Have you tried putting a puff of airborne particles (e.g., smoke) through the system? I suppose it would only yield a minimum estimate of airspeed but it would be a good sanity check on your other measurements.
Well, for sure the velocity is too high at 6.38 m/s, since that would result in overunity (roughly COP 2) with an ohmic heater. The velocity does drop a little bit with increasing temperature based on a quick test last night. But it does not drop enough.
One possibility is there are jets of high velocity in the air stream that spin the anemometer blades much faster than the average velocity would. So a longer tube or some diffuser may be required to allow the air flow to even out.