GHCN's Dodgy Adjustments In Iceland

Guest post by Paul Homewood

Take a look at the two graphs above. They are both mean temperature plots for Stykkisholmur, a small town in the west of Iceland.

Now take another look!

The top one is the official Icelandic Met Office (IMO) plot, and it clearly shows a gradual upward trend since the early 19thC, interrupted by cold periods between 1860 and 1880, and again in the 1960’s and 70’s. (The latter are famed in Iceland as the “Sea Ice Years”, when agriculture and fishing suffered so heavily that unemployment soared and the currency devalued by 50% for a full history, check here.) Current temperatures are about the same as they were in the 1930’s and 40’s. (The graph only runs to 2007, but temperatures have dropped since then).

The IMO’s narrative that appears with the graph states.

The time from 1925 onwards is dominated by a very large cycle that does not show an overall significant warming, although the temperature rise of the last 20 years is considerable.

The 20th century warm period that started in the 1920s ended very abruptly in 1965. It can be divided into three sub-periods, a very warm one to 1942, a colder interval during 1943 to 1952, but it was decisively warm during 1953 to 1964.

The cold period 1965 to 1995 also included a few sub-periods. The so called “sea ice years” 1965 to 1971, a slightly warmer period 1972 till 1978, a very cold interval during 1979 to 1986, but thereafter it became gradually warmer, the last cold year in the sequence being 1995. Since then it has been warm, the warmth culminating in 2002 to 2003. Generally the description above refers to the whole country, but there are slightly diverging details, depending on the source of the cold air.

Now take a look at the second graph. This is from GHCN. Ignore the bits on the left and check the graphs on the right. The top, red, plot is the actual temperature record, which follows the IMO trend. The second, yellow plot, however, is the adjusted GHCN record, which is the one actually used for global temperature calculation by both GHCN and also GISS. The bottom graph shows the value of the adjustment, with blue indicating adjusting down up to 1964 and red being upwards adjustment since. The scale is a bit unclear, but the overall effect of the adjustment is to add a warming adjustment of 0.74C.

You might ask “where did the Sea Ice Years go?”, and you would be right. GHCN have pretty much adjusted these out of existence. The GHCN algorithm has obviously mistaken the sharp fall in temperatures in 1965 as some sort of aberration caused by a change of location or equipment, and therefore wiped it from the record. However it has not made the same assumption about subsequent and equally sharp rises. The result is that temperatures since 1990 appear to be consistently higher than the 1930’s and 40’s, which, according to the IMO, simply is not the case.

Let’s stop for a moment, and review what GHCN have to say about why they adjust.

Surface weather stations are frequently subject to minor relocations throughout their history of operation. Observing stations may also undergo changes in instrumentation as measurement technology evolves. Furthermore, observing practices may vary through time, and the land use/land cover in the vicinity of an observing site can be altered by either natural or man-made causes. Any such modifications to the circumstances behind temperature measurements have the potential to alter a thermometer’s microclimate exposure characteristics or otherwise change the bias of measurements relative to those taken under previous circumstances. The manifestation of such changes is often an abrupt shift in the mean level of temperature readings that is unrelated to true climate variations and trends. Ultimately, these artifacts (also known as inhomogeneities) confound attempts to quantify climate variability and change because the magnitude of the artifact can be as large as or larger than the true background climate signal. The process of removing the impact of non-climatic changes in climate series is called homogenization, an essential but sometimes overlooked component of climate analysis.

So have there been any such changes in Stykkisholmur in 1965? Well, not according to Trausti Jonsson, a Senior Climatologist with the IMO, who tells me

“There were minor relocations at Stykkishólmur in May 1964, May 1966 and April 1968. None has been found important at the 0.2°C level. “

Furthermore, as Trausti makes clear on his blog Iceland Weather, the IMO has calculated series with some appropriate adjustments for each individual station. In other words, if there have been changes in location or recording methods, they have adjusted for these on a specific basis already. There is no need for GHCN to make further adjustments.

So, back to GHCN. To isolate the inhomogeneities they talk about, they compare a station with other stations, via a “Pairwise Homogeneity Algorithm”. Put simply, if Stykkisholmur shows a sudden drop in temperature while other nearby stations don’t, it suggests that the “drop” is due to local, non-climatic factors, and is therefore adjusted back to the trend of other stations. But is there any evidence that this is the case in Iceland?

There are altogether six Icelandic stations listed by GHCN that are still current: Reykjavik,Vestmanneyja, Akureyri, Keflavik, Hofn and Stykkisholmur. The IMO have produced the following plot for Reykjavik, Akureyri and Sykkisholmur.

Figure 2. 7-year running means of temperature at three locations in Iceland, Reykjavík (red trace)), Stykkishólmur (blue trace) og Akureyri (green trace). Kuldakast = cold period. The first of the marked periods was the coldest one in the north (Akureyri), the second one was the coldest in Reykjavík

So not only are the cold periods seen at Stykkisholmur repeated at Reykjavik and Akureyri, they were actually more pronounced at the latter. And what do GHCN say?

Just as with Stykkisholmur, the temperature has been adjusted down prior to 1965 and/or up since. And not only in Reykjavik and Akureyri. Every single site in Iceland has been adjusted in the same fashion. (You might also notice that in some cases the very warm period around 1940, that the IMO refers to above, has been adjusted down).

The “Pairwise Algorithm” is claimed to isolate non-climatic changes by comparison with other stations. But in Iceland this clearly has not happened. Every single station exhibits the same trend and at every one the algorithm has adjusted it out. There are no stations that the algorithm could possibly have used to have come to the conclusions that it did.

It is impossible to come to any other conclusion than that the software is hopelessly flawed.

How significant is all this? GHCN state that the latest set of adjustments have added 0.13C/century to global land temperatures, and of course this is on top of adjustments arising from earlier versions. This is a quarter of the reported warming across the globe since 1980.

But if the adjustments made in Iceland are patently false, can there be any confidence that adjustments made elsewhere are not also fatally flawed?

BTW – check out Reykjavik. Not only has GHCN added a massive adjustment, but GISS have actually made matters worse by adding to it when supposedly adjusting for UHI – see here.