Thanks very much for the recommendation of caution, Eric, when using ERA-40 for the daily climo normals. I will do so. I do see a similar day to day pattern of changes in anomalies. Regardless, I clearly see discrepancies in the anomalies when comparing the two datasets for a particular day. For example, consider the period 9/1/16-4/30/17:
whereas the ERA-40 base shows not even a single day colder than normal, the JRA-55 data shows there were about 28 days colder than normal. This actually suggests a couple of degree C WARM rather than cold bias for the ERA-40 based daily anomalies vs the anomalies shown in JRA-55 mainly in DJF. The other months look pretty close. May looks pretty close up until just the last 3 or so days as both were mainly 1 to 3 C colder than normal for May up til then. So, even your link's data suggests May has been chilly in the Arctic most of the month. However, there clearly is a difference now. So, I agree that caution is needed.
Aside: I had already noted earlier in this thread that I do have a concern about the 10 or so years having a disconnect of 7C+ between 12/31 of one year and 1/1 of the subsequent year as regards actuals. That still doesn't make sense.
whereas the ERA-40 base shows not even a single day colder than normal, the JRA-55 data shows there were about 28 days colder than normal. This actually suggests a couple of degree C WARM rather than cold bias for the ERA-40 based daily anomalies vs the anomalies shown in JRA-55 mainly in DJF. The other months look pretty close. May looks pretty close up until just the last 3 or so days as both were mainly 1 to 3 C colder than normal for May up til then. So, even your link's data suggests May has been chilly in the Arctic most of the month. However, there clearly is a difference now. So, I agree that caution is needed.
Aside: I had already noted earlier in this thread that I do have a concern about the 10 or so years having a disconnect of 7C+ between 12/31 of one year and 1/1 of the subsequent year as regards actuals. That still doesn't make sense.
Last edited: