Sorry I'm late to the party......
I am not an NTSB investigator, but here is my thought on this situation...
Ice build up typically causes loss of power & lift with some warning on prop driven people aircraft. If we apply this to drones, you should see a loss of lift and clearly the telemetry and the visuals do not indicate this but rather an increase in altitude. Again, the video feed also supports no loss of altitude. That said, I suspect the drone would try to maintain altitude as the props iced up until the motors could no longer keep up and then the drone would start to descend. I would expect slowly at first and then probably increasing in rate.
Now, why is the aircraft not responding to input? Nowhere do I see a connection status for video or control signals? Maybe this is due to some configuration settings, but unusual to me. Sat and battery I see. Why?
All times are listed as Youtube video and NOT flight time:
Now, at video(6:45) the pilot of the accident aircraft states "100M from home and quite high- struggling to see. Clearly it's not in a critical state of falling at this point. I assume that he REALLY can see the drone occasionally.
When the video is at 7:29, the pilot says"the flight stopped at 5:30" I assume a complete loss of connection or the drone is already on the ground.
Now at 7:38 the pilot states "it's at 6meters but 116 meters elevation" Now his comments are a bit confusing because he is speaking in post crash terms about the telemetry. He states that it crashed right in front of him(I assume 6 meters away +/-). I am thinking that these two events, the telemetry(6M out and 116M high) and the visible crash occurred in reverse order or more likely simultaneously. I don't see the drone damage as free-fall damage despite hitting the vehicle. Also, a calculated 2.5 M/S decent seems normal to me.
Another interesting issue is the ability to control some of the drone functions but not all. I believe that when in RTH, you still have full control over lateral direction but not up and I'm not sure about down. When I had an auto RTH due to low battery, the controls were very confusing and sluggish compared to non RTH. This could explain the control issue.
My theory is that a comm issue to the drone occurred possibly due to icing on the antennas. I'm not sure if DJI uses different antennas for different control inputs. RTH was started, but the telemetry was possibly wrong again due to antenna icing or some other comm issue.
At 6:45 it appears that he could see the drone high and 100M out. I assume that although struggling to see it he did in fact see it at this point.
45 seconds later when the flight time stopped, the telemetry showed 6 meters out. I believe that in those 45 seconds the drone descended at about 2.5 meters/second without the pilot noticing due to the numerous warning occupying his attention.
My final conclusion is controlled flight into terrain due to pilot error by not keeping an eye on the aircraft during the emergency/RTH decent. He attention was diverted to not only the warnings, but trying to make sense of the input issues. This is a classic pilot error issue. Many a pilot have flown into terrain while diagnosing a faulty indicator.
ALWAYS fly first!!!!! As they say... Aviate, Navigate and Communicate.