ASOS/AWOS (United States METAR)

Synoptic’s oldest and most widely used dataset is the ASOS/AWOS network, mesonet ID 1.

ASOS/AWOS combines two independent data sources into a single time series.

1. The official METAR feed for the United States and neighboring countries - covering every airport that produces METAR and SPECI observations. These come from 3 types of platforms. ASOS stations, AWOS stations, and manual reports. Synoptic does not know which station is which type of equipment.

2. A 5-minute sample of the High Frequency ASOS dataset produced by NOAA and shared publicly. A “METAR” string is created for these messages, but this is not a genuine METAR. This is referred to as “HFMETAR” in this document. About half the dataset has these observations.

Dataset nomenclature

The name “ASOS/AWOS” is a compromise among US federal agencies, where the dataset itself is maintained by a combination of NOAA, FAA and US DOD entities. ASOS refers to the Automated Surface Observing System, which is a US-Government managed type of airport sensing platform maintained generally by the National Weather Service, and held in the highest standard.

AWOS stands for Automated Weather Observing System, and is the general term for any aviation-grade (which is an extremely high standard) airport observing system. Most DOD sites, and many smaller airports have AWOS’s, which may be maintained by government or private maintainers. There are multiple grades of AWOS, and this network will share observations from comissioned units which are certified to send out automated METAR.

METAR is a foundational meteorological format, which has several possible definitions, https://en.wikipedia.org/wiki/METAR but, is used in aviation as the official indicator of the weather conditions at an airport. These reports are sent hourly at a minimum, but can be as frequent as every 15 minutes in some sites. METAR typically reports at the same time each hour, and in the united states a common time to report is 53 minutes past the hour.

A SPECI is a type of METAR report which is in the same dataset as METAR, but is distributed as a result of a significant change in the meteorological conditions between regular METAR reports. These can occur at any minute.

Latency

The two data sources for this dataset have substantially different latencies. Formal METARS are available from the Synoptic platform within 3 minutes of their sampling in most conditions. This is in line with most global sources for METAR observations, and we are working to improve on this (June 2026). SPECI observations have the same latency characteristic.

The HFMETAR component has latency of 5-9 minutes or more. This data is produced upstream of synoptic, and is delivered in 5-minute batches. This latency can also vary from time to time depending on the workload of that upstream source.

Our latency monitoring tools ( Data Availability Dashboard ) work on a per-dataset basis, meaning their latency reports will be a combination of the HF and normal METARS. This is not a valid representation of the latency of METARS alone.

Weather events do not generally impact availability of ASOS/AWOS data, however local disruptions can prevent specific stations from being accessible.

A note about HFMETAR observations

Some stations in the ASOS/AWOS dataset are augmented with HFMETAR data. HFMETAR have all the same drawbacks as the High Frequency ASOS dataset, and more. They have higher latency (even than the HFASOS feed itself), lower precision, and fewer variables. Synoptic has been producing a METAR string from these lower-quality observations, that has led to significant confusion regarding these observations. To be clear, these are not valid METAR observations, they are not based on the sampling conventions for METAR, and they have not received the quality control performed on METAR.

Weather API supports an argument specially for this dataset of &hfmetar=0 to prevent HFMETAR values from being returned from the API.

Presently (June 2026) the push streaming service does not support this argument and there is no way to exclude the HFMETAR observations.