The preparation of a new global temperature dataset (HadCRUT4) is underway with the release a few weeks ago of CRUTEM4 – the University of East Anglia’s Climatic Research Unit’s updated land temperature data. The paper describing CRUTEM4 is here.
To compile the dataset the researchers looked at 5583 land stations and took data from 4842 of them. Much more data is included from the Arctic, especially the Russian Arctic.
The differences between the Northern and the Southern Hemisphere are apparent. Between 1901 – 2010 the NH warmed by 1.12 deg C and the SH by 0.84 deg C. CRUTEM4 also claims to have found an extra 0.1 deg C warming in the NH since 2001.
Fig 1 shows the differences between CRUTEM3 (red) and CRUTEM4 (blue) – click on the image to enlarge.
The data are yearly starting in 1975 and ending in 2010.
The record-breaking 1998 El Nino (not much effect in polar regions) is reduced in CRUTEM4 whilst the 2010 El Nino is not. Generally CRUTEM4 is a little cooler pre-2000 and, if anything, a little warmer post-2000, although that is marginal. Although the CRUTEM4 trend is many times that of CRUTEM3, there is no much indication of a statistically significant trend in the past decade.
Overall it is much as expected and doesn’t change much.
The UK Met Office derives a series of diagnostics from CRUTEM4. In one of them they calculate the global mean as the land area weighted average of the northern and southern hemisphere annual averages (2NH+SH)/3. This is ostensibly to prevent the data being dominated by the Northern Hemisphere which is more extensive and has better sampling than the Southern Hemisphere. It is finally adjusted by smoothing with a 21-point binomial filter. I expect we will see more of this graph although I consider it to be processed to much and not as useful as looking at separate hemisphere data, and using common sense!
Many consider that HadCRUT3 (combined land and sea) has underestimated the global warming trend over the past 15 years or so because it has, by chance it is claimed, sampled the areas of the Earth’s land surface that show more cooling meaning that the data was not representative of the ‘real’ situation. There is a somewhat less important effect concerning the change from ocean measurements made by ships and more recent ones made by buoys. Ships tend to be warm-biased.
We await the update of HadCRUT3 to HadCRUT4 which will incorporate the new HadSST3 (ocean) dataset with CRUTEM4. How it will affect the temperatures estimated for the 1940s and 1950s and the past 15 years will be eagerly awaited.
We must be careful though about adding apples to oranges. The incorporation of more polar data from the rapidly warming Arctic will affect estimates of warming trends. But this should not be automatically taken as an AGW trend because it is not clear that AGW is behind the warming seen in the Arctic – leaking of warm water from the Northern Atlantic is a possibility. It would be too cynical to blame a natural phenomenon on mankind.
Because of this HadCRUT3 will remain an important dataset in monitoring the temperature of the planet. After all the non-Arctic regions must increase in temperature due to the AGW effect and that will show up well in HadCRUT3. If HadCRUT3 continues to show no increase in temperature, as it has done over the past 15 years, while the Arctic warms, that will be a significant diagnostic about what is going on.