Rad Alt - Bendix King KRA 405
Normally pretty accurate (<1m ≤ 200m Rad Alt).
The issue of course is the density or variation of the trees below you. Would you like it to indicate the height of each tree you have just passed over? The antennas are on the tailboom after all, it doesn't look forward. If the trees were ~ 100' each I guess as the indication would be wildly fluctuating you would assume that you were over trees! Having used them for geophysical work and from memory at a cover of >50% they would read the distance to the tree tops. That information was "smoothed" in the indicator to the pilot. The raw data which was recorded was all over the place and was smoothed in the data post processing. As opposed to your requirement we actually wanted to know where the ground was so estimated the tree height and flew at a lower Rad Alt accordingly until passing the trees. There is obviously a damping algorithm but who knows what it is. I would guess that it is probably the same as your old Bell was. Sounds like you are pretty low on NVG's in uncharted territory and relying on data that can be "incorrect". Rather you than me.
If you have some time on your hands you may be abe to work it out from here -
TSO C87
As for the EPC - see the previous. Only so many kept "on file". An EPC trend is required if you operate "Cat A" - see RFM Supplement. Be aware EPC data, ECD Ground Power Check, and engine manufacturers data/method may vary. You cannot mix the 3. ECD EPC data is acquired at 2 x 60% Q at S & L. ECD Ground Power Check done using airframe data and "eyeball" with one engine in idle. Engine Manufacturer check normally on the ground with one engine in idle and acquired by ground based software using EEC parameters. See RFM Section 5.