Measuring the Surface Air Temperature (Part 2)
Posted by jennifer, May 8th, 2008 - under Uncategorized.
Tags: Climate & Climate Change
In part 1 of this series ‘Measuring the Surface Air Temperature’ I wrote that James Hansen of NASA’s Goddard Space Institute explains that it is not easy to measure surface air temperature particularly in the presence of vegetation because the temperature above the vegetation may be very different from the temperature below the vegetation.
So, I thought, maybe it is easier to measure the surface temperature where there is not much vegetation, for example, at the Antarctic.
But apparently knowing the average temperatures on Antarctica has its own challenge including the sparseness of ground-based weather stations particularly in the continent’s high altitude interior and the harsh environment also takes its toll on equipment.
So NASA relies on satellites that measure energy radiated from the ice surface and estimate a level of uncertainty in these measurements between 2-3 degree Celsius (read more here).
Given this level of uncertainty I find it extraordinary that NASA can suggest a warming trend of a fraction of a degree over the last 20 years in the following image.
Bill Kininmonth, former head of Australia’s Bureau of Meteorology’s National Climate Centre, recently emailed me that because it is very difficult to assess surface temperature over ice surfaces using satellite radiometers it is more realistic to consider sea surface temperatures and to also exclude regions of seasonal sea ice.
In the same email he provided the following image of the sea surface temperatures.
Bill also explained that the Larsen B ice shelf at the Antarctic shattered rather than melted earlier this year, with the comment “shattering is not related to melting”.
creep; we are at cross purposes; let me express it philosphically;
Russell’s paradox concerns things which do not belong to any class but belong to a class of things which do not belong to any class. The paradox is, if it is a member of itself, it must not possess the defining property of the class, which is not to be a member of itself. If it is not a member of itself, it must possess the defining property of the class, and therefore must be a member of itself. This is a semantic issue resolved in reality by the actual quality of uniqueness. So, with temperature (and other climate paremeters); a base period purports to resolve the paradox of a trend of unique temperatures; but anomalies are still ‘coarse graining’ the reality of each temperature data. Any attempt to establish a trend on the basis of a base period is still compromised by dechoherence because any different base period will produce different anomalies (the Russell paradox). Can this be resolved by statistical significance? No, because any trend has a geographical algorithm as well as a time one. Given that the geographical algorithm is dependent on the time algorithm no level of certainty can mitigate the dechoherent element of the ‘trend’.
“a base period purports to resolve the paradox of a trend of unique temperatures” – it does?
Why is there a paradox. It’s only a simple relative manipulation. A different base will change the anomaly amplitude.
Geographic variation, processing technique or instrument variation (e.g. satellite vs thermometer) is fine – it’s a different methodology/technology. A different sampling of the same issue. A different set of data.
I’m not with you I’m afraid. I don’t see why this is that complex.
Temperature is defined as the thermal state of an object in thermal equilibrium. As the earth isn’t, it can’t have a temperature.
Oops!