TCS Daily

Is the Arctic Melting?

By Willie Soon - March 24, 2003 12:00 AM

Reliable information for the public about changes in the Arctic's climate is hard to come by. Different newspaper and media reports, even when quoting the same scientific studies, can tell different stories. Worse, reports about the state of the Arctic give contradictory pictures of what has actually happened to the Arctic's sea ice, ocean, air pressure and temperature.

Such reporting - including the claim that the positive swing of the normal Arctic atmospheric oscillation may have intensified during the 1990s - creates the notion that something fundamental has changed, when there is little evidence to that effect.

The Arctic has often been used by advocates of the hypothesis of anthropogenic - man-made - climate change as a poster child for human-induced global warming. So, an accurate assessment on the state of the Arctic climate variation is important.

The aim of this article is to eliminate some of the confusion about such news media sensations by doing three things:

  1. Provide the most up-to-date and reliable information about Arctic climate, such as air temperature, for the past 125 years as a benchmark in this Tech Central Station CHARTiFACT forum.
  2. Describe the actual observed changes in temperature and explain some common pitfalls in drawing sweeping conclusions based on either short or specific regional records. 
  3. Discuss the implications of the observed temperature trends in terms of the often imprecisely argued claim that carbon dioxide- (CO2 -) induced global warming is amplifying warming in the polar region.

The Basic Data

Figure 1: Arctic-wide temperature anomalies (in °C) from 1875-2001 relative to the mean of 1961-1990 interval, with the number of stations producing the temperature set in each decade. (Courtesy of Igor Polyakov of IARC at the University of Alaska)

Figure 1 shows the annual time series of the Arctic surface air temperature from 1875 to 2001 as it was recently reconstructed by Igor Polyakov and colleagues at the International Arctic Research Center (IARC) in Fairbanks, Alaska and the Arctic and Antarctic Research Institute in St. Petersburg, Russia.

The sources of this new temperature record include measurements from land stations, floating buoys on the ocean and even drifting stations on sea ice. Detailed documentations of the methodology and spatial sampling strategy had been published in papers that appear in Geophysical Research Letters, Journal of Climate and the American Geophysical Union's EOS.

Figure 2: Distribution of surface air temperature stations on land, ocean or sea-ice for the composite Arctic-wide temperature record in Figure 1. (Courtesy of Igor Polyakov of IARC at the University of Alaska)

Figure 2 shows you all the locations poleward of about 62°N (with the Arctic circle defined as the zonal ring around 66°N) where the air temperatures are sampled to produce the Arctic-wide temperature history shown in Figure 1.

What's Happening?

So what do we see in Figure 1?

First note that the maximum annual Arctic-wide temperature anomaly - the difference from the mean temperature for 1961-90 as plotted by the blue dash line - reached a maximum of 1.7°Celsius in 1938. That compares with a maximum of 1.5°C in 2000.

Next, notice the blue solid curvy line. It gives a 6-year running average of the annual temperature anomalies plotted as a dotted blue line. This line helps focus on the climatic changes of longer time-scales, instead of year-to-year weather "noise" in the dash-line.

Now, for a more interesting part: Just for the sake of discussion, contrast two views of the record. Compare the red curve that was drawn by a straight line from 1875 to 2001 versus the four green lines drawn over four intervals in Figure 1.

The red curve describes the longest-term temperature variation resolvable in this Arctic record, and it shows a change over the period of about 1°C per century.

What does the trend mean? Some people take it and argue, "See, the Arctic climate is warming; it's warmer today than 125 years ago. As CO2 from the burning of fossil fuels has increased during that period, that likely has contributed to that warming."

But there's another way to look at the record than a relatively straight line. That is to consider multi-decadal shifts of the temperature, as seen in the four green lines, from a cooler condition in 1875-1920, to a warmer condition in the period1921-1955, then returning to a cooler condition for the years 1956-1985 and finally a warmer phase from the mid-to-late 1980s onward.

This latter view is considered more natural. More than that, it is also considered more consistent with our current understanding of how the sea ice, ocean temperature, salinity and circulation, air circulation and temperature, as well as many important land processes, including river runoff and snow, interact and produce the responses of the Arctic climate system.

From this perspective, one finds an Arctic climate that has a preferred tendency to produce variability that oscillates in decadal and multi-decadal periods. Several careful analyses of the sea ice changes over the Arctic also point to the dominant role played by atmospheric circulation. That component affecting the climate appears to be locked in a 50-80 year cycle - a natural see-saw - that is both large in amplitude and persistent in its timing. During these 50-80 years cycle, certain regions in the Eastern Arctic will warm a lot (as in the 1990s), while parts in the Western Arctic will cool, and vice-versa with the alternating phases of the oscillation.

Yet what should we make of the 100-plus years' warming trend marked by the red dash line in Figure 1? Is this really indicative of effects from the global warming caused by increasing CO2 in the atmosphere? To examine that question more precisely, we need some additional information.

Claiming Climate 'Trends'

Figure 3: Sensitivity of the temperature trend values (in °C per year) for Arctic (green) and Northern-Hemisphere (red) records on the length of fitting intervals starting from 17 years (1985-2001; rightmost) to 126 years (1875-2001; leftmost). (Courtesy of Igor Polyakov of IARC at the University of Alaska)

The green solid line in Figure 3 shows how one can get different answers for the trend in the Arctic temperature record depending upon the starting and end points for deriving that trend. The same is true for the Northern Hemisphere-wide temperature record, in the red solid line. Depending on how you plot the trend, it is either alarming or not, even looking at the same data.

The chart clearly shows an alarming warming rate for the Arctic - as large as 4 to 6°C per century (or 0.04 to 0.06°C per year) - seems to have occurred in the last 20 to 30 years. And, in fact, an example of such an alarming conclusion, drawn from very short climatic records, can be found in a Nov. 27, 2002, NASA/State Department-related press release detecting a change at the very alarming rate of 1.2°C per decade (or 12°C per century) for sea-ice temperature over the Arctic. This large rate of warming was implicated for the rapidly declining perennial sea-ice cover.

But there's a problem with drawing such conclusions. The temperature trend - and the perennial sea-ice trend - is very unreliable. The reason: It was based only upon the climate record from 1981-2000. That is a mere 19 years. If the temperature trend was taken for 60 to 80 years back, we would see that the trend for Arctic temperatures was actually negative. It is only a modestly positive change annually over the last 120 years.

Figure 3, in this light, becomes an important reminder of the danger of drawing a climatic trend line based on short records. And it also provides a cautionary tale for those who would attribute Arctic climate change in recent decades to the burning of fossil fuels.

The association of the observed warming trend of about 1°C over 100-years for the Arctic temperature, as seen in Figure 1, to CO2-global warming is implausible for two important reasons.

First, 70 to 80 percent of the rise of man-made CO2 in the air to date came after the 1960s. Yet, Figure 1 clearly shows that a large part of the 100-year warming trend was contributed by a pre-1960s increase in temperature. That was at a time when the air's CO2 content was still low.

Secondly, and this is a somewhat surprising fact for scientists, when the long-term temperature trend was calculated in Figure 3 using at least the 100-year long record, both the Arctic- and Northern-Hemisphere-wide warming trends have similar values.

What is so surprising about that? Well, it contradicts all the known predictions in the amplification of the polar warming. Those predictions from climate models that consider anthropogenic greenhouse gases - primarily CO2 from burning fossil fuels - to be forcing global warming say that the Arctic should warm by 1.5 to 4.5 times the global mean warming. And that is not happening.

And there's no explanation for why it is not. One explanation typically invoked to argue why there has been less rapid warming in the mid-latitudes of the northern hemisphere (Asia, Europe and America) in previous decades is so-called man-made sulfate aerosols - soot and smog - put out by industry provided a cooling factor. (Don't ask why or how the greenhouse-warming promoters are so sure of this aerosol cooling possibility while considering only one particular kind of aerosols out of many more.) But that effect is expected to be minimal, and it isn't present in the remote Arctic, thus offering no explanation to the lack of warming amplification there.

The complaint the "Arctic is melting" as a result of fossil fuel use thus has no basis from the climate records of that region and that for the Northern Hemisphere. So, it is no wonder that reports purporting to prove that are confusing and contradictory.


TCS Daily Archives