select search filters
briefings
roundups & rapid reactions
Fiona fox's blog
A modelling study published in Geophysical Research Letters looks at Arctic sea ice loss in the last 20 years.
Dr Michel Tsamados, Lecturer, Centre for Polar Observation and Modelling, University College London (UCL),
“By contrasting model results and satellite observations over the 2005-2024 period this study documents clearly a ‘slowdown’ or ‘pause’ in Arctic sea ice extent (the area where sea ice fraction is above 15%).
“Similar pauses or hiatuses are natural in the climate system and were extensively scrutinised in the past in the context of global climate warming (see 1998-2013 period).
“In the case of sea ice decline this is made all the more surprising because the Arctic is warming up to 4 times faster than the rest of the planet and one would expect a continued sea ice decline.
“I think the study is good and sound but might hide some more nuanced aspects of sea ice decline. I have always found the metric ‘sea ice extent’ quite simplistic in that it encompasses a complex 3D field into a 2D projection. Other metrics, such as sea ice age for example, paint an even more alarming picture. The authors briefly address this by looking at a model results on sea ice thickness but don’t dive deeper into how their results hold when looking at the total sea ice volume. Satellite data on sea ice thickness exist and would be worth looking into [1].
“Another aspect of the study that is lacking some discussion and would be very interesting to investigate more is the so called ‘model response uncertainty’. In other words how much of the model trends are due to deficiencies and biases in the models. From my experience [2] sea ice is too thick in climate models and there are also issues with the feedback processes in both summer and winter in the models.
“To sum up, I wouldn’t call this a good news story as the recent slowdown of sea ice extent is only one part of a complex Arctic landscape (i.e. ice volume, ecology, temperature). I also don’t think that we can draw any confident projections from the finding of this study that the current slowdown will persist in the coming years.”
Dr Gaëlle Veyssière, Sea ice and Snow Physicist, British Antarctic Survey, said:
What should readers take from this new research?
“This study shows that, over the past 20 years, the observed amount of Arctic sea ice loss has been smaller than in the decades before (up until 1979; when satellite imagery made this type of observations possible), and this slowdown is noticeable in all seasons. Climate models are also involved in the analysis, and they show that such pauses can happen naturally, even while the global temperature continues to increase due to human activities. The authors find that natural climate variability have likely masked much of the human-driven ice loss during this period, but this does not mean global warming has stopped or that Arctic sea ice is “safe” now. Once the natural variability shifts, the ice is expected to keep shrinking, and possibly at a faster rate than before.
Are there any important caveats to be aware of?
“There are a few important points to keep in mind here:
“It’s a ‘pause’ in the rate of loss of sea ice. Arctic sea ice is still far lower than it was in the 1980s and the decline has overall slowed down for the study period 2005-2024.
“Natural climate variability is likely to be the driving factor. Some human-related influences might have played a role too, but that part is uncertain.
“The climate modelling analysis suggests that this slowdown could continue for a few more years. However, it will remain temporary and once the natural variability shifts, the sea ice is likely to keep decreasing, potentially faster than before.
“This pause should not be confused with a sign of recovery. It doesn’t indicate that Arctic sea ice is improving or that we are seeing reversing climate trends. As the authors indicated, without human-caused warming, sea ice would likely have increased over this period.”
Could the slower rate of sea ice loss be seen as good news, in the sense that it gives some more time for climate action to ramp up?
“This slowdown is very likely temporary and caused mostly by natural variability, not by a significant change in the underlying human-driven global warming. We might have a little more time before we reach critical milestones like the first ice-free Arctic summer, however, once that variability swings the other way, ice loss could accelerate, meaning we could lose any apparent “extra time” quickly. It’s more like a brief intermission in a storm, not the storm passing.”
Dr Ed Blockley, Leader of Polar Climate Group, Met office Hadley Centre, said:
“It is well known that multi-decadal variability can exert considerable influence on the evolution of Arctic sea ice. Although we have very high confidence that anthropogenic warming is responsible for the majority of Arctic sea ice decline observed in the satellite record, we know that long-term variability patterns can play a role in modulating that decline.
This paper nicely highlights the important role that multi-decadal variability plays in the evolution of Arctic sea ice, that such slow-down events are not unexpected, and that, collectively, the climate models are able to represent this important behaviour.
“The influence of such long timescale variability on the system poses a considerable challenge for the projection of Arctic sea ice in the future. For this reason we are only able to make projections with around 10-20 years uncertainty of the date in which the Arctic Ocean will first be ice-free.
“Understanding long-term variability patterns such as this, their underlying driving mechanisms, and the models’ ability to represent them, is therefore very important to improve both our understanding of past changes in Arctic sea ice, and our confidence in future projections.”
Prof Julienne Stroeve, Professor of Polar Observation & Modelling, University College London (UCL), said:
“I would say two things, one it is not surprising that the September minimum extent hasn’t changed much the last decade as we know pauses in climate records, be in global temperatures or sea ice can remain the same for several years in a row as a result of internal climate variability. This has been reported in previous studies and thus I do not find the new study particularly novel. Second, there is a negative feedback whereby at the end of summer when temperatures drop again new ice forms and this ice grows thermodynamically to about 1.5 to 2m over winter. This ice growth helps to offset continued summer ice loss. So until we start to impact the amount of ice that can grow over a winter season by (1) shortening the ice-growth season and (2) reduced freezing degree days, then this negative feedback will help to provide a buffer to summer ice loss. I was a co-author on the Notz and Stroeve (2016) paper where we related the loss of September sea ice to cumulative CO2. If I update that plot through 2024 it looks like the one below. The relationship is still linear but it’s not as strong as we previously reported.”
Prof Andrew Shepherd, Director of the Centre for Polar Observation and Modelling, Northumbria University, said:
What should readers take from this new research?
“The intended message from this research is that the recent variability in arctic sea ice extent is not evidence that the historical trend has stopped, because the past climate simulations suggest that 20-30 year pauses are relatively common. There is about a 20% chance that such a situation is natural. I suspect most will view this to be a slim chance, and so the main message will likely not land in the way the authors had intended.
Are there any important caveats to be aware of?
“Yes. The authors are in my view not considering two important issues.
“First, there are two extreme melting years in the record they use – 2008 and 2011 (see Figure 1). In addition to long term trends, a known feature of climate change is the increased propensity for extreme events. Historical simulations are unlikely to capture these, and so a comparison of the recent trend to those simulations is not ideal. An alternative statistical analysis could be to assess the impact of those extreme events; if they are excluded from the record then the recent trend is almost certainly negative and significant. My concern is therefore that the 20% chance is almost certainly an underestimate.
“Second, the paper is only considering trends in extent. The amount of sea ice is also dependent upon its thickness, and that is an important consideration for two reasons. First, as the ice pack retreats the first ice to be lost is at the margins and this is the thinnest, and it is expected that the pace of retreat will slow down over time. So this is not an indication that the forcing (climate warming) has changed; rather it is consistent with it. Second, we know that the Arctic sea ice pack is thinning also, and so even if the extent were not reducing the volume still is. Our data show that since 2010 the average October thickness has fallen by 0.6 centimetres per year.”
Could the slower rate of sea ice loss be seen as good news, in the sense that it gives some more time for climate action to ramp up?
“No, for the second reason above”
‘Minimal Arctic sea ice loss in the last 20 years, consistent with internal climate variability’ by England et al. was published in Geophysical Research Letters on 5th August 2025.
Declared interests
For all experts, no reply to our request for DOIs was received.