I received a note from a reader of this web-log who was a bit cranky with my post advertising the current review by Warwick Hughes, see post and thread here.
The really relevant piece of information from the long email was perhaps this graph:
View image (about 80 kbs).
Comment included:
The temperature trend maps on the BoM’s website alone attest to the fact that UHI [Urban Heat Islands] have nothing to do with the warming as greatest warming has occurred in areas with the lowest population densities (the subtropical arid zones – which just so happen to be those which are predicted to warm most rapidly under global warming).
Anyway, if the above isn’t enough, Sea Surface Temperature information has been added to the BoM’s website at http://www.bom.gov.au/cgi-bin/silo/reg/cli_chg/trendmaps.cgi . You will see that there is little to no difference to the rate of warming of land and over the oceans.
Jennifer Marohasy BSc PhD is a critical thinker with expertise in the scientific method.
I don’t know whether to laugh or cry about the graph and the comment.
1. I know that the graph comes from within the Bureau of Meteorology, that organisation which is trying to take over the CSIRO’s climate model. I have serious reservations about the validity of CSIRO’s work and this information from the BoM has given me more even cause for concern.
2. Okay, it’s an average sea surface temperature (SST) but how was it calculated? Australia’s west is influenced by the Indian Ocean, the south by winds from the Southern Ocean and the east by the Pacific Ocean, so on what measurements is the average SST based and how does it balance the temperatures of the three oceans and the Timor and Arafura Seas?
3. You might also like to ask about the integrity of the sea surface data. Did we really have accurate and wide-spread measurements of SSTs in 1950, let alone 1900? I very much doubt it.
4. Take the bigger picture – how many people honestly expeted that the land would warm slightly but not the oceans? You might say that it stands to reason that warmer air or an increase in solar emissions wouldn’t distinguish between land and sea.
5. On the other hand, earlier this year someone at Monash Uni who has links to CSIRO climate modelling told me quite forcibly “Amazingly, the climate of the northern hemisphere and the southern hemisphere are somewhat different. The land mass split is about 70:30 in favour of the northern hemisphere, which means that the hemispheres will respond differently to different forcing.” He seems to be saying that GW theory says that land and sea temperatures should be different, but if that’s what the models say and this BoM data contradicts this then doesn’t that cast doubt on the model that the BoM is trying to get its hand on?
6. Now look at the graph in detail. The temperature anomalies also don’t always go in the same direction in any year, let alone vary equally or even in ratio with each other. The graph has a total range of 2 degrees and I can see about 7 instances where the difference between sea and land temperatures exceeded 0.6 degrees, about one-third of that range. In the early 1980’s land temperatures increased (probably due to something called El Nino) but average SSTs scarcely changed. (Given that the Pacific off eastern Australia tends to warm during El Nino events maybe the averaging from the other oceans hides the impact.) Also look at 1992, SSTs temperatures fell back to 0 while land temperatures climbed to 0.7. Same driver of temperature? Hardly!
Put it all together and the graph tells us very little at all, save perhaps that in a general sense land and sea have probably been warming (unless of course the warming in the surrounding oceans is is very localised and averaging has distorted it). If we assume that both have warmed then the inconsistency in temperature changes suggests that certain “background” warming has occurred and something else is happening to cause that short-term inconsistency. In neither case case does teh graph provide any clear indication of a driver.
As for the acompanying comment about UHI being disproved by the greatest warming occuring in subtropical arid zones, I really have to question the intelligence of these people. (Just ignore for a moment the fact that trends can prove warming, cooling or no change depending on the year one starts with and that one or two abnormal years late in the trend period can really skew the result.) According to the BoM website the greatest has ocurred along the NSW-Queensland border. Wow, that’s amazing! It coincides incredibly well with a region that is greatly impacted by El Nino conditions and even more amazing, the recent warming there has occurred during a series of El Nino years.
As I said, I don’t know whether to laugh or cry if this is the best the BoM can do.
Temperature is, in itself, an intensive variable. It cannot, ever, form the basis of any statistically valid, real, computation.
Period.
However as a number, in isolation, temperature may be manipulated into all sorts of results.
“Temperature is, in itself, an intensive variable. …”
What’s an example of a “non”-intensive variable then.
“It cannot, ever, form the basis of any statistically valid, real, computation.” … why not ?
Just asking …
Intensive and Extensive variables – they’re an old way of classifing quantities – basically intensive variables are ones that are independant of the size of the system it describes ie density while extensive variables are size dependent ie volume. Intensive varibles are the product of dividing two extensive variables. A lot of scientific equations include both intensive and extenive variables ie PV=nRT.
I’ve argued with Louis before on this topic and refuse to engage him again – Louis for some unknown reason refuses to accept that intensive variables can be ‘statistically’ valid or averaged.
If intensive variables weren’t able to be averaged then the mining companies won’t be able to determine the value of ore bodies ie a body of ore is determined to be 50000 m3 – you take say 50 ore samples from the ore body, determine the concentration (an intensive variable) of whatever in each, average the concentrations then multiply the average by the volume and hey presto you’ve got how much metal is present.