is doppler effect important?

hi
in wireless communication when T > T_D Doppler effect leads fast fading, where T is symbol duration and coherence time is T_c = 1/f_D.
f_D has the maximum value of 300 Hz, so T_c has the minimum of 3 ms. so Doppler phenomena only affects systems with symbol rate below 300 sym/sec.
thus in real wireless communication Doppler effect does not exist.
if above assumptions are true, we can only consider frequency selectivity effect of the channel and model the channel by matrix H.
so why MATLAB use JAKES model (considering Doppler effect) in rayleighchan function?

First, Doppler frequency can exceed 300 Hz, depending on your platform speeds and the carrier frequency. For example, at a 5 GHz carrier, the Doppler frequency on a high-speed train (~ 70 m/s) can exceed 1 kHz.

Keep in mind that coherence time is really just a rule-of-thumb based on a very specific model. The Rappaport text, for example, lists several formulas for relating Doppler frequency to coherence time. In any case, the channel is continuously changing, so how much change you can tolerate really depends on your configuration. For example 256-QAM tolerates less Doppler than QPSK for a fixed SNR.

Also, even if the channel doesn't change appreciably from symbol to symbol, it may change enough over time to degrade links. For example, if you are using coherent demodulation (non-differential QAM constellations), then coherence time gives you some idea for how often the channel estimate should be acquired. If you are simulating your link with channel estimation, you would want to use the rayleighchan function to see how this would degrade performance.