The IPCC and climate science community are telling the world that the Arctic ocean may some time soon be free of sea ice in summer. Countries bordering the Arctic basin, and some beyond, are clamouring to lay claim to oil and gas reserves rumoured to lie below the seabed.
The sea ice record of the Arctic is in two parts. The satellite era record begins in 1979. Prior to that it is based on human, surface observations. Splicing the two methods together may clearly present problems. The pre-1979 reconstruction preferred by the IPCC is that by Walsh and Chapman that essentially shows no sea ice extent variance. The satellite record essentially shows continuous loss since it began and splicing these two together produces a perfect sea ice hockey stick.
In this post, a number of Arctic sea ice reconstructions are reviewed and compared with Walsh and Chapman. These other reconstructions tend to show much more natural fluctuation in historic sea ice levels. Furthermore, these other reconstructions tend to agree with the surface thermometer record of air temperatures while the Walsh and Chapman / IPCC view of history does not.
A number of countries bordering the Arctic seem to be convinced that the Arctic Ocean will shortly become partially or completely ice free and are taking steps to stake their claim to the huge volumes of oil and gas that will supposedly become exploitable as soon as the ice disappears. As the Daily Mail recently put it:
The Arctic is the only region in the world where borders remain unregulated, and Russia, Norway, Denmark and Canada have all made submissions to the U.N. claiming ownership of the Arctic seabed and its buried treasures. Russia is beefing up its military presence in the Arctic, sending troops and missiles to strengthen its position in the competition for the region’s extensive oil and gas reserves. As well as deploying advanced anti-aircraft missiles to the region, President Vladimir Putin is overseeing the completion of six new bases designed to see off foreign competition for the natural resources. But Russia is not the only nation to have stepped up its military designs on the Arctic. British and American submarines have recently been located in the Arctic Ocean, while Canada has also increased its military presence and President Obama has proposed to launch the U.S. icebreaker fleet …..
The US Navy has already set forth its strategy to contain Russia in its Arctic Roadmap . Here is the unclassified version:
Ensure United States Arctic sovereignty and provide homeland defense; Provide ready naval forces to respond to crisis and contingencies; Preserve freedom of the seas ….
And a war over oil would not be unprecedented. The oil embargo the US imposed on Japan in 1940 was what triggered the Japanese attack on Pearl Harbor.
It is of course to be hoped that all this is saber-rattling and will never come to anything. Yet it marks a first for climate change. We’ve all heard the wild claims of how climate change has already contributed to wars in Africa and to the emergence of IS in the Middle East, but here we have a projected climate change impact that has not yet happened – and which may never happen – that’s already setting the groundwork for possible armed conflict in the Arctic.
The question this post addresses is, how good are the data that back up the IPCC’s claim that the Arctic will become ice free at some time in the foreseeable future? Do we know enough about the historic behavior of Arctic sea ice to be certain of anything? Or is the world being led around by the nose by climate model projections based on inadequate data?
We begin with the IPCC’s Arctic Ice extent time series, which is used as the basis for ice extent model projections. It’s shown in Figure 4.3 (a) of the IPCC’s Fifth Assessment Report and reproduced below as Figure 1. The series is constructed by combining the post-1978 satellite record with the pre-1978 Chapman and Walsh and HadISST1 surface observation reconstructions. It’s basically another hockey stick, with the downward bend after 1950 coinciding with the claimed increase in man-made radiative forcings after the middle of the 20th century. Note that Figure 4.3 (b) is omitted for clarity:
Figure 1: Annual Arctic ice extent since 1870 according to the IPCC .
And Figure 2 shows a model projection of future Arctic ice concentrations developed from it:
Figure 2: Model-simulated Arctic sea ice concentrations projected to 2100
The IPCC series shown in Figure 1 does, however, exhibit a few suspicious features. One is that the blade of the hockey stick is constructed from satellite measurements and the shaft from surface ice observations, but here we will give the IPCC the benefit of the doubt and assume that these features are not artifacts of the use of two different data sets. Instead we will concentrate on two other potentially suspicious features:
1. Other Arctic ice-extent time series do not match the IPCC’s series.
2. The IPCC’s series conflicts with surface temperature observations.
Discussing these in sequence:
1. Other Arctic ice time series do not match the IPCC’s series.
There are a large number of Arctic sea ice reconstructions based on historic ice-edge observations that rarely see the light of day but which don’t show the same trends as the IPCC reconstruction, and below I show a selection of them. But first a map of the Arctic seas for orientation purposes:
Figure 3: Location of the Arctic seas
The first reconstruction is shown in Figure 16-3 of the 2007 AR4, reproduced below as Figure 4 (the original data are from Vinje 2000) . It shows a gradual ice loss in April sea ice extent over the Nordic seas since 1865, but the rate of ice loss decreases rather than increases with time and there is no sign of an acceleration after 1960. The series covers only the Nordic seas and shows only April data, but it is nevertheless difficult to reconcile with the IPCC series, which bends in the opposite direction. (Vinje’s conclusion, incidentally, was that “ the recent decrease in ice extent is within the range of variability observed since the eighteenth century”):
Figure 4: Vinje ice extent time series, Nordic Seas, 1865-2000
A more directly comparable data set is that of Zakharov, as described in Johanessen et al. 2004. Zakharov covers about two-thirds of the total area of the Arctic Ocean – a much larger fraction than the Nordic Seas plot shown in Figure 4 – and is compared with the “Walsh” contribution to the IPCC reconstruction (the red line in Figure 1) in Johannessen et al’s Figure 5. Similar short-term trends are visible in both data sets but the longer -term trends are quite different, with Walsh showing rapid ice loss after 1950 and Zakharov showing hardly any.
Figure 5: Time series derived from Zakharov sea ice data set, 1900-2000
The Polyakov et al 2003 data set also covers the four “arctic” (Siberian) seas and shows present-day ice extents which are not that different to those of 100 years ago. Polyakov et al. conclude as follows: Examination of records of fast ice thickness (1936–2000) and ice extent (1900–2000) in the Kara, Laptev, East Siberian, and Chukchi Seas provide evidence that long-term ice thickness and extent trends are small and generally not statistically significant.
Figure 6: Polyakov time series, Arctic seas (Y scale in 1000km2)
Figure 7 shows Polyakov’s series for the four seas.
Figure 7: Polyakov time series for the four Arctic seas (Y scale in 1000km2)
Next come Divine & Dick (2004), who present the Nordic Sea ice edge anomaly reconstructions for April, June and August shown in Figure 8. Like the Vinje data in Figure 2 they show a gradual overall decline in sea ice extent since the late 1800s but no evidence for a recent acceleration in the rate of ice loss:
Figure 8: Divine & Dick April, June & August ice edge anomaly series, 1860-2000
I have more plots of Arctic ice extent reconstructions that I can no longer find the links for but which I am going to show anyway because I have no reason to believe there is anything wrong with them. First, Miles et al (2007) show an increase in the rate of sea ice loss, but starting around 1890, not in 1970 (Figure 9):
Figure 9: Miles et al. ice-edge anomaly series, Greenland and West Nordic Seas
Ogilvie (1992) also shows an acceleration in the rate of ice loss in the Nordic seas, but in this case beginning around 1850:
Figure 10: Ice extent, Nordic Seas, “mainly based on logbook observations and Icelandic coast observations by Ogilvie (1992):
Then there are the thumbnails shown in Figure 11, which I believe are from Vinje: Do they define the Little Ice Age?
Then there’s the Hill reconstruction of winter sea ice extent off Newfoundland, which shows sea ice beginning to retreat after 1920 but staying roughly stable since 1970:
Figure 12: Hill time series of ice extent off Newfoundland
And finally this reconstruction by Ogilvie & Jónsdóttir of sea ice conditions off Iceland, which goes all the way back to 1600 and shows more ice off Iceland now than there was 400 years ago:
Figure 13: Ogilvie & Jónsdóttir sea ice index off Iceland since 1600.
Above we have a dozen or more ice-extent plots that don’t match the IPCC’s version of events (Figure 1). What do they tell us? That the IPCC’s version is potentially suspect. And as discussed briefly in the second part of the analysis it becomes even more so when we inject surface temperature observations:
2. The IPCC series doesn’t match surface temperature observations.
To illustrate this point I use the four series shown in Figure 14, all of which except for Miles have been shown earlier (I have the annual ice extent values downloaded for all of them). The Vinje, Zakharov and Miles series are constructed from ice edge observations. The IPCC series is a version of the Chapman-Walsh series that closely matches the IPCC series shown in Figure 1. I have juggled the ice extent scales so that all four series plot in about the same place but have made no attempt to relate ice extents to temperature change in a quantitative sense:
Figure 14: Vinje, Zakharov and Miles ice extent series compared with IPCC ice extent series (Walsh-Chapman). The scales have been adjusted to line the plots up.
Figure 16 now plots the four series against Arctic annual mean surface air temperatures since 1880 (my estimates based on 49 Arctic surface station records). The scales are again adjusted so that they line up and the ice extent scales are inverted so that ice extent and temperatures move in the same sense:
Figure 15: Figure 14 ice extent series compared with Arctic surface air temperatures since 1880.
Vinje, Zhakarov and in particular Miles all show a tendency for ice extent to decrease as temperatures increase, which is what we would expect to see. The IPCC series, however, shows ice extent tracking temperature only after 1970 – a period roughly coincident with the satellite measurements. It shows no visible response to the equally large temperature changes before that. According to the IPCC neither the ~1.5C increase in Arctic surface temperatures between 1880 and 1940 nor the ~1C decrease between 1940 and 1970 had any significant impact on Arctic sea ice extent. This is implausible.
So what’s the bottom line? Basically that we don’t have any reliable records of how Arctic sea ice behaved before the satellite data begin in 1978 (assuming that the satellite records are themselves giving us the right answer). So when the IPCC’s computer models project that Arctic sea ice will disappear at some point in the foreseeable future they are once again extrapolating into the next one hundred years without any understanding of what happened in the last. The countries that are now casting avaricious eyes towards the oil and gas bonanza that an ice-free Arctic would liberate should bear this fact in mind.