Regarding the sonde from yesterday morning with the 219-kt spot wind, highest dropsonde wind ever seen in a TC, AOML/HRD shared the raw data and I took a look this morning. They’ll have the final call, but I see nothing wrong with the ob and I suspect it’s going to hold up.

Was 234 knot sonde at the 866mb level in Katrina unreliable? tropicalatlantic.com/recon/recon.... Nothing really matches close to it for that storm though, expect one with 182 knots on another mission at the 839mb level. Melissa had another mission with one with 210 knots & other high readings.

Also, do you know what the wind reported in the dropsonde message is averaged over? (The ones released in WMO messages) Such as a half second perhaps? Or 0.25, or 1? I would like to put that information on my site. Maybe it has varied over the years and I can put a range if so.

The raw wind data are taken every 0.25 s. Before transmission they’re run through a filter with a 5-s wavelength. Until a couple years ago it was a 10-s filter.

Thanks. Does that mean it currently takes 2.5 seconds of data above the pressure altitude and 2.5 seconds below it and averages it all together for the reported pressure altitude? And then for surface wind is it less than 5 seconds or the last five seconds before the sonde terminates?

Not quite. The processing code has an almost 30-year history through three main developers, of which I was the first, but believe it’s still a low-pass filter with a half-power amplitude set at any desired wavelength - in this case 5 s. Near endpoints filtering tapers to zero.

I was trying to put together a disclaimer for my website to add to dropsonde data. I'm not a meteorologist, instead I code things, so I inputted what you had said to ChatGPT to help me understand it some. That long conversation is here & probably has errors: tropicalatlantic.com/reconnaissan...

But the conversation isn't important. I was wondering what the result sounded like to add. I have to ask scientists when I add something like it: Part 1 of 3: "A dropsonde's reported winds represent roughly a five-second smoothed average derived from measurements taken about every 0.25 seconds."

Part 2 of 3: "Before transmission, the data is processed through a low-pass filter whose half-power amplitude occurs at a 5-second wavelength. This means that wind fluctuations occurring faster than about every five seconds are increasingly damped, reducing short-term noise and turbulence."

Part 3 of 3: "Near the beginning and end of the descent, the effective averaging period shortens as the filter tapers to zero. Earlier dropsondes used a 10-second filter." Thanks.

I thought the AI response was on the mark, except maybe the part about 20 samples near the beginning - I no longer remember exactly how the filter is coded. The 3-bullet summary you have here seems good to me.