How Old is the Greenland Ice Sheet, Anyway?

File:Topographic map of Greenland bedrock.jpg
Greenland with its ice sheet removed.

Lying awake last night, the question in the title above rattled around in my head: how old is the Greenland ice sheet? Turns out it’s about 110,000 years old.  Before that, during the last interglacial period 125,000 years ago, Greenland was covered by forests. Why does this matter? Well:

– We’re in an interglacial period at the moment, and have been for about 12,000 years.

– If the Greenland ice sheet melts during interglacial periods, we should expect that it might melt at some point before the next glacial period (due anywhere from soon to 50,000 years out). Maybe it will, maybe it won’t, but it’s a possibility we should consider.

– there’s enough ice in the ice sheet to raise sea levels about 20′, which would make life very difficult for most coastal cities – if it happens fast. If it happens gradually over centuries, not so much – gradually moving your city to higher ground or abandoning it or building dikes around it aren’t a big deal, really, if you have centuries to execute.

But the real point here is that the ice sheet has formed and melted away at least a few times well before humans did much of anything to influence the climate, or, indeed, were around at all. So, it’s something to consider. The concern is that the whole sheet might melt fast enough – say, over less than a century – to cause problems for coastal dwellers. How likely is that?

Not very. Here’s a little bit from the Oracle Wikipedia. Note especially the last part:

If the entire 2,850,000 km3 (684,000 cu mi) of ice were to melt, global sea levels would rise 7.2 m (24 ft).[3] Recently, fears have grown that continued climate change will make the Greenland Ice Sheet cross a threshold where long-term melting of the ice sheet is inevitable. Climate models project that local warming in Greenland will be 3 °C (5 °F) to 9 °C (16 °F) during this century. Ice sheet models project that such a warming would initiate the long-term melting of the ice sheet, leading to a complete melting of the ice sheet (over centuries), resulting in a global sea level rise of about 7 metres (23 ft).[6] Such a rise would inundate almost every major coastal city in the world. How fast the melt would eventually occur is a matter of discussion. According to the IPCC 2001 report,[3] such warming would, if kept from rising further after the 21st Century, result in 1 to 5 meter sea level rise over the next millennium due to Greenland ice sheet melting (see image below). However, in a study published in Nature in 2013, 133 researchers analyzed a Greenland ice core from the Eemian interglacial. They concluded that GIS had been 8 degrees C warmer than today for 6000 years. The large and long-lasting warming had a modest effect on the ice sheet, leaving it largely intact.[11]

I’m continually struck by how often – virtually always – the least panicky information is put at the end of paragraph or essay, even in august sources like Wikipedia – I mean, like the NYT and SciAm. If you skimmed, or didn’t read to the end, you’d most likely miss it. As someone who dabbles in marketing professionally, the drill in marketing writing is to put what you want the customer to remember up front in as catchy (meaning: as emotionally charged) a manner as you can effect.  Anything you’d rather they not pay attention to but you must say for some reason is put in small print at the end, or spoken really fast.  I’ll make special note of such occurrences going forward.

Getting back to the topic, let’s take a detailed look at how that paragraph is structured:

Item 1: “If the entire 2,850,000 km3 (684,000 cu mi) of ice were to melt, global sea levels would rise 7.2 m (24 ft).[3]“. Very good – this is just math: X volume of water in the ice sheets spread over Y area  of oceans = rise of Z length – adjustment (if any) for increased total ocean area as low lying land become shallow ocean. 

Item 2: “Recently, fears have grown that continued climate change will make the Greenland Ice Sheet cross a threshold where long-term melting of the ice sheet is inevitable.” Fears have been felt by unnamed fear-feeling people: if temperatures continue to rise (note the awkward but evidently-required-by-some-law circumlocution: “continued climate change will make the Greenland Ice Sheet cross a threshold”)  long-term melting of the ice sheet becomes inevitable. (Never mind that long-term melting of the ice sheet is inevitable anyway, given enough time.)

Item 3: ” Climate models project that local warming in Greenland will be 3 °C (5 °F) to 9 °C (16 °F) during this century.” Climate models predict an increase in temperature in Greenland. Note that we’ve gone, so far, from good solid math in item 1 to fear being passively felt in item 2 to model output. Model output is not data. Also, this is a little nit-picky, but it bears keeping in mind that models don’t predict – people using the output of models predict. Saying models predict gives everything an undeserved aura of detachment, making the claim a passive inevitability rather than something somebody is saying.

Item 4: “Ice sheet models project that such a warming would initiate the long-term melting of the ice sheet, leading to a complete melting of the ice sheet (over centuries), resulting in a global sea level rise of about 7 metres (23 ft).” The author or authors are nearly repeating themselves here with the sea level rise numbers, but that’s not the best part – that would be the parenthetical “over centuries”. So, OK, we don’t have to panic, then? And also note we have another autonomous model making predictions all by itself.

Item 5: “Such a rise would inundate almost every major coastal city in the world.”  Yep – over the course of centuries.

Item 6: “How fast the melt would eventually occur is a matter of discussion. According to the IPCC 2001 report,[3] such warming would, if kept from rising further after the 21st Century, result in 1 to 5 meter sea level rise over the next millennium due to Greenland ice sheet melting (see image below).” Pop Quiz: How many of the world’s larger coastal cities existed at all 1,000 years ago at anything like their present sizes? How large was the coastal population 1,000 years ago? In Greece, there’s this interesting thing about the locations of port towns: they move. Marshes and deltas silt up, leaving the original city far from the coast, or filling up the harbor to the point where the port is moved.  Cities often cease to be coastal cities, or move to follow the new cost or harbor.

If the ancient Greeks could do it (often, it seems, without really noticing that that’s what they were doing), then I think, given centuries, we could do it, too. And it won’t even be a big deal – hardly noticed by any one generation – because our building with few exceptions don’t last for centuries, so, as they can be replaced over the centuries by building farther inland without much fuss.

But – fear being felt again – we are warned that the temperatures could increase even more. Yep, sure could. With that in mind, we should be very careful to check our model outputs against actual data, to make sure we get good, usable answers. If there were a big hike in temperatures over a short time – the graph would look something like a hockey stick – then we’d know we might not have the whole millennium to move our cities uphill a bit. We might only have a few centuries.

Item 7: “However, in a study published in Nature in 2013, 133 researchers analyzed a Greenland ice core from the Eemian interglacial.” Oh, look! Some real data!  Actual ice cores – physical evidence! – to examine, rather than model output. We get all the way to item 7 before any data is dragged to the witness stand.

Item 8: “They concluded that GIS had been 8 degrees C warmer than today for 6000 years. The large and long-lasting warming had a modest effect on the ice sheet, leaving it largely intact.[11]”  Looking at the physical evidence, upon which science is based and to which all models must conform or die and all modelers bow, we see it ain’t gonna happen, it being the ice sheets melting suddenly. 8 degrees C warmer than now for 6,000 years didn’t melt the ‘GIS’. So, barring a hockey-stick event even more scary-looking than those that haunt Al Gore’s dreams, that drives temperatures in Greenland up more than 8 degrees C, it looks like, based on the evidence of historical (non)meltings, we have plenty of time to take prudent steps in response to the all but inevitable eventual melting of the Greenland ice sheet.

So, not to pick too much on the authors of the paragraph above, but it would be nice, if unconventional, if once in a while when discussing science we followed a paragraph structure not taken directly from tabloid writing. I would humbly suggest that this could be achieved by two simple steps:

– Start with the data and math. They are king. Model output, if necessary, follows. Always recognize the difference between physical evidence/data and model output. One is real, the other wishful thinking unless and until backed up by the data.*

– Skip the unnamed sources who ‘fear’ or have other emotions. Emotional appeals make for good marketing and bad science.

What scientific conclusions can a layman make here? There’s a couple, always keeping the tentative nature of scientific conclusions in mind:

– The Greenland ice sheet is likely to melt at some point;

– That point is likely at least thousands of years out UNLESS;

– Temperature in Greenland increase a lot and stay hotter for a long time, therefore;

– We should keep a close eye on temperatures in Greenland.

A life lived in fear is a life half lived.

* We’re talking *future* data. In other words, a model is only and exactly as good as the degree to which predictions based on its output match what really happens. That a model matches the historical data used to build the model is a prerequisite, not a proof.

Advertisements

Author: Joseph Moore

Enough with the smarty-pants Dante quote. Just some opinionated blogger dude.

4 thoughts on “How Old is the Greenland Ice Sheet, Anyway?”

  1. It is also acceptable to build the model using half of the historical data, then test it against the other half.

    It is never acceptable to test it against the same data used to build it.

    It is double-plus ungood to test the model output against the outputs of a bunch of other models.

    1. From the model-building point of view, testing against historical data that wasn’t used to build the model is good. It helps, at the very least, to weed out the bad models. But what I’m trying to say here, not much of a point, really, is that even a model that accounted for separate historical data would still only prove useful if it usefully matches new data going forward. It’s a sort of good model versus useful model distinction. But, yea – I’ve often wondered how the current models handle the long periods of warmer temperatures that the ice cores suggest happened in the past long before humans were contributing meaningful CO2 to the atmosphere. It seems CO2 is neither necessary nor sufficient to cause temperature increases. Therefore, at the very least, the model should account for whatever other factors are involved in such a way that, by putting in reasonable values for those factors, we can ‘predict’ the past. Instead, it seems the models only go back a few decades, really.

      BTW, this raises another point that’s been rattling around my head: in this context, isn’t the distinction between ‘weather’ and ‘climate’ one of useful size? Climate being nothing more than weather over time, what is the minimum amount of time needed for a sample of weather to become a sample of climate? I’m thinking decades might just not be enough – heck, centuries might not be enough. Conversely, there’s some discussion around how the current interglacial has been blessed with extraordinarily mild weather, that the weather fossilized, as it were, in ice cores shows that previous interglacials were punctuated by some fairly wide swings in weather/climate that went on for centuries at a stretch. In that case, a sufficiently wide sample of weather will say that the climate fluctuates widely over such and such a range – which seems to me not what we mean by climate, generally.

      1. The importance of predicting the future is that every model is built upon assumptions, and many of these assumptions may be so widely accepted that they go without saying. And then the real-world system slips over into a different regime. Not only are the Xs and Ys different magnitudes, but the very relationship between X and Y may change. If Y is the length of a rubber band and X is the weight hung from it, increased weights lead to longer lengths, right up to the tensile point, where instead of stretching, it snaps. A similar example is the Newtonian model which made very nice predictions, but which was predicated upon large slow objects. Once you get to microscopic particles or to relativistic speeds, it breaks down.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s