Greenland Melt

This year, Europe was hit with multiple extreme heat waves. After inflicting pain on Europeans, some of that heat sauntered over to Greenland. The result? Greenland showed greater melt this summer than ever before.


But this year’s ice melt in Greenland should be put in context. Estimates of Greenland’s total ice mass tend to agree pretty well, as summarized by Bamber et al. 2018

The graph shows three different estimates: from altimetry, from an input-output model (IOM), and from GRACE (the Gravity Recovery and Climate Experiment). All three agree that since about 2003, Greenland ice is in decline. Serious decline.

I retrieved the GRACE data, which only go to May of 2019 so they don’t show the super-melt of this year. But, we’re not so concerned with this year’s super-melt as we are with what happened before. There are some missing months in 2017 and 2018, but that won’t prevent us from examing what there is. This (ice mass in gigatonnes):

Not only is there a clear decline over time, there’s also a seasonal cycle. That’s no surprise with ice: it tends to increase in winter and decrease in summer. But we’re more interested in the trend than in that seasons, so I’ll remove the annual cycle to create de-seasonalized Greenland ice mass.

If we fit a straight line (by least squares regression), it shows that from 2003 until now, Greenland ice has melted at an average rate of 223 Gt/yr (gigatonnes per year). But that’s the average over the entire period; has it been constant? If we look at what’s left over after subtracting that constant trend, we find how the data depart from that constant trend. Like this:

Those residuals don’t look like just random noise! Statistical tests confirm, it isn’t. So what has Greenland ice melt really been up to?

I fit a couple of models to these data (statistical models, not computer simulations). One is the piecewise-linear fit, which models the trend as a set of straight lines connected at their endpoints. The times to switch from one model to the next are chosen by finding the best-fit times, as long as they meet the “statistically significant” criterion — there’s no sense changing the line if there’s no evidence that that actually happened to the trend.

I ended up with a model consisting of three straight-line segments, shown here in blue:

I also fit a lowess smooth to the data:

Both models enable me to estimate the rate of change of Greenland ice.
Here is what they say about how the rate of decline has changed over time (the given uncertainty ranges are 2 standard errors above and below the estimates):

The piecewise linear fit gives constant rates over each straight-line segment, of course — because the model consists of straight lines. The lowess smooth shows more detailed changes, but also less precision, because the uncertainty in the rate right now is bigger than the uncertainty in the average rate over a longer time span.

There was a time (from August 2009 through November 2012) when the ice declined faster than before and after, at the rate of 361 Gt/yr. The rate of ice loss since 2013 has “only” been 131 Gt/yr.

This year’s super-melt is likely to change the trend yet again. Unfortunately, it’s likely to decline even faster. That bodes ill for the people who live in Greenland, and it bodes ill for the rest of us too. That meltwater ends up in the ocean, raising sea level (Greenland has enough ice to raise sea level by 7 meters — 23 feet — if it all melts) and interfering with the Gulf Stream current in the Atlantic ocean.

Folks who live on the coast are already feeling the heat, with local flooding an ever-more-serious problem. But the real worry is that it’s going to get worse. Greenland will continue to melt, probably even faster than before. How fast it melts, and how hard it is to deal with, depends on us. If we cut carbon emissions quickly and dramatically, we’ll save ourselves billions of dollars by reducing tidal flooding and storm surge.

But, doing that would require that we actually practice wisdom, making choices based on facing the truth of what the future will bring. In general, that’s not a strong point for most politicians. As for President grab-’em-by-the-pussy? Wisdom has he none, truth he fears.


This blog is made possible by readers like you; join others by donating at My Wee Dragon.


2 responses to “Greenland Melt

  1. Asked on Twitter, why the Greenland ice melt varies like that. I could have guessed the answer. Ruth Mottram wrote: the North Atlantic Oscilation. https://www.researchgate.net/publication/334445470_Decelerated_Greenland_Ice_Sheet_Melt_Driven_by_Positive_Summer_North_Atlantic_Oscillation

  2. The steep decline in your residuals coincided with a slight uptick in summer sea ice during the 2014 and 2015 years. These can all be attributed to the volcanic eruption of Bárðarbunga in Iceland.