Don't bet against physics

Don't be surprised by short term fluctuations in Earth systems as the planet warms. In the long term, heat is hard to overcome.

A sunny day in the Arctic in September of 2009. Photo by Patrick Kelley, U.S. Coast Guard, via Flickr. Public Domain.

As readers of ReportEarth likely already know, 2023 and 2024 were the two hottest years on record for the globe. And at least according to the European Union’s Copernicus Climate Change Service, the temperature of the latter year came in at more than 1.5 degrees Celsius above pre-industrial levels, defined as the average temperature between 1850 and 1900.

As of now, 2025 is keeping pace pretty well, running warm enough that it seems likely to come in 3rd or possibly even 2nd in the rankings. While temperatures moderated a bit over the northern hemisphere summer, September temperatures have spiked back upwards again, very nearly into 1.5 degrees C territory on a monthly basis. That makes this the third warmest September on record, according to Copernicus.

Here’s a chart I’ve made to visualize this, with helpful feedback from Copernicus scientists. You can see that each decade since the 1970s is notably warmer than the last. You can also see that the decade of the 2020s is accumulating a number of quite hot years, possibly enough of them that, if 2026-2029 come in warm as well, it will put the decade as a whole very close to the 1.5 degrees C threshold in this dataset. (Note that on mobile devices, the labeling for this and other charts below may not work as well, and I recommend viewing them on desktop or in email if you are a subscriber.)

It takes more than a single year for the entire Earth to cross the critical 1.5 degrees C threshold. Scientific institutions are holding out for the long term planetary average temperature to cross this line, although this average can be defined in different ways.

Either way, the temperatures of 2025 are very important to watch. Why? Because we’re living right now in a year that will probably be going into that average.

Monitoring the world’s ice amid fast warming

So what’s all that heat that doing to the rest of the Earth, and in particular, to its iconic ice regions, or what scientists call the cryosphere?

The striking thing is that, amid this recent bout of major warmth, the world’s polar ice has not been marching entirely in lockstep. To be clear, there’s nothing going on that undermines the big picture view, which is that it’s warming, we’re going to blow past 1.5 degrees C, and in the long term, the planet is going to lose more ice.

However, we need to acknowledge and understand what’s happening, if only to be better able to respond to the hundredth “Antarctic is gaining ice” Facebook meme.

So let’s look at sea ice first. Then we’ll go on to the world’s ice sheets.

I recently made the following chart. It is inspired by the National Snow and Ice Data Center’s wonderful “Charctic” interactive sea ice graph, but shows all years individually back to 1978, and shows the Arctic and Antarctic at the same time, amounting to more than 34,000 daily data points overall. This brings out a striking pattern, in which the two poles’ seasonal cycles create a shape resembling an infinity symbol (make of that what you will). The chart is currently updated through October 20, but highlights the September highs and lows in particular:

Turning first to the Antarctic, sea ice extent was quite low throughout the year. The summer minimum extent effectively tied for the second lowest on record (with 2022 and 2024), according to the tally of the National Snow and Ice Data Center. The winter maximum extent was the third lowest on record.

Here’s another way of looking at this, emphasizing Antarctic sea ice extent has recently moved to notably lower levels:

I think it is very important to consider this Antarctic record in the context of the current moment. Back in 2014, Antarctic sea ice hit a dramatic new high, lending fuel to those who are skeptical of some aspects of climate change science. This was also around the era in which there was great talk of a global warming “pause.” And yet, in both cases, look at where we are now.

Which brings us to the Arctic.

In the Arctic, 2025’s winter maximum extent came in as a new record low. But the summer low, which falls in September, only ranks at 11th place when you consider it as a monthly average (it tied for 10th if you consider the single lowest day of the year). Here’s a chart showing the monthly average values:

And here’s the mystery: September Arctic sea ice levels have been pretty flat, going on nearly two decades now. That’s not what you’d naturally expect on a warming planet. Here’s what happens if you fit a simple linear trendline to the September values from 2007 and on, which I’ve done my best to show in Datawrapper with little squares (Datawrapper is a wonderful tool, but it can be hard to get it to do custom trendlines):

As you can see, Septembers are basically flat lately. This is something the National Snow and Ice Data Center itself notes, writing, “While the last 19 years, from 2007 to 2025, are the lowest 19 sea ice extents in the satellite record, there has been no significant trend in September minimum extents over this time.” So why don’t we see a continuing downtrend in this critical month? Why haven’t we seen a new low even lower than the September 2012 record?

Scientists are on the case, and several studies have now addressed this. One suggests an Arctic “regime shift” occurred in 2007: The ice moved into a distinctly shrunken overall configuration, but then stayed there, without any additional downward trend. Another recent paper suggests this is the kind of thing that can happen even on a continually warming Earth, once you take natural climate variability into account.

A question often asked is, what about the three dimensional volume or thickness of Arctic sea ice, rather than its two dimensional extent? Well, other research suggests that the loss of sea ice volume has also slowed recently.

I asked Walt Meier of the National Snow and Ice Data Center, with whom I check many of these charts, for his take on what’s going on with Arctic Septembers lately. The losses up to around 2007 or so, Meier noted, took out the “easy to melt” ice.

“What’s left is the older, thicker ice that is near the pole, so not as much sunlight, colder temperatures,” Meier explained by email. “And the remaining ice is towards the coast of Greenland and the Canadian Archipelago, which the ice drift tends to replenish and build up thickness through ridging. So, I think the September extent may have reached a bit of a wall where it will take further warming to give the climate enough kick to melt through the thicker ice that gets less energy.”

This means that fortunately, we probably still have a ways to go before seeing an ice free summer Arctic ocean in summer, which would be an enormously consequential event. The IPCC Sixth Assessment Report, the latest installment of the mega-consensus document, says this will likely happen for the first time sometime before 2050, or within the next 25 years. So while there’s good reason to worry about the long term future of Arctic sea ice, the emphasis is on long term.

In sum: Arctic sea ice losses in September have slowed, or paused, depending on how you want to phrase it. But as the Earth traps more and more heat due to greenhouse gases, this is expected to eventually melt the harder-to-melt sea ice in summer, too. We’re not there yet, and yes it takes time, and yes, there is natural variability…but don’t bet against physics.

The ice sheet picture

With that, let’s turn to the world’s ice sheets.

I find that my charts get a fair amount of traction these days on Facebook. This was most recently the case with the infinity shaped sea ice chart above. But presumably because of this, Facebook also constantly feeds me other charts claiming that Antarctic recently has “gained ice,” or something along these lines. It’s another apparent anomaly, not unlike the Arctic one discussed above.

Once again, then, we turn to the data to see what we are dealing with:

As you can see, Antarctica and Greenland have, together, lost trillions of tons of ice in the last two decades. But more recently, Antarctic losses have flattened, and there’s arguably a fairly small gain back of some ice.

I discussed, back in May, the cause of this, so I won’t go into as much detail here as I did for Arctic sea ice. But briefly, the cause is very heavy snowfall, which may itself be linked to climate change. In recent years, the continent has been accumulating as much or more ice from snow as it has been losing at its edges.

The recent losses of Antarctic sea ice, shown above, are probably themselves a contributing factor to the new weather conditions that have been driving more snow and ice accumulation inland. These sea ice changes are “bringing open marine (albeit polar marine) conditions closer to the ice sheet edge,” said Ted Scambos, an Antarctic expert at the University of Colorado, Boulder, by email.

This anomaly may continue for a while. It’s even possible we could go into a major ice accumulation state and see the trend reverse substantially. Scientists aren’t sure.

But they once again expect that ocean heat, which is already having major impacts in West Antarctica, will ultimately take its toll. “In the long term, the geologic past (from ice core records) shows us that warmer conditions lead to a smaller ice sheet in Antarctica,” Scambos said.

The conclusion, overall: Even on a warming Earth, not everything moves in lockstep. The system remains complex, with various natural cycles and patterns, feedbacks, and so on. If you have a system strongly influenced by heat, like the cryosphere, it’s still apparent which way things are headed in the long term. But short term fluctuations, hiccups, slowdowns, pauses, whatever you want to call it: They happen.

We acknowledge them, in context, and we move on.

For more of my charts, which are growing in number weekly, check on the chart archive here! Thanks for reading!