The course has dried out, and I did a complete remeasurement on Sunday
For 7k reference point to 8k reference point:
when flooded 968.4m
when dry 966.9m
Increase when 200m under water = 1.5m
So it looks as if the constant increases when under water by approx 1.5/200= 0.75%
However caution is required since there could easily be errors in lining up with the reference points at 7k ( a tree trunk 5m away across the road) and at 8k (an electric pole about 2m away). In fact I could imagine it possible that allowing for worst case errors would give the increase as 1.5m +/- 1.5m. So my measurement is not accurate enough to be really useful, except to say that a 200m stretch of flood wont impact on a measurement of a 10k very significantly- it is small enough to be accommodated within the SCPF, if there were no other problems on the measurement.
Of course I would really needed two nail-marked reference points either side of a flood in order to make a more accurate estimate. At various times in the last 7 weeks there has been a 20m long flood on my cal course at home. I did consider going out in the rain, nailing it, and trying to compare cals on that 20m with something else not under water, but I decided that 20m would be too short to get an sufficiently accurate calibration.
All the roads have pretty well dried out now, but now knowing what local areas are prone to flooding in extreme conditions, I suppose I could prepare by marking out with nails a cal course which is at risk of flooding and a close by course which would always be dry. I would not need the exact lengths, just the ratio of cal counts when both are dry and then I would just have to wait for another flood. My only worry is that they say this has been the wettest winter at least since 1995 and possibly even since records began. I could have a long wait.