As expected, NOAA global anomaly reduced in January, from 1.12°C to 1.04°C (report here). It was expected because grid methods like theirs give lower weighting to the Arctic, which was very warm, and so relatively more weight to Eurasia, much of which was cold. But it has been a month which emphasises these divergences. I'll show the recent months plot from the results page:
The NCEP/NCAR index pretty much follows GISS, upward. The satellite indices rise in unison. NOAA follows TempLS mesh down, with TempLS grid dropping further. Normally I would expect TempLS mesh to go with GISS, and grid with NOAA. The difference is there, but with TempLS generally dropping more. I'm not sure why.