top of page

Estimating Arctic sea ice trends before the satellite era

Satellite observations indicate that the average Arctic sea ice extent has generally decreased since the start of the satellite records in October 1978. However, is the 1978-present satellite record really long enough to allow us to:

  1. Assess how unusual (or not) the recent trends are?

  2. Determine how much of the recent climate change is human-caused vs. natural?

In 2017, three of us published a study in Hydrological Sciences Journal in which we extended the Arctic sea ice estimates back to 1901 using various pre-satellite era data sources. In our study, we found that the recent Arctic sea ice retreat during the satellite era actually followed a period of sea ice growth after the mid-1940s, which in turn followed a period of sea ice retreat after the 1910s. This suggests that the Arctic sea ice is a lot more dynamic than you might think from just considering the satellite records.


Sea ice trends during the satellite era

The Arctic and Antarctic sea ice extent satellite data can be downloaded from the US National Snow & Ice Data Center (NSIDC) here. In the graphs below, we’ve plotted the average annual sea ice extents from this satellite data for both the Arctic and the Antarctic. For comparison, we’ve also shown Arctic air temperature trends since 1900 (adapted from our HSJ article).

We can see that, yes, the average Arctic sea ice extent has generally decreased since the start of the satellite record. Although, interestingly the average Antarctic sea ice extent has generally increased over the same period. However, when we look at the much longer Arctic temperature record we can see that this is not surprising. The Arctic region has been warming since the late 1970s (when the satellite records began), but this followed a period of Arctic cooling from the 1940s to the early 1970s! In other words, if the satellite records had begun in the 1940s and if the Arctic sea ice extent is related to Arctic temperatures, we would probably have detected a period of Arctic sea ice growth.


Arctic sea ice changes during the pre-satellite era

One of the reasons there has been such interest in the satellite-based sea ice records is that the satellites are monitoring most of the planet and provide almost continuous coverage. But, people were also monitoring Arctic sea ice before the satellite era using various land, ship, submarine, buoy and aircraft measurements.

The main problem is that the pre-satellite data is unfortunately very limited. If a ship travelled through a particular region in a given season, then they could have reported how much ice was in that region, or whether it was ice-free. But, what do you do if there were no ships (or airplanes, buoys, etc.) in that region?

So, we realised that the pre-satellite data needs to be re-calibrated to account for the limited observations and also the changes in different data sources (airplanes vs. ships vs. buoys, etc.) for different regions and times. Essentially, we used Arctic temperature records from weather stations on land to ensure that the sea ice measurements from each of the data sources show a similar response to Arctic temperatures to that observed in the satellite era.

After re-calibration, we obtained the following result including error bars:


There are several points to notice:

  • While Arctic sea ice has indeed been generally decreasing since the start of the satellite era, this coincidentally followed a period of Arctic sea ice growth from the 1940s to 1970s!

  • Indeed, the Arctic seems to routinely alternate between periods of sea ice growth and sea ice retreat.

  • If we ignore the error bars, perhaps you could argue that sea ice extents since 2005 are lower than they have been since 1901. However, we shouldn’t ignore the error bars. We can see that the lower error bars for the pre-satellite era have been lower at several stages than the upper error bars for the entire satellite era. In other words, the recent low values are still consistent with our estimates for the pre-satellite era.


For more information, see:

R. Connolly, M. Connolly and W. Soon, 2017. Re-calibration of Arctic sea ice extent datasets using Arctic surface air temperature records. Hydrological Sciences Journal, Vol. 62, pp 1317-1340.; (🗄️).

bottom of page