I started my PHD in July 2010. On March 12, 2016, the paper that summarizes over five years of work was finally published. Before going on what it is about, you can download your own copy of the paper from this link until May 1st. After that, it is behind a paywall, but feel free to email me if you want a copy. I am always happy to help out, don’t be shy!
A childhood dream
When I was a kid, I was totally obsessed with Earth sciences. I had notebooks full of dinosaur names and facts. I had a big rock collection! I completely wanted to be a geologist or paleontologist when I grew up. When my parents went on a trip to Las Vegas, the gift they got me was a collection of “The rocks and minerals of Nevada” (I actually still have that). Sometime after the age of about 8 or 9, I sort of grew out of those ideas.
I didn’t really get back into the Earth sciences until I randomly took an introductory geology course as an undergrad student as an elective. I was completely enthralled, and remembered the love I had of the Earth that I had as a kid. The class was taught by Jim Teller, an expert on glacial Lake Agassiz.
When I was growing up, I went to school on top of one of the largest features in the post-glacial landscape in North America – the Upper Campbell strandline. It formed when there was a large lake known as Lake Agassiz at the south end of the Laurentide Ice sheet. This lake existed because the natural northward flow of water was blocked by the ice sheet. When I was a kid, I played in the gravel pits of this wave sculpted piece of the ancestral Assiniboine River delta, finding interesting rocks and fossils. I always wondered, how did this come to be? Of course, everyone here knew that this was the shoreline of Lake Agassiz, locally known as the Arden Ridge. I knew the Arden Ridge stretched way to the south, and that you could drive a long way to the south on top of it. I remembered the dip in topography just behind our elementary school. I always thought this was really odd. I also wondered why there was another ridge about 1 km to the east, and other smaller ridges further on the road to my parents’ house.
So I went to Australia, and found out about these things.
Going to Australia
I started my PHD in 2010, after it was pretty clear that the downturn in the mining and oil industries was unlikely to improve after the 2008 market crash. I can’t say I regret anything, as the resource sector has not really improved since then. I decided to continue along the vein of my masters’ degree, and study ice sheets and the deformation of the Earth caused by their weight (known as glacial-isostatic adjustment). The Australian National University (ANU), where I did my PHD, is well known in this field, because of the influential work of Kurt Lambeck and his colleagues. Alas, he had just retired at that point, so I was supervised by Paul Tregoning. Although I had initially applied to the ANU on the basis of doing work in Antarctica, when I arrived they asked if I was interested in looking at the Laurentide Ice Sheet. I’d like to thank Paul and all my other collaborators on this project for giving me the resources to complete it!
The first thing I looked at were the datasets available that gave indications of when ice free conditions existed in the Laurentide Ice Sheet, and the vast collection of relative sea level indicators (i.e. sea level was higher in the past in places the ice sheet covered, due to the weight of the ice sheet pressing down on the Earth and causing the hot rocks deep in the mantle to flow away). Somewhat unconsciously, I started by focusing on the western half of the Laurentide Ice Sheet, a place that was quite controversial in previous reconstructions of the ice sheet. I finished my mid-term examination in November 2011, and my thesis committee suggested I continue to focus solely on the western Laurentide Ice Sheet so that I can finish in a reasonable length of time.
The first part of my PHD that got published was a paper entitled “An assessment of the minimum timing of ice free conditions of the western Laurentide Ice Sheet” in 2013. In that paper, I looked at a combination of the inferred direction of ice sheet retreat (i.e. where the edge of the ice sheet was as it was melting) and the chronological data from a variety of sources to make a map of how certain it was that there was no ice at a particular time. This served as a template of where I put the ice margins through time in my final model.
In the end, I managed to get my PHD thesis handed in for examination in April 2014, some three years and nine months after I started. This was pretty tight, you are free to read about my experiences on the ANU Earth Science student blog, On Circulation.
The paper (linked in the first paragraph) just came out, but I started writing it soon after the corrections to my PHD thesis were completed in the middle of last year. It is basically a summary of what I did in my PHD. I have another paper that is currently under open review that describes the software I wrote to produce my ice sheet model. I may do another paper regarding some of the data, but that will depend on whether or not I get time to write it. There are basically two parts to the paper, a description of the glacial-isostatic adjustment data (i.e. observations of how the weight of the ice sheet deformed the Earth), and the resulting ice sheet model itself.
The data consisted of observations of modern uplift from permanent GPS stations, the tilting of lakes within the area I studied, the tilt of continuous, well dated glacial lake strandlines (i.e. beaches), and relative sea level data. I used 11 GPS stations, which have uplift rates up to 12.5 mm/yr. One of the big disappointments is how few permanent GPS stations there are in Canada, considering it has one of the most substantial uplift rates in the world. For the lake tilts, I took lake level data from lakes that had at least two gauges in them, and calculated the change in height over time. Some lakes, like Lake of the Woods and Lake Winnipeg are large enough that there is a measurable change in lake level between different ends of the lakes over the period of years because one side is going up faster than the other. The glacial lake strandlines (i.e. beaches, scarps) are a pretty new constraint, at least as far as Lake Agassiz is concerned. Four levels of Lake Agassiz have been both well mapped and well dated to be useful in my analysis. The largest of those strandlines is the Upper Campbell, which ran right through my home town, and there is about 150 m difference in elevation between the north end in north-central Saskatchewan, and the south end near the intersecting border of North Dakota, South Dakota and Minnesota. That “dip” I mentioned earlier is actually the physical location of where the shoreline was. Finally, I used a collection of database of relative sea level indicators from the Arctic and Hudson Bay coasts. Since most of this area is going up, sea level has been falling since the ice melted away.
The second aspect of my paper is a description of the actual model, and how well I was able to fit the database of observations of glacial-isostatic adjustment. The model I made had a specific set of Earth model parameters (which I will not elaborate on here), of which the entire reconstruction is dependent on. I set up a series of timesteps, ranging between 2000 and 500 years for the past 30000 years. Here is an example of the ice thickness and topography of the ice sheet at 20000 years before present:
As you can see, the area in the core area of the ice sheet, located southeastern part of Northwest Territories had 4000 m of ice covering it! The topography of the ice sheet is not that high, because the weight of the ice sheet caused the earth to be pressed down about 1 km in that area. The ice sheet gently sloped downwards south of that. There is a ridge of ice that sort of goes around Hudson Bay, which stands in contrast to some other ice sheet models that suggest a dome of ice existed over Hudson Bay. My model has way less ice that previous reconstructions, adding to the so called “missing ice” problem, where there is not enough ice to account for the large drop in global sea level that reached its minimum about 20000 years ago.
I developed the ice sheet model on the basis of fitting observations of glacial-isostatic adjustment. Above shows one of those comparisons, with the Upper Campbell strandline. The measurements are basically the elevation of points along the strandline, from North Dakota, through Manitoba and into Saskatchewan. The modelled value is the amount of elevation change from the age of when the strandline formed (about 10500 years ago), relative to a point at the south end of the lake. Ideally, you would want the measured and predicted values to be the same, or along the black dashed line. I adjusted my ice sheet model to try and do that. I have also plotted three other ice sheet models to show that they do not achieve this equivalence.
Why is this important?
Of course, to the non-scientist, it is important to get across why this is important. Here are some points:
- This ice sheet model can be used as an input for climate models, since it was constructed without invoking climatic parameters.
- The modelled modern-day uplift from my model can be subtracted from satellite gravity data (the GRACE satellites) to determine changes in the amount of water on the surface of western Canada for the past 15 years. There’s been some major floods in western Canada in recent times, but the gravity data are contaminated with the effects of glacial-isostatic adjustment, so it makes interpreting difficult. I’m sure someone will do it!
- One of the sudden climate change events in the past 20000 years was the Younger Dryas event, which happened about 13000 years ago. One of the recent hypotheses of the cause of this event, a rapid draining of Lake Agassiz into the Arctic Ocean, is deemed to be unlikely, given the requirements to match the tilt of the Lake Agassiz and Lake McConnell (another glacial lake) strandlines. I always like to tell people that the Earth’s climate system is probably far more sensitive than we currently assume. The fact that we have not unequivocally determined the exact cause of the Younger Dryas after decades of research is a testament to this.
- The ice sheet reconstruction I presented has quite a bit less ice than previous ice sheet reconstructions. This result adds to the mystery of where all of the ice was to account for the 120-135 m of global sea level fall from 19000-26000 years ago. People are still looking!
- The western Laurentide Ice Sheet was likely a key contributor to the rapid (10-20 m in a few hundred years) sea level rise event known as Meltwater Pulse 1a.
You can download the ice thickness and paleo-topography tiles in the downloads section of this website.