PlumeSmartEPA Regulatory Air Dispersion Modeling
‹ Runs
Air Modeling GuideCovers AERMOD 24142 & 26135 — every excerpt version-stamped · current NAAQS

AERMINUTE: recovering the wind hours METAR throws away

The standard hourly weather archive was built for aviation, not dispersion modeling — and its coding rules silently delete exactly the hours a dispersion model cares most about. AERMINUTE exists to get them back. Its User's Guide states the problem completely:

A potential concern related to the use of NWS meteorological data for dispersion modeling is the often high incidence of calms and variable wind conditions reported for the Automated Surface Observing Stations (ASOS) in use at most NWS stations since the mid-1990's. In the METAR coding used to report surface observations beginning July 1996, a calm wind is defined as a wind speed less than 3 knots and is assigned a value of 0 knots. The METAR code also introduced the variable wind observation that may include wind speeds up to 6 knots, but the wind direction is reported as missing if the wind direction varies more than 60 degrees during the 2-minute averaging period for the observation. The AERMOD model currently cannot simulate dispersion under calm or missing wind conditions.

Source · AERMINUTE User's Guide (EPA-454/B-26-006), version 26135 §1 — captured verbatim in the source library.

Read that as a modeler: light-wind hours are high-impact hours — in AERMOD's concentration equation the effective wind speed is in the denominator (C = Q/ũ · P<sub>y</sub> · P<sub>z</sub>; Model Formulation EPA-454/B-26-003 §5, eq. 51), so less wind means less dilution. Yet METAR's coding turns a 2-knot wind into a "calm" and a meandering 5-knot wind into a missing direction — both hours AERMOD cannot model. A met record assembled only from the standard archive quietly loses many of its highest-concentration hours.

The fix: average the whole hour instead of sampling it

ASOS stations actually measure much more than the hourly report shows — NCEI archives "running 2-minute average winds, reported every minute, for commissioned ASOS stations" (§3). AERMINUTE turns those into a real hourly average:

To reduce the number of calms and missing winds in the surface data, archived 1-minute winds for the ASOS stations can be used to calculate hourly average wind speed and directions, which are used to supplement the standard archive of hourly observed winds processed in AERMET.

Source · AERMINUTE UG (EPA-454/B-26-006) §1.

standard hourly archive (METAR)one 2-minute average stands for the hour< 3 knots → reported as calm (0 kt) · direction varying > 60° at ≤ 6 kt → direction missing — AERMOD cannot model either hourAERMINUTE from 1-minute ASOS datanon-overlapping even-minute 2-minute averages tile the hour → a true hourly mean speed & directionhh:00hh+1:00hh:00hh+1:00result: far fewer calm and missing-wind hours in the AERMET record — the hours AERMOD would otherwise drop
Why 1-minute data helps — one observation vs an average over the whole hour (AERMINUTE UG §1, §4.6)

The averaging is careful about overlap: "Even minute winds are given first priority in calculating hourly average winds" — each even-minute value is a 2-minute average ending at that minute, so the even minutes tile the hour without double-counting; minute 1 of each hour is never used because its 2-minute average spans two hours; and the hourly averages are hour-ending to match AERMET and AERMOD's hours 1–24 (§4.6). The 1-minute values also "have not been subjected to the METAR coding for calm and variable winds" (§3) — which is precisely why the recovered hours exist. Where 1-minute files are missing, AERMINUTE can fall back to the 5-minute ASOS archive (since version 15272, §1).

The details that carry regulatory weight

  • Appendix W expects this. The Guideline's meteorology section says "The AERMINUTE processor, in most cases, should be used to process 1-minute ASOS wind data for input to AERMET when processing NWS ASOS sites in AERMET" — 40 CFR 51 App. W §8.4.2(a) (quoted in full in Meteorological data & representativeness).
  • Truncation is fixed downstream. All ASOS wind speeds "are truncated, rather than rounded, to whole knots"; AERMINUTE does not correct that bias — AERMET applies the truncation adjustment "for all ASOS wind speeds, including those derived from 1-minute data" (§4.6). The two tools are designed as a pair.
  • The anemometer type matters. Stations in the Ice Free Winds (IFW) group "use sonic anemometers instead of cup and vane anemometers," and "AERMINUTE treats wind speeds less than 2 knots differently based on whether the station uses a sonic or cup and vane anemometer during the data period" (§4.1.2) — so the station's IFW commissioning date is a real input, not trivia.
  • Limited QA is the trade-off. The 1-minute archive has "been subjected to only limited quality control measures" (§3) — AERMINUTE applies its own quality checks (its Appendix A) precisely because the raw feed is not the quality-assured METAR product.

Version 26135 of AERMINUTE adds a reader for the GHCNh format — the archive that replaced ISHD at NCEI — "for comparison to 1-minute and 5-minute values" (§2.5), keeping the tool aligned with where the surface data now live.

In PlumeSmart

When the platform builds meteorology for an ASOS station with available 1-minute data, AERMINUTE runs in the pipeline automatically and its hourly averages supplement the standard archive in AERMET — with the station's data period and processing recorded alongside the met dataset.