Better data
I went back to the BTR records and interrogated the data labels, to get the calibration factors - that shows me what the values actually are - I think I was quoting data numbers before, rather than actual temperatures.
Anyway, I was wrong.
I now measure 12 K daytime and 8 K nighttime difference in the northern extended areas and for the southern areas something like 6 K daytime and 4 K nighttime.
It troubles me that the difference is less at night than by day. I would have expected the standout between cold nightime ground to be better than for warm daytime.
Any ideas?