G'day Chris,
Hmm, the el Nino 1998 being "weird". Not a word I would use. And 2005 being a record, huh. Sorry. Don't agree at all. What dataset are you looking at. Going by the 1979 to date Surface Air Temperature and low altitude air temperature averages from the two satellites tasked to gather such data, 1998 was truly a very large el Nino affected year, but 2005 was not even close to being a record. It too had significant el Nino effects aiding any increase.
So we have 1998, probably a one in a century event because of factors that contributed to making the el Nino the largest effect it could be on climate for the year. That sort of data, in datasets that are not extensive enough are normally excluded to prevent distortion due to a one off. If I was studying the breeding cycles of tree frogs and had ten years of data but in one of the years an outbreak of some disease killed most of the rampaging cats in the area where the study was being done, I would have to either exclude that year, wait another twenty years or so until my dataset was actually of a reasonable period or mark up the data summaries very heavily with a warning concerning that particular year. It might not be the best of analogies but since it happens to be one that actually happened, and my brain isn't functioning more than it's 10%, it will have to do.
I do not believe that 1998 should be included in any datasets when arguing average world temperatures simply because the datasets are not extensive enough to show any cycles of this type of rare occurence. It is somewhat like those stupid statistics for dam constructions where they indicate it will survive anything but a 1:10,000 year flood. How could anyone have any idea what the rainfall pattern or amount is over a period of 10,000 years since in most cases of dams the records go back perhaps 50 years and even then usually not for the exact area of the dam.
In a dataset that goes from 1979 to 2007, even a 30 year cycle event may not show up, let along a 100 year cycle, and of course climate does not just have cyclical events but random events that have average frequencies between events only. A one in a hundred year event, for climate, does not mean that it happens once every hundred years, only that it happens, on average, one time for a hundred years where it doesn't happen. That doesn't mean that it doesn't actually happen three times in ten years and then none at all for several hundred years.
The AVERAGE for this Ice Age for Interglacial periods is that they will only last for 10% of the time for the Ice Age and should be ON AVERAGE 5,000 years or less long before again reverting to a glacial period. Using those figures this glaciation is 6,000 years overdue to return to an Interglacial Period and thus the odds are far higher that the Interglacial period will end in the short term than not. That is not true of course, just as the odds of throwing a heads is the same after thowing ten heads in a row or after thowing ten tails. Human brains are not wired to recognise this as true because pattern recognition is so hardwired into our head. Does not make it incorrect though, just difficult to grasp.
By the way even if you INCLUDE 1998 in the satellite data you still do not get any significant warming trend and extruding it you actually get a cooling trend, although too slight to argue anything.
The surface air temperature data from ground stations does indeed show a large increase in average temperatures from about 1975. Suspicion should fall on any data where the most accurate method of collecting the data, seriously contradicts that data, yet the SAT ground collection data seems to be used as if it is accurate to several decimal places. Trouble is the data is grossly inaccurate, the collection system is seriously flawed, the averaging system is whatever the relevant authority for an authority decides it to be and can and has been changed over very wide areas on a whim. Measuring devices are changed. The positioning of the equipment itself is changed. Minor little things such as urban heat effects are ignored or somehow are written off as too minor to be worthy of adjustment (although an error in the satellite system that could have meant the averages were out in the order of 10,000th of a percent have been argued by a great many as reason enough to consider the satellite data to be fundamentally flawed).
So if we use the North American climate dataset for continental US, we find that not only has the rather more than insignificant chunk of the earth has not warmed over the past century, it has actually cooled a bit. It has not warmed significantly from 1975 on.
If we us the British Admiratly records of Surface Air Tempertures at sea over the last 200 years we also find that no significant warming trend has occurred and that only a pretty minor warming trend is shown from 1975.
The satellite data from 1979 shows no warming with the 1998 data, slight cooling without it.
Oh and Chris, the dataset you used is partly modelled and very much based on assumed data and assumed corrections to the data that is available. It takes a long time to do it but have a look at the raw data without corrections for things that are assumed to need correction for, as I have actually done, and you don't end up with a graph even remotely like the one you posted.
When looking at AVERAGE world's temperatures, the field becomes immensely complicated. It shouldn't be because we have a pretty good dataset from 1979 onwards. It just gets ignored because it does not provide the nice rising graph that seems to be wanted. Most people would agree that you really should compare apples with apples. That data used should be sourced the same way, using the same methods, the same calibration to equipment, the same positioning if it has some importance, the same mathematics to summarise the data, etc, etc.
So you can rely on the UK Admiratly Figures for water temperatures for the last 200 years becaus the method has not changed much. It cannot ever be fully comparable because the measuring tools have been updated and not calibrated against the equipment being retired or against some standard. You also have the problems that the ships were not in the same position each year and that some years concentrations occurred in say latitudes or particular oceans etc. And the problem of increased ability to travel in difficult sea areas such as areas subject to severe weather or very cold areas or areas with icebergs also disturbs the distribution of collection points. However, at least some of the methodology was consistent.
You can rely on the US weather collection for small towns of less than 5,000 population where the weather stations have not moved over the last 100 years if you want relatively comparable figures without the problem of massive changes because of heat islands etc. But this is a very small dataset indeed because there are very few stations even in the US where such things as wars and civil uprest have not had an impact.
You can rely on Australian weather collection in the same categories as that above and pretty much for the same reasons and with the added bonus that weather stations have been moved less often than their US cousins.
NZ weather data is even more comparable.
As mentioned before you have the world's satellite data for not only temperature but also for ice coverage, at least two dimensionally, from 1979.
And that's about all.
It is amazing just how non comparable weather data collected around the world for whatever reason has been. The US may have at least the plus of not having been interupted by wars or civil unrest and the like. But there has been seven changes in the times of temperature collection in the past century, to give just one example of the hundreds of reasons why the data is just not camparable. You just cannot rely upon a figure in the records that says Springfield had a temperature of 38.7 degrees in 19 July 1919, a record high. That's a bit like saying your body temperature is exactly 98.4 degrees because that is what it should be when not sick. I've been in enough hospitals to know that variations as much as half a degree out can occur simply because of different themometers used. And that excludes different methods of collection. If one side of my mouth can give off a temperature reading 0.3 degrees different to the other side using the same themometer and being only several seconds apart in the measurements then just how far out could readings be when measuring devices are moved 100 metres up hill, or from the shade at the side of a building to full sunlight.
And any argument that is along the lines that all these variations cancel out really hasn't thought the comment through.
If the prior data is that unreliable, I find real problems with comments such as "studies show" and then suggesting that something might happen to the temperature in 2009 or at any other time because that is not a study at all but a guess based on a model. Compounding the problem is when such comments are linked to what CO2 might do in the future and how this is somehow linked to the future temperature.
If the link between the small fraction of CO2 in the atmosphere varying and temperature varying because of it is in any way true, please point out any study that demonstrates the link in any way at all.
This is a particularly difficult area in relation to climate change, but actually the most important aspect. To argue that the world is warming actually requires that one knows what the world's temperature is and was. If that is not known then the change must be very significant indeed so that anecdotal evidence is overwhelming, or you just cannot establish warming (or cooling for that matter) at all.
Regards
Richard
