Islands in the ice: Models, math, penguins & a circumnavigation
In amongst the ice sheets and glaciers, there are small rocky outcrops and mountain tops that form suitable habitat for Antarctica’s biodiversity. Climate change could drive expansion of these ice-free areas by as much as 25% by the end of this century, leading to new opportunities for both native and non-native species…
What’s the first thing that comes to mind when you picture Antarctica? Probably ice, snow, penguins or killer whales, right? Well, those things are definitely there - yet the frozen continent is also home to a wide array of extremely tough and well adapted plants and invertebrates, including moss, lichens, springtails, rotifers, nematodes and tardigrades. Like elsewhere on the planet, these species face multiple threats, including human activity, invasive species and of course climate change.
Cryptopygus antarcticus - a native Antarctic springtail (Photo: Melissa Houghton)
I am a PhD candidate in the Centre for Biodiversity and Conservation Science at the University of Queensland in Brisbane. My PhD research is all about determining how we can best conserve Antarctic biodiversity in the face of these multiple stressors, which begins with understanding how these threats will impact the terrestrial communities.
Over 99% of terrestrial biodiversity survives in small islands of suitable habitat in a sea of ice. Ice-free areas make up less than 1% of the continent and are scattered as small patches around the coast, as cliff faces, or as the tops of mountains that emerge out of the ice sheets. Being essential to life in Antarctica – understanding how biodiversity habitat would be impacted by climate change seemed the obvious place to start my research.
To my surprise, no one had really looked at how ice-free areas might be affected by climate change, especially not on a broad scale. This may be because a lot of the research about climate change impacts on the Antarctic environment has focused on how much the Antarctic ice sheets will contribute to global sea-level rise, which is hugely important to coastal biodiversity and societies around the world.
Coastal ice-free area in Marie Byrd Land, West Antarctica (Photo: Jasmine Lee)
Figuring out how to build bio-physical models to predict how ice-free areas might change was always going to be a challenge – my excellent supervisory team has a lot of experience and expertise, but none of that included ice melt modelling, or climate change projections…
We (by we, I mean my fantastic supervisor at the Australian Antarctic Division: Aleks Terauds) reached out to a well known climate scientist at the British Antarctic Survey, Tom Bracegirdle.
Tom was excited and enthusiastic and committed to providing us with the Antarctic specific temperature projections we would need for the project.
Meanwhile, I was busy trawling through the literature. I eventually came across a method of measuring ice melt that was right for our study. Temperature-index melt modelling relies on the strong correlation between air temperature and ice melt to predict the amount of melt over a region/catchment.
A plan was formed – we would use the newly available 10km resolution AMPS (Antarctic Mesoscale Prediction System) air temperature records to determine how much melt there was in Antarctica today and then add the temperature projections to these to determine how much melt there would be in the future. Finally, we could subtract the amount of melt from the Bedmap2 layer of ice-thickness to see the extent of ice-free area change.
Lichens in an ice-free area, Mt Siple, West Antarctica (Photo: Jasmine Lee)
This was when my love-hate relationship with R really began. I spent months figuring out how to read in the thousands of NetCDF files that contained the AMPS temperature records, how to manipulate raster files with the Raster package, how to write complex loops and functions to apply the mathematical models, etc…
It was only after I had a first round of preliminary results that everyone took notice – maybe there would be a larger change in ice-free areas than we had imagined. And it makes sense right? Temperatures increase, ice melts, ice-free areas expand – very logical.
I decided it would be good to tailor the models to incorporate spatial variation in rates of ice melt (eg. north-facing slopes receive more sunlight and therefore melt faster than south facing slopes). So after some more literature searching, I decided adapting the methods of Hock (1999) to incorporate solar radiation to vary the melt rate would be ideal. The only problem was that I didn’t know of any radiation layers for Antarctica. But no worries, Ben (one of my co-authors) came to the rescue (again). He literally just ‘whipped’ up some radiation data in a number of weeks using a complex model he found that incorporates elevation, latitude, slope, angle etc.
I began to rerun the models, apparently it was far more computationally expensive now. So much so that it was taking 5 or 6 days to run a single model…
After a few weeks of much hair pulling and swearing, I had managed to get the models down to about 5hrs each! That’s how much the code could be optimised (ps. pro tip – change the default output in Raster to save heaps of memory). And then once R had finished the melt models, it was time to feed it into my old favourite, ArcGis, to measure the impact on ice-free areas.
However, it wasn’t over yet. I then reran all the models to incorporate updated layers of the current ice-free area.
And then I reran them again when we realised that we should tailor the projections specifically for the Antarctic Peninsula.
And then yet again when we realised a whole extra year of AMPS data was available (because it had been so long since I’d started…).
My primary supervisor (Rich Fuller) gave me the advice “I hope you have put really good notes in your code incase you have to rerun it again”, I replied “I am not re-running them again! I’m going to quit my PhD if I have to run them again”.
But there I was – running them again so that we could incorporate precipitation…
Eventually there was nothing else we could think of to improve upon so I got stuck into writing it up. Which I was managing to focus on because I was looking forward to an upcoming field trip. And it wasn’t just any field trip – it was to Antarctica!!! For the first leg of the Antarctic Circumnavigation Expedition (ACE), which was carrying scientists from all over the world to collect data on a research cruise via all of the sub-Antarctic islands and the continent. Like many conservation projects, all of my work was carried out behind a desk – so the chance to go into the field (to collect data for a colleagues project on microplastics) was exhilarating. We submitted the paper and when it got sent for review I was so excited I could barely wait for an appropriate hour to ring family and supervisors (I get up early for the gym). I managed to wait until about 0600 before I rang dad…
Fast forward a couple of months and I visited an ice-free area for the first time when we stopped at Mt Siple (Marie Byrd Land, West Antarctica) to count the Adélie penguin colony. Flying over the ice-free areas in a helicopter and then getting to actually stand in one (plus penguins) was one of the best days of my life. You can imagine it couldn’t get much better when a few days later I received the first round of reviewers comments on the paper and they were all quite positive. I then madly set about trying to make all the changes and coordinate with co-authors via the satellite internet on a Russian Icebreaker (actually we are lucky we had internet at all), plus juggling benthic trawling, microplastics sampling and several hours of seabird surveys a day. Fortunately I don’t get sea sick. Somehow I managed to get it all sent off and submitted within the two week deadline, though the real kudos has to go to seabird ecologist Peter Ryan – who had to do bird surveys by himself for a few days and who somehow put up with me constantly complaining about being tired…
In the field!!! And standing in a real life ice-free area - Mt Siple, West Antarctica (Photo: Peter Ryan)
The effort that went into improving the models paid off. Our results were new and startling – ice free areas could expand by over 17,00km2 by the end of the century (a 25% increase). The biodiversity impacts of ice-free area expansion are likely to be complex and substantial. The extra habitat will provide opportunities for native species to colonise new places, but it will also provide these opportunities to non-native species. Antarctica’s biggest protection from alien species is its harsh climate and extreme weather, but as it warms and conditions become more mild, then the non-natives will be able to more easily establish and then spread via the increasing connectivity between ice-free areas (some patches are predicted to expand so much that they start joining together). Because the Antarctic species are so specialised (to survive the harsh conditions), it is likely that they may be outcompeted by the invaders. So like everywhere else in the world there will be both winners and losers of climate change, and only time will tell which will be which.
However, in the mean time we can use our ice-free area models to help prepare strategies to help reduce the impacts on native species. They could be incorporated into protected area planning to help identify key sites, and they can also be used to help pinpoint sites where we should increase biosecurity – as humans are one of the primary vectors of non-native species to the continent.
It took two years to do this research, so it’s good to see it finally out there. Though really, it’s only just the beginning – now I have the other three chapters of my PhD to contend with…
You can read the full paper published in Nature here: http://go.nature.com/2spjh07