Glaciers in almost every part of the world are getting smaller, often at jaw-dropping rates.  That seems like a pretty clear sign that the world is warming, right?  It’s good enough for most people, but at long last, there is paper to tell you exactly what ‘good’ and ‘enough’ mean.

Let’s back up for a minute.  A glacier is a big piece of ice that slowly flows downhill.  If it collects more snow in the winter than it loses to melt in the summer, it gets thicker, and, eventually, grows longer, and if the opposite happens, it thins and shrinks.  Because it takes years for glaciers to grow or shrink, their changes tell us not so much about the weather this year or last year, as about how the climate has changed over the last few decades.  What’s more, glaciers are big and easy to measure, so even in places where we don’t have great long-term records about weather, we often know how the glaciers have changed for centuries.

With this in mind, if we see a glacier getting smaller, we can tell that the climate around it is warming, and if glaciers all over the world are thinning, we know the global climate is warming, right?  Yep, pretty much right, but in today’s anti-science world, you can find arguments that this isn’t the case.  Maybe all these shrinking glaciers are just shrinking because of the end of the Little Ice Age, a cold period that ended in the nineteenth century.  And any one glacier might just be thinning in response to a few warm recent years.

Enter Gerard Roe, Marcia Baker, and Florian Herla.  In their paper from earlier this year, they looked at snowfall, melt, and glacier length measurements from a set of 37 glaciers scattered around the world.  They used a simple mathematical model to calculate how sensitive the length of each glacier should be to short-term and long-term fluctuations in temperature and snowfall.  Their model showed that changes in weather, like a couple of extra-snowy or extra-warm years, don’t lead to noticeable changes in glacier length, but even small changes in temperature or snowfall can make glaciers grow or shrink if they persist for decades.  Their models, presented with random climate records that had the same year-to-year fluctuations as the real record but no long-term trend, changed by less than a kilometer.  Presented with climate records that included the real-world warming trend, the models did what the real glaciers did, and retreated rapidly.  Comparing the size of the no-climate-warming fluctuations with the size of the glacier retreats, the authors could work out the chance that the glacier retreats were caused by random climate variations.  And the result was pretty convincing. There was a small, 11%, chance that the change in one of the 37 glaciers was due to random climate fluctuations.  For most of the rest, that chance was less than 1%.

That’s a pretty strong, unambiguous finding, and the paper ends there.  As James Brown used to say, “Kill ‘em and leave.” Glaciers around the world are getting smaller, and the ones that have good climate records aren’t doing it by accident.  But the point I wanted to make by bringing up this paper is that any idea that seems obvious can be checked by careful science, and that for all the scientists going to beautiful places making observations, there are a few sitting in offices checking to see if the observations mean what we think they mean. In this case they did, but if they didn’t, Gerard and company would have happily published a different paper telling us we were blowing smoke.