Surface hydrology of the Greenland Ice Sheet

The Greenland Ice Sheet (GrIS) (Sermersuaq in Greenlandic) is the second largest body of ice in the world. It is almost 2,400 km north to south, and 1,100 km wide, covering a total of 1,710,000 km2, approximately a fifth the surface area of Canada. At its deepest point, the ice sheet is 3.43 km thick, and in total it contains 2,942,393 km3 of ice, equivalent to a potential 7.4 m of global sea level rise and 10% of the Earth’s freshwater resources.

Although not without spatial and temporal complexity, there has been a generalized warming trend over Greenland in the last three decades. This warming has been widely associated with increased meltwater production, culminating in 2012 when almost the entire surface of the ice sheet underwent melting.

Common sense suggests that if temperature rises over the Greenland Ice Sheet, more melt will occur, and more runoff into the ocean will cause an increase in sea levels. If this relationship were straightforward, we could build a simple relationship between temperature and ice mass loss, and hence predict sea level rise. However, like all glacial systems the mass loss in the GrIS has a complex and non linear relationship to climate. Conceptually, that means that we cannot say that X degrees of temperature increase results in Y meters of sea level rise; the system contains certain responses such as thresholds and feedback responses which complicate the relationship between climate and mass loss.

One of these complicated relationships is that meltwater produced on the ice surface does not simply drain off the sides – it actually enters the ice sheet through near-vertical pathways called moulins and is able to access the bed of the ice sheet. There, basal (bed) meltwater can, under certain circumstances, increase pressure at the bed of the ice sheet, reducing friction and allowing the ice to flow more easily. This means that ice travels to lower elevations where it can a) melt more rapidly (because temperature is generally higher at lower elevations), increasing mass loss, or b) calve off into the ocean (if the outlet glacier terminates in a body of water), again increasing mass loss.

However, this relationship is again not very straightforward. It turns out that throughout the melt season, the subglacial hydrology of the ice sheet evolves to be able to convey draining meltwater more effectively, and ice speed and melt begin to lose their relationship to each other (they become decoupled). For these reasons, it is very important that we understand where, when, and how much meltwater from the surface reaches the base of the ice sheet.

But… we don’t know that. Not really. The glaciological community has a lot of assumptions that we have embedded in models to try to simulate moulin formation, but we don’t have a very good grasp on where moulins form, and how their distribution affects the where, when and how much of meltwater penetration into the ice sheet. In my research, I attempt to build insight into this problem by investigating controls on where moulins form and how this affects the spatial variability of surface hydrology on the GrIS.

What specifically am I researching?

My research focuses on the south west of the GrIS, where there are a lot of supraglacial (on top of the ice sheet) hydrological features (lakes, rivers, etc…), and where there has been the most previous work on the topic. I use remote sensing methods to answer my research questions. I use imagery and digital elevation models to identify moulins, river channels and their catchments on the ice sheet, and I then use geospatial methods to investigate the spatial distributions of these hydrological features.

What have I found out so far?

  1. If we want to study surface hydrological features (such as lakes and streams), we need to be able to figure out where they are, and we will mostly do this using remotely sensed data. That is actually not that straightforward to do! There are various approaches to doing it, each with its own limitations. In my contribution, I add one more approach to the toolkit, and compare my results to a dataset from Laurence Smith’s UCLA lab group. You can read the paper here. 
  2. My contribution so far is also methodological. I compare different methods of remotely determining how much water is supplied through supraglacial rivers to their terminal moulins. This is important because the how much, where and when of moulin discharge (volume of water in some given amount of time) is important for predicting how ice sheet velocity will respond to surface melt patterns. However, we don’t currently have good, generalizable methods for remotely estimating moulin discharge over large areas. This chapter is published as a Letter in the Journal of Glaciology