About
Tracks, time series and quality control for every Saildrone, Oshen and Chance Maritime uncrewed surface vehicle published on the NOAA PMEL ERDDAP — 153 deployments, 2017 to now, active and historic.
Where the data comes from
Everything here is NOAA's. The vehicles are operated by PMEL, AOML and NEFSC with Saildrone Inc., Oshen and Chance Maritime Technologies, and published on the PMEL ERDDAP. Each record's page links to its own ERDDAP dataset, which carries the licence and the acknowledgement that govern it; this site republishes neither more nor differently.
This site is not live, and it cannot be.data.pmel.noaa.gov sends no Access-Control-Allow-Origin header on any response, so a page in your browser cannot read it at all. The data here was fetched by the build, checked, and written out as JSON this origin serves. Every page prints when its record was fetched. If you need the current hour, follow the ERDDAP link on the record and take it from the source.
What was done to the numbers
Nothing was removed and nothing was smoothed. Three things were changed, and all three are stated on the page where they apply:
- Units were converted where the archive is inconsistent. Every Oshen publishes wind in knots and every other vehicle in m/s; one Chance column publishes a bearing in radians. Each record's quality report lists every conversion applied to it.
- Records were decimated to at most 8,000 points, by keeping every nth real sample rather than by averaging — an average would put numbers on screen that no instrument reported. The quality checks ran before that, on the finest resolution the build fetched.
- Values were rounded to what each instrument resolves: sea temperature to 0.0001 °C, pressure to 0.01 hPa, position to five decimals, and so on.
What was computed here
Seven quantities are derived rather than published, and each is marked with an ƒ in the variable list:
- U₁₀, the wind adjusted to 10 m through a neutral logarithmic profile from the sensor's own height — 0.66 m on an Oshen, about 3.4 m on a Saildrone and published per record because the wing moves. Without it the two cannot be compared:the Oshen adjustment is +31 %. Neutral is an assumption, not a measurement; in a strongly stable surface layer it over-corrects.
- Wind stress, ρa CD U₁₀², with the Large & Pond (1981) drag coefficient.
- Dewpoint and specific humidity, from the Magnus form with Alduchov & Eskridge (1996) coefficients.
- Absolute Salinity, Conservative Temperature, σ₀, spiciness and sound speed, from TEOS-10 at the surface. Every variable named pressure on a USV is atmospheric, so the sea pressure is a constant 0 dbar — feeding a barometer to a seawater routine as decibars inflates in-situ density by 4.3 kg/m³ and sound speed by 17 m/s, and every number that comes back still looks like seawater.
What the quality checks do and do not claim
The archive publishes no quality flags on almost all of this data, so every quality statement here was made by this site. The quality page says what each of the nine checks looks for and — more usefully — what each one is not.
A finding is a second opinion, not a correction. It marks the data and never alters it. A record with nothing found is a record these nine checks found nothing in, which is a smaller claim than "clean".
What is not here
- Analysis. Intercomparison, storm context and glider collocation live in NOAA-USV-analysis. This site shows what the vehicles measured and what is wrong with it; it does not draw conclusions from it.
- Vertical structure. These are surface vehicles. Eleven Chance Maritime file listings hold CTD casts, echosounder files and imagery behind per-cast files — real profile data, and the only vertical structure this archive has. Reading them is open work.
- Every column. Each record carries at most 40 of its variables here, chosen canonical-first. The ERDDAP link on every page has all of them at full resolution.
Credits
Built by Donglai Gong (Virginia Institute of Marine Science, William & Mary) at the Collaboratory for Physical Oceanography, with Claude (Anthropic). Source at oceansensing/usv; all rights reserved.
Sibling sites: gliders, for the IOOS Glider DAC, which shares this one's plot engine and TEOS-10 implementation. Basemap by Esri — GEBCO, NOAA, National Geographic.