It would be useful to have anomaly data as well as absolute data, for example temperature anomaly versus the climatological normal for a given day. I know that the GFS produces a 2m temperature anomaly and both GFS and ECMWF produce a 850mb temperature anomaly.
