Hi,
Modern "intelligent" radio chips are so sophisticated (but cheap) with capabilities for data bufferering, FSK, channel hopping and even spread spectrum, etc. that the transmiisions may be very difficult to "hack", even if the designer didn't intentionally aim to encrypt the transmissions. So I don't think I'd try, even if the Netatmo were a "useful" Wind Gauge.
But IMHO a modular (wireless linked) wind gauge which needs 4 AA cells to be changed every year is not a "serious" sensor. Even the Fine Offset stations have a few metres of cable to their wind sensors, so the "battery box" (transmitter) can be more conveniently located, and it needs only two AA cellls to be replaced every few years. Also, the "active" (ultrasonic transmission) method must consume much more power than a simple "counter", so I don't see how it could detect wind gusts, but must just take a "snapshot" every few (five ?) minutes. For a wind sensor like the Netatmo, solar recharging seems the "obvious" choice.
However, the "ultrasonic time of flight" is an excellent method for measuring wind speed/direction and is not too difficult for a hobbyist to implement, using a microcontroller. So a "tear down" of the Metatmo might be rather interesting, but you'd likely end up with some knowledge, but a non-functioning sensor.
I did actually develop a microcontroller-based ultrasonic time of flight air speed (not direction) sensor about 30 years ago. Now, very low cost ultrasonic "rangefinders" are available for "Arduinos", but I don't know if they would be sufficiently weatherproof (or accurate). So I would probably start with the type of ultrasonic transducer used in automobile reversing sensors.
Normally, the "flight path" would be some tens of cms to give an adequate delay/variation, but an alernative method (which I used) is to compare the phase of the received to transmitted signal (which can be done easily with a flip-flop, low pass filter and an ADC). Basically, three "universal" transducers (i.e. able to transmit or receive) on an equilateral triangle could produce six path measurements (including "forwards" and "backwards") to enable calculation of two airspeed vectors (at right angles) and eliminate variations of the speed of sound (due to temperature and humidity, etc.).
Maybe something I'll try eventually, but I really want to get my sunshine sensor project "finished" first.
Cheers, Alan.