Some radiosonde signals are successfully decoded at distances well in excess of the specified range for the radiosonde.
Experience reported here relates to Vaisala RS41-SGM radiosondes which are specified at 60mW max transmitter output and range up to 350km.
Free space model of the radio path
Whilst the path is over land, the radio path at high altitudes approaches a free space path, albeit with some ground reflection effects at the receiver.
The transmit antenna is probably close to a vertical dipole in performance, lets assume that it has a gain of 2.13dBi in the path direction.
Receive antennas will commonly be omni directional, and fairly low directivity, so lets just assume rx antenna system gain 0dBi.
If the path was free space, the rx field strength would be 13.8dBµV/m, tx power -115.5dBm. This model ignores ground reflection gain and losses, but it is in the ball park.
Let’s assume the rx system Noise Figure is 8dB and there is 1000K of ambient (ie external) noise. We can calculate the rx Signal/Noise ratio (SNR).
Above, SNR is 10dB. This is about 3dB above the experienced threshold of detection reported by radiosonde_auto_rx.
So the specified 350km range seems sane, and has a little margin for lower tx power and higher ambient noise.
RS41 is commonly received out to distances well over 500km
Is this just exploiting the margin calculated above, or could it be other propagation mechanisms?
One of the observed effects is a very slow variation in SNR, of the timing one might associate with Aircraft Enhancement (reflection of radio waves from aircraft).
Casual observation and comparison to flightradar24 maps of aircraft near the radio path showed some correlation.
Nowra to the East 10/12/2023
A recent flight from Nowra to the East provided an interesting case study.
Above is a 3D plot of SNR vs (lat,lon) for the latter part of the radiosonde flight for 13m starting at burst. We might expect that SNR would decrease from burst to loss of signal, and it does fall immediately after burst, but then increase significantly and remains fairly high for around 9m… even though it is descending rapidly.
Above is an approximate map of the radio path and a large jet (A320) passing through.
Wagga to the East 13/12/2023
Above is a chart of packet counts from VK1OD of a radiosonde flight from Wagga to the East, 250-280km path length. Note the burst nature of packets, propagation is not consistent and occurs only when suitable aircraft (flying the main Sydney – Melboure corridor) support AE propagation.
Conclusions
This one experiment reconciles with several informal observations, and possibly explains variations in SNR that would not be otherwise expected may be due to Aircraft Enhancement.
Aircraft Enhancement might also explain unexpected long paths that of durations suggestive of Aircraft Enhancement rather than Sporadic E or Tropospheric Ducting.