Recent posts in the 2016/2017 freezing thread about centring the DMI 80N+ temp graph at winter time (rather than summer) prompted me to look back over the previous years.
Assuming the graphs display the full 365/366 days (and no days have been truncated) I noticed a large jump in the order of +16 K when you compare the last day of 1999 (~246 K) with the first day on the 2000 year graph (~262 K). Graph attached.
Niall, I can't speak to the graphs, but the raw data files don't show that jump. Here's what I have from the data files:
1999 360 246.118582 243.869741
1999 361 249.668921 243.743231
1999 362 256.79156 243.774908
1999 363 260.427556 244.006179
1999 364 258.812742 243.900104
1999 365 260.085622 243.804996
2000 1 261.983298 243.894301
2000 2 260.105925 244.023126
2000 3 258.758819 243.664103
2000 4 258.181293 243.20243