Actually it isn't. That's USHCN data, so stns are the same. That dataset doesn't flip stns around. If data is missing, they estimate for the month. Secondly, I've seen a comparison of aftrn n morning stns for max temps, and the pattern is very similar. Sorry I don't have those graphs.
The issue we are seeing now isn't max temps so much as higher mins, and less winter cold.