News of Arctic warming nothing more than hot air

Published October 1, 2000

The rhetoric is heating up out there.

Item 1. Press Release, University of Colorado, Boulder, August 9, 2000, titled “Arctic temperatures warmest in last four centuries, study says.”

Arctic temperatures in the late 20th century . . . were the warmest in four centuries . . . [T]he changes appear to be at least partly a result of human activity, said . . . Mark Serreze, the paper’s principal author.

Item 2. From the paper itself, “Observed Evidence of Recent Change in the Northern High-Latitude Environment,” by M.C. Serreze et al.

Temperature records for 1990-1995 . . . place the results . . . into a longer-term perspective. It is apparent that our interpretation of temperature trends changes substantially if decades prior to 1970 are included. Annual mean temperatures fell during the period 1940-1970. . . . [I]t appears that annual temperatures from 1920-1940 rose even more markedly than during the post-1970s period.

Item 3: Figure 3, from Serreze et al. It is obvious that Arctic (latitude 55°N-85°N) temperatures are no warmer now than they were 70 years ago, and in fact, are cooler in the fall.

So where did the press release’s statement about “warmest temperatures in the last four centuries” come from? Simple. The temperatures they refer to are not annual temperatures: They’re reconstructed summer proxy temperatures, based on indirect indicators such as tree rings, pollen records, and so on.

Reconstructing the truth

How exactly do temperatures get “reconstructed”? By comparing the variations in, say, tree-ring width with observed temperatures, where there are concurrent records, and then “back-calculating” (or reconstructing) temperatures before such records exist. They don’t match up perfectly (or even very well, some might say). In general, tree rings explain about one-half of the observed temperature variation during the summer. That means some information is lost when scientists go from measured temperatures to “proxy” records.

Figure 5 in Serreze et al. is a compilation of Arctic temperatures from proxy records. That figure shows recent decades to be warmer than the previous 400 years, including the period more than seven decades ago that we know–from very real temperatures, not reconstructions–was as warm as today. The “proxy” record blew it, missing the warm period of the early 20th century.

Here’s the sad fact: You can call recent decades the warmest in 400 years only if you throw out the known, measured temperatures and substitute “proxy” records that are themselves based on an incomplete match to those temperatures.

Choose your timeframe carefully

Serreze’s work raises yet another question: When looking at the observed temperatures, why did Serreze and company begin their analyses in 1970? Very simple: Using all of the data back to 1900 paints a completely different picture.

We would like to say this is an isolated incident, but it isn’t: We found the same process in operation in a Science paper on Greenland’s glaciers (see “Is Greenland really melting?” page 1). There, the authors compared the last five years’ (1994-1999) temperatures to the average since 1979 and assumed that for each increment above that average there would be an increased melting. That’s all well and good, but in fact the overall temperature trend in Greenland since 1925 is negative, and the last five years aren’t particularly unusual.

So this latest report on Arctic temperatures isn’t news after all: just another repetition of the same old story that if you ignore inconvenient data, you can say anything you want about global warming.


Krabill, W., et al., 2000. Greenland Ice Sheet: High Elevation Balance and Peripheral Thinning. Science, 289, 228-230.

Serreze, M.C., et al., 2000. Observational Evidence of Recent Change in the Northern High-Latitude Environment. Climatic Change, 46, 159-207.