By Matt Chernos
Bridge Glacier in the summer of 2011. Photo by L. MacKenzie
Bridge Glacier is located in the Pacific Ranges of southwestern British Columbia, Canada, between the Coast and Chilcotin Mountains. The glacier is an outlet of the Lillooet Icefield, about 175 km north of Vancouver. Beyond being the source of my MSc thesis, Bridge Glacier is important because it is the source of the Bridge River, and the Bridge River Hydroelectric Project, which supplies British Columbia with 6-8% of its electrical supply, leading some to call it a giant “melting battery“.
Bridge Glacier is also an interesting study because it sits in a lake, known unofficially as “Bridge Lake”, allowing calving to be a significant source of ice loss from the glacier. This wasn’t always the case though. Until the late 1980s, Bridge Lake was relatively small and shallow. Since 1991 though, Bridge Glacier has retreated substantially, and has allowed the lake to triple in size (up to 6 km^2 in 2012). As we have already seen on this blog, and throughout scientific literature, glaciers that flow into large bodies of water retreat somewhat less directly than the traditional response of warmer temperatures driving retreat. The question is, how much does the lake actually affect Bridge Glacier’s retreat?
One easy way to figure out glacier retreat rates is to check Landsat Imagery (available easily through their aptly named “LandsatLook Viewer“). Satellites capture images roughly every 16 or so days. Unfortunately, if you’re studying an area that is pretty cloudy, 80% of your images will be blobs of cloud (which can look really cool, but doesn’t really help reconstruct glacial histories). Images for Bridge Glacier go all the way back to 1972, with a few gaps in the data due to clouds and low temporal resolutions in the 70 and early 80s. By collecting an image from late September every year, I was able to compare two sequential images and measure how far the glacier retreated. Putting all the years together, here is an animation of Bridge Glacier 1972-2012.
Bridge Glacier 1972-2012. If the animation isn’t running on your browser, click the picture.
And again putting into a “science form” in a graph (below):
(a) Bridge Glacier cumulative retreat (b) Winter Precipitation Anomaly (snowfall) (c) Summer Temperature Anomaly (d) Mean Annual Flow Anomaly (Bridge River Flow). All figures explained in the text.
To get a better handle on the graph, (a) is the cumulative retreat of Bridge Glacier from its position in 1972. Cumulative values are often seen because individual years have a lot of variability, which can distort what is otherwise a really clear trend of retreat. (b) is the winter precipitation anomaly (November-April inclusive). Anomalies are often used in climate science to clearly illustrate whether any given year was above or below average. For this graph, average winter precipitation (snowfall as water equivalent to eliminate snow density discrepancies) was taken for 1972-2012. (c) is the summer temperature anomaly (May-October). Again, positive (red) values indicate hotter than normal summers. (d) is the mean annual flow of Bridge River at the edge of the glacier. Mean Annual Flow (MAF) is a measure of how much water passes through the Bridge River, and is an excellent indicator of how much melt there is in any given year.
The first thing that jumps out to me in this graph, is that the retreat rate is ‘step-like’. Until 1991, the glacier is retreating pretty consistently, but that summer, things become more irregular. 1991 and 1992 have fairly large retreats, but then the glacier stays pretty constant for a couple years until it retreats again dramatically in a couple bursts. In fact, most of the retreat since 1991 occurs in a couple years of large bursts 0f dramatic retreats.
One of the largest retreats occurred in 2005 between July 27 and August 15:
July 27, 2005
August 12, 2005
September 5, 2005
Hopefully this photographic evidence is convincing enough that this ‘step-like’ or ‘staircase’ retreat is a product of calving as the glacier terminus is destabilized by the lake that it is floating in. The question that remains, though, is how much of this can we blame on Bridge Lake, and how much of it is a product of a changing climate.
To do this, we can run a numerical experiment: How would Bridge Glacier have retreated if Bridge Lake wasn’t there?
We’ve already discussed how a glacier’s change in length is a product of changes in temperature, buffered by its individual climatic sensitivity, and the period it takes for this change in temperature to fully affect the terminus, known as the response time. If we calculate these variables for Bridge Glacier, and combine it with temperature data, we can predict how the glacier would have retreated had it not been for the lake.
Modeled retreat for Bridge Glacier. The LOESS line is the modeled retreat rate smoothed to even out the statistical ‘noise’ of temperature data.
What this shows is that until 1991, climate projected retreat is very close to what we actually observed at Bridge Glacier. However, from 1991 onwards, once Bridge enters the ‘staircase’ stage, it retreats a lot faster than projected. If we go back to the animation of Bridge Glacier’s retreat, we can see that 1991 is, not coincidentally, when we start to see icebergs in the lake. All of this suggests, once again, that once the glacier begins to calve icebergs, its retreat becomes less a product of warm temperatures, and more of calving events.
If calving is in fact driving the accelerated retreat of Bridge Glacier, as this model suggests, it isn’t all bad news. Eventually, the glacier will retreat far enough back, and high enough, that it will no longer spill into the lake. Once the glacier reaches this point, it should, once again retreat more in line with the gentle climate-controlled retreat (albeit still retreat!) projected in the model. Once it does so, it will have ended its calving phase, and will once again respond more directly to climatic trends.
Matt’s Blog: Bridging The Gap