High (latitude) anxiety

Published August 1, 2000

For the second time in the past four years, Polish climatologist Rajmund Przybylak has a put a chill on climate model projections of global warming temperature patterns.

Fellow global warming “skeptics” (we prefer “realists”) no doubt recall Przybylak’s 1997 paper, in which he found that the high-latitude warming all climate models so happily generate was in fact not evident in the temperature records.

In his most recent foray into high northern latitude records, just published in the International Journal of Climatology, Przybylak makes a strong case that the models are doing a lousy job of predicting the observed patterns of temperature change. He further challenges the prevailing paradigms about how the globe should warm. All this in an easily readable 29 pages.

First, Przybylak carefully identifies the “Arctic” region based on climatological rather than astronomical definitions. He notes that the inclusion of non-Arctic data is a major problem that can induce spurious trends from locations that have subarctic or even midlatitude climates. Using his definition (which he in turn derived from two other sources), the true “climatic” Arctic actually extends as far south as 55° N latitude, in the vicinity of Baffin Bay in northeastern North America; but only as far south as 77° N over the North Atlantic.

Next, he proposes that the existing model-derived notion that most of the warming will be over the poles and in winter is simply wrong. For example, as we include more and more lower-latitude data, the pattern of warming observed in the Arctic becomes unrecognizable. Over the highest latitudes, temperatures were highest in the 1930s, whereas current mean temperatures are near the minimum for the available period of record.

Przybylak was able to relate a significant portion of the 1930s warming to changes in atmospheric wind patterns “probably connected with natural climatic forcings.”

Further, Przybylak compares data from his 37 available Arctic and seven sub-Arctic stations with the IPCC gridded data set, which contains 30 useable 5° latitude by 5° longitude grid boxes over the same region. When temperature anomalies for winter (December, January, and February) are compared between the stations and grid boxes, markedly different patterns emerge. Since the correlation between the two records is only 0.55, that means 70 percent of the temperature variability in the grid record cannot be accounted for in the station record. That is disconcerting, since the grid cells are presumably based upon the best available station temperature records.

While the grid cell record shows a general warming tendency during the 1990s, the station record shows significant cooling since the mid-1980s, to the extent that the five-year running mean temperature is lower now that at any time since the beginning of quality station data availability in 1950.

The implication here is that some errors have crept into at least one of the two records that are producing major temperature biases in the high Arctic. Indeed, the author cautions that

“the quality of [the grid box] data in its present state is significantly lower than the station data used in this work. In contrast to grid-box data, the temperature series from stations have no gaps. Taking these factors into account, identification of characteristics of Arctic temperature variations should still be based on stations’ data.”

Przybylak’s analysis of temperatures, both annually and seasonally since 1950, reveals similar surprises. Overall, temperatures were highest in the 1950s, lowest in the 1960s, and exhibit no trend since 1970–despite the rapid increase in human-induced greenhouse gases that allegedly will produce the most obvious warming, and thus be most easily detectable, in the Arctic.

In summarizing his temperature analyses, Przybylak states:

“In the Arctic, the areal average annual temperature [from 1991 to 1995] was higher than normal . . . by only 0.1° C. Analogical values for the Northern Hemisphere, the Southern Hemisphere, and the whole globe were equal to 0.3° C, 0.2° C, and 0.26° C, respectively. . . . However, according to climatic models, the doubling of CO2 in the atmosphere should have caused two to three times greater warming in the Polar Regions than the global average.”

Furthermore, after finding that most of the Arctic warming has occurred in spring and summer, he points out that “This seasonal pattern in air temperature change is not consistent with climate model outputs, which suggests that an enhanced greenhouse effect will result in the greatest warming occurring in winter and autumn.”

So why aren’t Arctic temperatures behaving the way the world’s greatest climate modeling minds say they should? Essentially, Przybylak has demonstrated that, since 1975, the real Arctic hasn’t heated up a lick. He proposes two theories:

1) With all the water, sea ice, and land ice in this region, it takes an awful lot of energy to warm things up. Conventional wisdom suggests greenhouse warming will be evident and most obvious in the very high latitudes first. Przybylak proposes it will take a lot more energy to warm the Arctic than a similarly sized region in the tropics, for example.

2) Natural factors, mainly changes in the circulation of the atmosphere, will reduce or completely dwarf the greenhouse warming. As support, Przybylak notes that an Arctic circulation shift that occurred in the mid-1970s corresponds with the lack of warming over the region. Other natural factors include changes in solar activity.

If these results are correct, they could have a profound impact on our current understanding of the role of the high latitudes in global climate change. To take one example, a predominant winter warming will reduce the extent of high-latitude snow cover. The longer snow- and ice-free period each year results in more radiation being absorbed by the Arctic Ocean because less of that highly reflective frozen water is hanging around the surface. The onset of winter freezing will likewise be delayed. That type of positive feedback (in which greenhouse gases ultimately, through a series of complex interactions, indirectly raise surface temperatures) is the basis for most climate models.

But Przybylak’s results suggest that this positive feedback, if it exists, has only a minor influence on Arctic temperatures.

In short, yet another paper, based on observations rather than model-derived projections, has called into question several of the fundamental premises surrounding the global warming debate–premises that serve as the basis of all the doom-and-gloom scenarios. Without those models, global warming apocalyptics are left hanging in the deep freeze.


Przybylak, R., 2000. Temporal and spatial variation of surface air temperature over the period of instrumental observations in the Arctic. International Journal of Climatology, 20, 587-614.