High Frequency ASOS
This is an Opt-in public dataset
Since the mid 2010's the FAA has permitted/enabled many major airport weather stations, which are part of the Automated Surface Observing System (ASOS) to measure data at a one-minute temporal frequency than the traditional METAR frequency, which is "hourly with reports of significant change at any time."
Because airport weather stations are operated as a cooperation between the FAA and NOAA, Synoptic receives HF-ASOS data via NOAA's MADIS service.
HF-ASOS are also periodically referred to as “One minute observations” or “OMO”
Considerations
There are many nuances to receiving and using the HF-ASOS data.
Not every airport
HF-ASOS are only available at most ASOS airport sensors. Many airports in the US are a different kind of station, known as an AWOS (automated weather observing station). The acronym may seem to mean the same thing, but the effect is that these other airport stations are often configured and operated differently than ASOS - which are all installed and managed by NOAA for the NWS and FAA. AWOS stations are not available for HF data.
At this time ASOS and AWOS data are intermingled in our network ID 1 "ASOS/AWOS" and simply looking at our metadata we cannot immediately tell you whether a site is an AWOS or ASOS. Most of the major commercial airports are ASOS, and do have the HF service.
Data organization and how to get it
Synoptic exposes HF-ASOS data in two ways. Since 2017 we have received and publicly delivered a 5-minute subset of the HF-ASOS data, provided to us via MADIS, and we have blended those measurements in with our ASOS/AWOS network data. This has always been public data, and a special argument was added to Weather API - hfmetars - to allow queries to ignore these values and only provide the traditional METAR observations.
Additionally, since 2019 Synoptic has received the full 1-minute data, as a low-latency provisional and real-time data stream from MADIS. Instead of combining these data with the existing METAR network, we created a new network to organize full HF-ASOS into, where the stations use their matching IDs (so Salt Lake is KSLC) with an additional "1M" - so KSLC is available for HF-ASOS as KSLC1M. This is only available for stations which have reported the HF-ASOS data.
We are in the process of revising this (September 2022) but for now the 1-minute HF-ASOS data is considered public, and can be accessed by anyone at no charge, however we require you to ask us to enable it, as it is a more complicated dataset to use, and the nuances noted in this document are important. So please submit a help ticket to have HF-ASOS data included in your API requests. The download service cannot offer access to HF-ASOS at this time.
Operational availability
The HF-ASOS data feed is considered experimental. This means that the FAA and NOAA are both not giving full operational effort towards sustaining the dataset, like they would for the regular METAR reports. This means that brief to several-day outages are not uncommon. Like most outages, these are completely outside of Synoptic's influence. We do not recommend developing applications solely based on the consumption of HF-ASOS data.
Our Demos site has a tool that indicates the if the HF-ASOS data feed is flowing into Synoptic.
Latency
For availability latency (the time between when the observation was made and when Synoptic receives it) HF-ASOS, when operating normally, usually arrives with between 7 and 15 minutes latency. Despite the 1 minute resolution, they arrive at synoptic in 5-minute batches. This batching means that there is no real-time benefit over a 5-minute version of the data we have had available within the standard METAR (ASOS/AWOS) data for several years.
Variable considerations
Many early users have noted that the HF-ASOS exhibit different variable characteristics from the METAR reports for the same stations. This is true, and is a result of the very different technology and process the HF observations take to get to us. But the most important considerations are:
Temperatures are expressed in whole degrees celsius
Wind speeds are a 2 minute average, even when looking at 1-minute observations.