Skip to main content

This webpage contains ready-to-run meteorological data suitable for the most recent version of AERMOD. Meteorological data were processed with AERMET version 22122 for years 2018 to 2022.

If past versions of meteorological data sets are needed, please email us and specify the surface and upper air station and AERMET version data you would like.

Note that the fulfillment of any request for a meteorological dataset is not an approval for use by MPCA air modeling staff. All meteorological data submittals will be subject to MPCA review. Justification for meteorological site selection should be included as part of the air modeling e-Service protocol or the protocol resubmittal form (AQDM-1.5).

Developing AERMOD-ready meteorological data

There are approximately 80 meteorological surface observing stations in Minnesota, consisting of Automated Surface Observing System (ASOS) and Automated Weather Observing System (AWOS) sites. The majority are located on airport property and follow guidelines laid out in the Federal Meteorological Handbook.

During meteorological processing, the MPCA examined more than 100 stations from Minnesota, Iowa, Wisconsin, South Dakota, and North Dakota to determine usability for modeling. Data usability was based availability of one-minute data and the amount of missing and calm hours per year for each surface station for the years 2018-2022. Stations that met requirements were omitted if a station with superior data quality was nearby. From the original list, 26 surface stations met requirements for further processing, including site-specific daily soil moisture conditions, daily snow cover, and corrected locational data.

The MPCA followed the draft Region 5 Meteorological Processing Guidance (see Resources section below).

Surface meteorological data

Surface meteorological data were obtained from the National Centers for Environmental Information (NCEI) Integrated Surface Database (ISD). This database is also referred to as the Integrated Surface Hourly Database (ISHD or ISH) or TD-3505. ISHD and TD-3280 are currently the only active formats after 1995 for National Weather Service (NWS) files. Additionally, most ASOS stations also have 1-minute average wind speed and direction data available to supplement the ISHD data, by using the pre-processing program AERMINUTE.

This data set, referred to as DSI-6405, consists of a running 2-minute average wind speed reported for every minute. These data are then used in Stage 2 of AERMET to either substitute missing on-site winds or replace standard winds from the ISHD file(s). For more details, see section 2.3.1 of the Addendum to the EPA’s “User’s guide for the AERMOD meteorological preprocessor (AERMET),” as well as EPA’s “AERMINUTE User’s Instructions.” 

At present, on-site meteorological data are not included on this webpage.

Upper air meteorological data

Upper air data (also known as radiosondes or soundings) were obtained from the National Oceanic and Atmospheric Administration (NOAA) and Earth System Research Laboratory (ESRL)’s Radiosonde Database. Data were collected for each individual year, for all sounding times and data levels, with wind units selected as tenths of meters/second. Upper air data were processed separately from the surface data in AERMET, resulting in upper air-specific Stage 1 files for three upper air stations:

  • Aberdeen, SD (WMO #: 72659) (Call sign: ABR)
  • International Falls, MN (WMO #: 72747) (Call sign: INL)
  • Chanhassen, MN (WMO #: 72649) (Call sign: MPX)
Surface characteristics

Surface roughness, Albedo, and Bowen ratio were determined using the EPA’s AERSURFACE processor, which relies on 2016 NLCD Land Use Land Cover (LULC) data for Minnesota, Wisconsin, South Dakota, and North Dakota. Land use was examined around each meteorological surface station’s location, to a radius of 1 kilometer (km). The 1 km buffer was divided into twelve 30-degree sectors and assigned an airport/nonairport flag. Surface roughness values were calculated for each sector. Albedo and Bowen ratio values were also examined using a default area of 10 km by 10 km. Monthly albedo and Bowen ratio values were calculated and applied to all sectors.

Seasonal determinations were made using the spring indices — models that represent the onset of spring season biological activity — in combination with first and last freeze dates.

For the AERSURFACE processing, the first day of spring was set to the leaf out date for the three species of lilacs and honeysuckles. The first day of summer was set to the bloom date for the three plant species, or the last spring freeze, whichever occurred latest. The first day of autumn set to the first fall freeze.

Starting with the 2016-2020 AERMET processing, snow cover and soil moisture are captured at daily temporal resolution. Daily snow cover is extracted from 1km and 4 km resolution geotiff files generated by the United States National Ice Center (USNIC). USNIC analysts produce these geotiffs using the Interactive Multisensor Snow and Ice Mapping System (IMS). Snow cover maps are derived from a variety of data products including satellite imagery and in situ data. Daily soil moisture percentiles are extracted from the Climate Prediction Center. 

Soil moisture percentiles less than 30% were set to "dry," 30-70% to "average," and greater than 70% to "wet". For days with snow cover, soil moisture was set to "average."

AERSURFACE was run four times per year for dry, average, and wet soil moisture with no snow cover and for average soil moisture with snow cover. Likewise, AERMET was run four times per year using these soil moisture and snow cover scenarios. Finally, a script using daily soil moisture and snow cover data generated final AERMET SFC and PFL files for each year. The new methodology generates AERMET files with surface roughness, albedo, and Bowen ratios defined on a daily basis, compared to monthly with the previous methodology.

All AERSURFACE outputs, after post-processing, were used in Stage 2 AERMET input files, as-is.


MPCA pre-processed meteorological data for AERMOD

Meteorological data sets are listed by the meteorological surface station’s three-letter call sign. Ex., MSP = Minneapolis/St. Paul Int’l Airport. Currently, each surface station is processed with one of three available upper air stations, based on proximity: MPX (Chanhassen, MN), INL (International Falls, MN), and ABR (Aberdeen, SD).

AERMET version 22112

Meteorological data were processed with AERMET version 22112 (for use with recent AERMOD versions), which includes the processors AERMINUTE version 15272 and AERSURFACE version 13016. Data are available for the 5-year period of 2018-2022.

Not all meteorological stations that were available for previous versions of AERMET are available for AERMET version 22112. This could be due to data availability, number of missing and calm hours in a dataset, and determination by MPCA of appropriate representative sites. Contact us if the dataset you are looking for is not listed below. Note that the processing of a particular meteorological dataset is not an approval for use by MPCA air modeling staff, and any justification for the use of a particular meteorological dataset is still subject to MPCA review and should be included as part of the air modeling protocol.

During March 2019, one-minute wind was missing for all ASOS sites. AERMINUTE allows five-minute ASOS data to be used when one-minute data is missing. MPCA incorporated these data into the processing of these sites making these approvable for modeling purposes.

Based on the current modeling guidance in Appendix W published in 2017, the adjusted surface friction velocity (ADJ_U*) approach is a regulatory modeling default and the MPCA has processed AERMET data using ADJ_U* as the default data set. A justification is required for using non-ADJ_U* AERMET data. The MPCA can process an AERMET data set using non-ADJ_U* as an option for projects upon request.

Call sign: surface stationElevationAERMINUTECall sign: upper air stationADJ_U*
AXN - Alexandria, MN432 mYesMPX - Chanhassen, MN
BDE - Baudette, MN330 mYesINL - International Falls, MN
BRD - Brainerd, MN372 mYesMPX - Chanhassen, MN
DLH - Duluth Int’l Arpt, MN435 mYesINL - International Falls, MN
**DYT - Duluth Sky Harbor Arpt, MN186 mNoINL - International Falls, MN
EST - Estherville, IA401 mYesMPX - Chanhassen, MN
FAR - Fargo Int'l Arpt., ND272 mYesABR- Aberdeen, SD
FCM - Flying Cloud Arpt, MN276 mYesMPX - Chanhassen, MN
FSD - Sioux Falls Int'l Arpt, SD433 mYesABR- Aberdeen, SD
GFK - Grand Forks Int'l Arp't255 mYesABR- Aberdeen, SD
HCO – Hallock, MN249 mNoINL - International Falls, MN
HIB – Hibbing, MN408 mYesINL - International Falls, MN
INL - International Falls, MN353 mYesINL - International Falls, MN
LSE - La Crosse, WI292 mYesMPX - Chanhassen, MN
MCW - Mason City, IA373 mYesMPX - Chanhassen, MN
MIC - Crystal, MN262 mYesMPX - Chanhassen, MN
MKT – Mankato, MN311 mNoMPX - Chanhassen, MN
MML – Marshall, MN359 mNoMPX - Chanhassen, MN
MSP - Minneapolis/St. Paul Int’l Arpt, MN256 mYesMPX - Chanhassen, MN
PKD - Park Rapids, MN439 mYesMPX - Chanhassen, MN
*RST – Rochester, MN396 mYesMPX - Chanhassen, MN
RWF - Redwood Falls, MN311 mYesMPX - Chanhassen, MN
STC - St. Cloud, MN310 mYesMPX - Chanhassen, MN
STP - St. Paul Downtown Arpt, MN212 mYesMPX - Chanhassen, MN

* Rochester (KRST) is missing one-minute and five-minute data from Nov. 29, 2018 to Aug. 19, 2019. During this period only the hourly observations are used.

** Duluth Sky Harbor (KDYT) METAR AWOS site exceeded EPA Region 5 allowable missing+calms, making 2018-2022 dataset ineligible for use. MPCA has substituted missing+calms with observations from Superior, WI (KSUW). The substituted dataset meets data requirements in 2017, 2018, 2020, 2021, and 2022.

Data files contained within each zip file follow a naming convention of SFCUPA_YYYY (ex., KMSPMPX_2020):

  • The first four characters, SFC, represents the meteorological surface station ID (ex., KMSP);
  • The fifth through seventh characters, UPA, represents the upper air station ID (ex., MPX for the Chanhassen, MN upper air station);
  • SOILS refers to the 4 different surface characteristics: DRY, AVERAGE, WET soils without snow cover, and SNOW (average soil moisture with snow cover)

Each zip file contains:

  • 5-year wind roses (SFCUPA_windrose_YYYYtoYYYY.bmp)
  • 5-year wind frequency counts (SFCUPA_FreqCount_YYYYtoYYYY.csv)
  • 5-year concatenated profile file (AERMET_SFCUPA_YYYYtoYYYY.pfl)
  • 5-year concatenated surface file (AERMET_SFCUPA_YYYYtoYYYY.sfc)
  • 5 subfolders for each individually-processed year, each of which contains:
    • Yearly wind rose (SFCUPA_YYYY_windrose.bmp)
    • AERMET stage 2 input file (AERMET-stage3_SFCUPA_YYYY_SOILS.inp)
    • Wind frequency counts (SFCUPA_YYYY_FreqCount.csv)
    • AERMET stage 2 message file (STAGE3_SFCUPA_YYYY_SOILS.MSG)
    • AERMET stage 1 surface report file (STAGE1SFC_SFC_YYYY.RPT)
    • AERMET stage 1 upper air report file (STAGE1UA_UPA_YYYY.RPT)
    • AERMET stage 3 report file (STAGE3_SFCUPA_SOILS_YYYY.RPT)
    • AERMET stage 3 output – surface file (AERMET_SFCUPA_YYYY_SOILS.sfc)
    • AERMET stage 3 output – profile file (AERMET_SFCUPA_YYYY_SOILS.pfl)
    • Final AERMET

Guidance for selecting the appropriate meteorological data set

40 CFR Part 51, Appendix W, 8.4.1.b. states ”The meteorological data used as input to a dispersion model should be selected on the basis of spatial and climatological (temporal) representativeness as well as the ability of the individual parameters selected to characterize the transport and dispersion conditions in the area of concern.” At a minimum, the following factors should be should be considered when determining the representativeness of a meteorological dataset:

  • The proximity of the meteorological monitoring site to the area under consideration;
  • The complexity of the terrain;
  • The exposure of the meteorological monitoring site;
  • The period of time during which data are collected.

MPCA air dispersion modeling staff has made a detailed effort to make available recent meteorological data, so as to more accurately reflect recent meteorological conditions. In order to determine the most representative meteorological station for use, it is recommended that the subject source consider meteorological sites with similar features. Features to consider include:

  • Wind direction and wind speed patterns
  • Terrain influences on wind patterns
  • Surface characteristics (albedo, Bowen ratios, surface roughness), based on surrounding land use
  • Proximity
  • Time period
  • AERMINUTE (to minimize calm hours) (if applicable, per EPA guidance)

In areas of the state where complex or highly variable terrain occurs (river valleys, lake shores, northeast Minnesota, etc.), it is recommended that a meteorological set be chosen based mainly on surface characteristics, wind patterns, land use, and terrain. In areas of Minnesota where terrain is relatively flat and invariable, then proximity, surface characteristics, land use, and wind patterns are to be examined.

To aid in meteorological site selection, the MPCA has created an online meteorological site selection tool (MSST). This tool was created to produce an objective analysis of the representativeness measures discussed above. In particular, the MSST analyzes proximity, terrain, surface characteristics, urban vs. rural and snow cover to determine meteorological similarity between a selected facility and surrounding meteorological sites. If a user disagrees with the site selected by the MSST, a different site can be selected using the justification provided by the tool. See the technical support document and user’s guide below for more details.

No matter the subject source and the reliance on past meteorological sites, please provide explanation/reasoning for your choice of met sets in the air modeling e-Service protocol.