Climate scientists admit they have a 90% chance of being wrong about Arctic sea ice

Guest Post by Javier Vinós

Arctic sea ice is lowest during the month of September, and its average extent during this month is a useful metric for measuring Arctic sea ice decline during the current period of global warming. During the 1980s and 1990s, September Arctic sea-ice extent (SIE) showed a moderate decline (Figure 1). After the 1997 climate shift, which involved a rather abrupt global atmospheric reorganization, the Arctic entered a period of rapid change that I call the Arctic Shift.^[1] During this period, Arctic SIE declined more rapidly. Scientists noticed this change in trend about a decade later and became increasingly concerned about the prospect of an ice-free Arctic.^[2]

The concern about the rapid decline of Arctic SIE in the early years of this century was due to the possibility of a runaway ice-albedo feedback. Loss of sea ice would reduce albedo, and additional solar energy would cause further sea ice loss. Models that reproduced the rapid loss predicted a tipping point that would lead to an ice-free Arctic by 2040, sparking public fears.^[3] However, recent work suggests that up to 60% of the decline in September SIE since 1979 may be due to changes in atmospheric circulation.^[4] In addition, the persistence of Arctic summer cloud cover significantly reduces the ice-albedo feedback.^[5] The realization that internal variability is a more important factor than expected explains why the rate of decline of Arctic summer SIE has slowed so much since 2007, contrary to all expectations.

The Arctic Shift, a period of adjustment of Arctic climate variables to the new atmospheric regime induced by the 1997 climate shift, ended for Arctic SIE in 2007. Since then, the September Arctic SIE shows no significant trend. However, climate researchers are still unaware of the effects of climate shifts and regimes on climate change, and they were surprised by the recovery of sea ice in 2013 when it became clear that there had been no net loss since 2007. Using models, they calculated a 34% chance of a 7-year pause (Figure 2).^[6]

However, the hiatus has now extended to 17 years and the probability has dropped to 10%. In other words, there is a 90% chance that climate scientists’ predictions about Arctic sea ice were wrong. If the hiatus continues until 2027, it will become statistically significant (p<0.05, or less than 5%) and no longer explainable by chance. For an explanation of the observed Arctic changes, see chapters 34 and 42 of my forthcoming book “Solving the Climate Puzzle. The Sun’s Surprising Role”.

The current state of affairs has led society to be alarmed by model predictions that have been proven wrong by the time they are published, but this often goes unnoticed. A recent example of this phenomenon is shown in Figure 3. In June 2023, news headlines around the world highlighted a scientific study that warned of the possibility of ice-free summers in the Arctic by the 2030s, regardless of our efforts to reduce emissions.

The article presents projections based on observations of an ice-free Arctic even under a low emissions scenario.^[7] However, it should be noted that the data in the article only cover observations through 2019, although data for 2020-22 were available at the time of publication. In addition, the model projections in the study begin in 2021. Figure 3 shows the results of the study under an intermediate emissions scenario similar to the current situation. However, a significant problem arises when considering the acceptance and publication of the paper, as the model projections for 2021 and 2022 differ greatly from the observed data, with a staggering difference of 1.3 million km² (0.5 million square miles) or 33% lower. This obvious problem, which undermines the entire study, raises questions about how the paper was accepted for publication.

How could such a blatantly flawed, and provably incorrect, article successfully pass the peer-review process? Moreover, who determines its suitability for widespread dissemination in a global media landscape that seems incapable of questioning or scrutinizing these predictions? The data refuting the article are readily available to anyone with an Internet connection and can easily be located with a simple search engine query. The current method of communicating predictions from highly uncertain climate models to the public is undeniably inadequate, and it is truly surprising that no authoritative scientific voice has addressed this issue and voiced disapproval.

Note: Part of the text and some of the figures in this article are taken from several chapters of my forthcoming book, “Solving the Climate Puzzle. The Sun’s Surprising Role,” to be published in November 2023.

1. Vinós, J., 2022. Climate of the Past, Present and Future: A scientific debate. 2^nd ed. Critical Science Press. ↑

2. Stroeve, J.C., et al., 2005. Geophys. Res. Lett. 32 (4). doi.org/10.1029/2004GL021810 ↑

3. Holland, M.M., et al., 2006. Geophys. Res. Lett. 33 (23). doi.org/10.1029/2006GL028024 ↑

4. Ding, Q., et al., 2017. Nat. Clim. Chang. 7 (4), pp.289–295. doi.org/10.1038/nclimate3241 ↑

5. Sledd, A. & L’Ecuyer, T.S., 2021. Front. Earth Sci. p.1067. doi.org/10.3389/feart.2021.769844 ↑

6. Swart, N.C., et al., 2015. Nat. Clim. Change, 5 (2), pp.86–89. doi.org/10.1038/nclimate2483 ↑

7. Kim, Y.H., et al., 2023. Nat. Commun. 14 (1), p.3139. doi.org/10.1038/s41467-023-38511-8 ↑