSD types of applications. When the AERMOD modeling domain overlaps
several grid cells of the prognostic meteorological data, such as for SIP demonstrations, the grid
cell that is most representative of the domain should be selected following guidance on
representativeness in Sections 8.4.1.b and 8.4.2.b of the proposed Appendix W.
3.3 Surface characteristics
MMIF will output surface characteristics, albedo, Bowen ratio, and surface roughness for input
into AERMET and also in the AERMOD ready surface meteorological file. When outputting
data for AERMET, MMIF outputs surface characteristics for one 360° sector at monthly
-------
resolution. The surface characteristics are based on the landuse data used by the prognostic
meteorological model. These surface characteristics should be used as they are representative of
the processed grid cell as discussed in Section 8.4.2(b) of the proposed Appendix W.
3.4 Treatment of low winds
When processing MMIF for input into AERMET, MMIF will generate the onsite wind speed
threshold option (THRESHOLD keyword) with a value of 0.5 m/s for the stage 1 AERMET
input file. The model user has the option to include this option when running AERMET. See the
AERMET user's guide and addendum (U.S. EPA, 2004; U.S. EPA, 2015b) for details about
AERMET this option. When generating MMIF output for direct AERMOD input, winds below
0.5 m/s are treated as calms in the AERMOD surface file.
3.5 Post-processing of outputs
When processing MMIF for AERMET files, a single MMIF run will produce an upper air file in
the FSL format, a surface data file that will be read into AERMET as site-specific data, and
surface characteristics (albedo, Bowen ratio, and surface roughness) at monthly resolution for
twelve sectors. These output files will cover the period processed in MMIF. In most situations,
a single MMIF run will not cover an entire three period or even a one year period. If that is the
case, the ONSITE files generated by the MMIF can be simply concatenated into a single file for
the three year period or individual yearly files before input into AERMET. The files must be
concatenated in temporal order. The same can be done for the FSL files. If using the MMIF
surface characteristics files, the AERSFC files cannot be simply concatenated. For a single
MMIF run, the surface characteristics are output for all twelve months and sectors. The months
outside of the data processing window set by START and STOP will have missing values, while
the months inside the window will have non-missing values. To create a valid AERSFC file
covering the entire three year period or desired period, an AERSURFACE file must be created
with non-missing values for all months and sectors. This can be created by simply cutting and
pasting the non-missing values for each month/sector combination into a single file. When
creating this file, the user should make sure to incorporate the header line from one of the files
into the concatenated file. This line is "** Generated by MMIF..." This line indicates to
AERMET that the meteorological data comes from MMIF and not an observed site-specific
dataset. This information is then passed to AERMOD via the surface meteorological file created
by AERMET. While this does not affect the data calculations in AERMET and AERMOD,
including the line ensures transparency when data files are reviewed.
An alternative approach to the file concatenation steps described above, is to run AERMET for
each period processed and concatenate AERMOD ready surface and profile files from the
multiple AERMET runs. For the profile files, the files can be simply concatenated together,
preserving the temporal order of the data (e.g., January 1, hour 1 of the first processed year is the
first line and December 31, hour 24 of the last processed year is the last record of the
-------
concatenated file). For the surface files, AERMET generates a header record for each file (the
record that lists the location, station identifiers, and AERMET version). When concatenating the
surface files, the header record for the first concatenated file should be retained. Only the data
records from the remaining surface files are needed. If the header records are retained for all
files, AERMOD will not run correctly. Again, the files should be concatenated in temporal
order. These steps also apply for processing AERMOD ready files when post-processing MMIF
output for AERMOD.
-------
4. References
Appel, K.W., Gilliam, R.C., Davis, N., Zubrow, A., and Howard, S.C.: Overview of the
Atmospheric Model Evaluation Tool (AMET) vl.l for evaluating meteorological and air
quality models, Environ. Modell. Softw., 26, 4, 434-443, 2011.
Environ, 2014: The Mesoscale Model Interface Program (MMIF) Version 3.1 User's Manual.
Grell, G., J. Dudhia, and D.R. Stauffer: A description of the fifth-generation Penn State/NCAR
Mesoscale Model (MM5), NCAR Tech. Note NCAR/TN-398+STR, 122 pp.
Skamarock, W. C., Klemp, J. B., Dudhia J., Gill, D. O., Barker, D. M., Duda, M. G., Huang, X.-
Y., Wang, W., and Powers, J. G.: A Description of the Advanced Research WRF
Version 3, National Centre of Atmospheric Research, Boulder, Colorado, 2008.
U.S. EPA, 2004: User's Guide for the AERMOD Meteorological Preprocessor (AERMET).
EPA-454/B-03-002. U.S. Environmental Protection Agency, Research Triangle Park,
NC 27711.
U.S. EPA, 2015a. Guideline on Air Quality Models. 40 CFRPart 51 Appendix W.
U.S. EPA, 2015b: Addendum - User's Guide for the AERMOD Meteorological Preprocessor
(AERMET). EPA-454/B-03-002. U.S. Environmental Protection Agency, Research
Triangle Park, NC 27711.
U.S. EPA, 2015c: Evaluation of Prognostic Meteorological Data in AERMOD Applications.
EPA-454/R-15-004. U.S. Environmental Protection Agency, Research Triangle Park, NC
27711.
10
-------
-------
United States Office of Air Quality Planning and Standards Publication No. EPA-454/R-15-004
Environmental Protection Air Quality Assessment Division July, 2015
Agency Research Triangle Park, NC
------- |