I saw this one on "Project Stats" :

http://climateapps2.oucs.ox.ac.uk/cpdnboinc/result.php?field=Temperature&resultid=30365&phase=AT#graph

Looks wrong somehow.

I had one like that as well. Interesting that it had 77 trickles. It may have something to do with that. Looks like it backtracked in the first phase and did the 54010 trickle 4 times. I'm not sure how that would affect an end of model graph though.