Click here for HadCRUT Data Rank Analysis (I)
Click here for HadCRUT Data Rank Analysis (II)
Click here for HadCRUT Data Rank Analysis (III)
Click here for HadCRUT Data Rank Analysis (IV)
Click here for Results of HadCRUT Data Rank Analysis (V)
This is the fourth posting in a series analyzing the information that can be obtained from the available HadCRUT data, recently updated to December 2007.
As in the previous blogs, the focus is on rank analysis, since it is widely claimed that global warming can be discerned by the fact that most of the warmest years have occurred very recently.
It is actually possible to obtain a rough indication on what is behind the recorded warming in the HadCRUT data by going one step below the usual globe-averaged, year-averaged figures.
(a) A strong hemispheric component is already visible in the yearly averages of the month-by-month ranks:
Note how for example SST/Southern-hemisphere is much more similar to Land/Southern-Hemisphere than to SST/Northern-hemisphere.
(b) Similar considerations apply at a seasonal level. See the graphs for the January-March period:
Obviously the Jan-Mar period is Southern Summer and Northern Winter. Let’s have a look at the Summer-to-Summer plots then:
I have computed the same graphs for all quarters, and for all seasons.
(c) It always looks more important to be in the same hemisphere, rather than in the same season or the same surface.
But visual inspection may be misleading, so a good round of correlations is in order (for the sake of clarity, the full list is at the end of this entry). These are the results:
(d) Correlation is highest intra-hemispherically (that is, when, say, the Northern Hemisphere’s land temperatures have placed near the top ranks, the NH sea-surface temperatures too have done the same) with a maximum of 98.6% (Southern Hemisphere, local Autumn) and a minimum of around 80% (Northern Hemisphere, local Winter).
(e) Same-season correlations are among the lowest, with a maximum of 74.5% (Spring) and a minimum of 68.8% (Summer).
(f) Among all the season-to-following-season correlations, the lowest values belong to the Oct_Dec-Jan_Mar periods (between 71% for Land, Northern Hemisphere and 80.5% for Land, Southern Emisphere).
(g) There is little, or perhaps even none, appreciable difference between Land and Sea-surface results
Conclusions and working hypotheses for the future will be discussed in next blog in the series.
SH SST/Land (V3)
NH SST/LAND (V3)
SST Seasonal NH/SH
Land Seasonal NH/SH
Jan_Mar : 75.95%