A guest post by Allen Pope
It sounds like a simple question, right? What color is a glacier? Snow is white, so a glacier is white…except after it gets pressed into glacier ice. Then, it’s a beautiful sapphire blue color (and I’m not the only one who thinks so – some artists do, too!). Glaciers can even be red where certain algae have started growing…or where debris, ash, or dust has accumulated on the surface, then they are a brownish/grayish color…and where the snow is older it will have melted a bit and so it looks a bit darker, too…
Suddenly, the color of a glacier isn’t such a simple thing to describe. Which is a problem, because the color of a glacier can be really important! Whether it is dark or bright will change how much energy the glacier absorbs and therefore how fast the surface melts. (Melting can sometimes darken the surface, enhancing this effect. Cool articles on that from NOAA and Meltfactor!) And if we are able to tell apart the slightly different colors of new snow and year-old snow, it provides a powerful tool for being able to measure the size of the glacier’s accumulation area, a value VERY closely related to the glacier mass balance.
My research is focused on getting more of a handle on the “color” of glaciers, not just in the visible wavelengths but also in the infrared. I compare measurements from the field to airborne and satellite imagery of the same glaciers (lots of pretty satellite images from NASA Earth Observatory). In this way, I am looking into what types of data will potentially be able to help us best monitor glaciers as well as how to take advantage of the satellite data record that is already archived.
Fieldwork pretty much entails walking around on a glacier’s surface and pointing our sensor at the snow or ice surface (my fieldwork blog from Svalbard; photos from Iceland 2011). By comparing every surface measurement with a reference panel at the same time (image above), we make sure that all the measurements are intercomparable even if light conditions are different. We record the reflectance of the surface of a range of wavelengths as in this graph.
Then, we obtain satellite images, each of which contains information in not only the red, green, and blue, but also near infrared, and shortwave infrared. Each image is broken up into “classes” of groups of similar pixels which are then averaged together. Here’s a shot of what this looks like on my computer screen.
Finally, we compare the field data with the satellite data. In this graph (below) the continuous lines are the field data and the bars are groups in the satellite imagery. You can see immediately that some bars and lines match up well while others do not. There are many reasons for this, the most important of which is that while field data are measured at a single point, the satellite images measure pixels which are 30 meters across – so there’s a lot of mixing which takes place! My job is to figure out the best way to match the lines and the bars so that the field data can explain the satellite data.
My PhD is still in progress, so I can’t give you any of the final answers yet. I presented my intermediate results of satellite and in situ measurements at a very large conference in December, and now I’m moving forwards with some fresh ideas! I’m looking forward to getting a good handle on the remote sensing techniques and applying them more widely. Until then, keep checking out GlacierChange.org
Allen Pope is a PhD student studying the remote sensing of glaciers at the Scott Polar Research Institute in Cambridge, UK. You can find out more about him and his research at about.me/allenpope or follow him on Twitter @PopePolar.