A little less than a month ago I posted about the significant differences between the second version (ERSST.v2) and third version (ERSST.v3) of the Southern Ocean SST data sets as presented in the Extended Reconstructed Sea Surface Temperature by Smith and Reynolds.
In this post I’ll illustrate the difference between the ERSST.v2 and ERSST.v3 versions for Arctic [60 to 90N] SST anomalies, and illustrate that the significant rise in the ERSST.v3 data after 1998 results from a natural event, the 1997/98 El Nino.
The source of ERSST.v3 data sets breaks them down into different latitude bands (13). Comparisons between other ERSST.v2 and ERSST.v3 data should hold no more surprises. But I will post on the others shortly to illustrate the differences.
ERSST.v2 AND ERSST.v3 VERSIONS OF ARCTIC OCEAN SST ANOMALIES
Figure 1 illustrates the ERSST.v2 and ERSST.v3 versions of the Arctic Ocean [60-90N] SST anomalies. The ERSST.v2 is illustrated from January 1854 to August 2008. The ERSST.v3 data begins in 1880 and ends in April 2008. (No updates have been issued since April 2008.) The data in Figure 1 have been smoothed with 12-Month running-average filters. There are some minor differences in the data prior to the 1990s, with greater polar amplification and indications of better sampling in the ERSST.v3 data, but after the 1990s, there is s significant divergence between the two data sets. The newer data rises sharply.
ARCTIC vs NORTHERN HEMISPHERE vs GLOBAL SST ANOMALIES
To help illustrate the reason for the significant rise in Arctic SST anomalies, I’ve compared the Arctic SST anomalies with Northern Hemisphere and Global SST anomalies in Figure 2. It covers the period of January 1978 to April 2008, and in it, the data is raw. Some of the instability in the Arctic signal results from Polar Amplification, some of it is from the Arctic Oscillation (AO), and some of it is from the fact that the Arctic is smaller in area than the other two data sets. Greater geographic areas tend to smooth data.
The Arctic data diverges from the Global and Northern Hemisphere data almost immediately after the 1997/98 El Nino. Studies show that the Arctic takes as much as 13 months to react to the additional heat released by El Nino events. The “El Nino of the Century” is obviously no exception.
The timing of the Arctic SST anomaly divergence appears to clearly indicate that the current run of elevated Arctic SST has a natural, not anthropogenic, basis.
Smoothing the data with 12-months filters, helps illustrate the Arctic SST response to the 1997/98 El Nino. Refer to Figure 3. The divergence comes approximately one year after the 1998 peaks in the Northern Hemisphere and Global SST anomalies.
By looking at the prior divergences of the Arctic SST anomalies, it could be argued that the Arctic appears to have begun its climb prior to the 1997/98 El Nino. Extending the data helps put the current Arctic SST anomalies and the starting point of their current warming into perspective. Refer to Figure 4.
Polar amplification is present in the higher trend in the Arctic SST during the rise from 1920 to 1940, during the decrease from the 40s to the 70s, and again during in the rise from the 70s to present. The Arctic SST anomalies didn’t skyrocket until immediately after the 1997/98 El Nino.
Smith and Reynolds Extended Reconstructed SST (ERSST.v2) is available through the NOAA National Operational Model Archive & Distribution System (NOMADS).
The more recent version of the Smith and Reynolds Extended Reconstructed SST (ERSST.v3), along with land surface temperature and combined (land + ocean) surface temperatures, are available in various latitudinal bands at:
Don’t let the PDO in the address confuse you. There’s much more there. The overview for the update is here: