I had worked quite a few years in SIMOCO with HF communication and had a bit of exposure to the whimsical nature of HF links.
The most critical problem with HF is to find a frequency suitable for a particular range, which will work reliably throughout 24 hours of the day and there is also seasonal variation. It occurs due to the continuous changes in the electron density of the ionospheric layers. Again, in HF range of frequency, AM is used to preserve the spectrum for more users.
Therefore, HF links are prone to any electrical noise generated by switching events like ignition of starters in motor-bikes. Hence, HF links are very noisy. These are the reasons why HF communication has not picked up for data communication and the data rate is very low (generally about 300 bps).
For HF, the squelch circuit also does not work reliably because of swings in noise of modulated carrier. I had tried somewhat successfully to overcome the noise problem by using the RSSI signal from the IF stage, while interfacing to VHF network.
Another approach, which should work, could be to change the carrier cognitively.