Tag Archives: sediment fingerprinting

G’day from down under!

I’m Sophie, a CECHR PhD student currently on placement studying at the University of Tasmania, Australia (UTAS). I am only a few days in but rapidly finding my feet thanks to being welcomed into a great department- and of course the good weather helps. Ordinarily I’m based in Wexford, Ireland studying sediment dynamics in agricultural catchments. My project is funded by the Walsh Fellowship Programme, Teagasc and collaborates with the University of Dundee/CECHR so really, I get the best of both worlds. This specific placement has been supported by the Australian Bicentennial Scholarship Fund from Kings College in London, UTAS and University of Dundee. I’m here to work on some exciting research so there’s a bit of background below to explain exactly what my project is, why I’m doing it and how my placement fits into my overall work programme.


The University of Tasmania

Sediment is an important and often underestimated pollutant in watercourses. It comes from eroded soils from a combination of places upstream. If we think about intensive agriculture, big bare fields which have just been seeded with a new crop may be at a high risk from heavy rain which causes soil particles to be washed in stream. The same can be said for rivers where stock can access the stream, the channel banks can be trampled causing sediment to move into the stream. However, the story isn’t quite so simple… There are many alternative sources of sediments such as naturally eroding channel banks which activate when river levels are high, and the edges of roads that can get eroded when vehicles pull in. My work is to identify where sediment is coming from and compare agricultural and non-agricultural sources in order to inform effective mitigation strategies and support increasing level of production, under increasing challenging climatic changes whilst protecting water quality.

It is difficult to measure soil loss from each source individually for a whole river catchment but what we can use a technique called sediment fingerprinting. This method can identify the contribution of sediment from the potential source areas of a mixed river sample based on the natural characteristics of the sediment particles. So the idea is, we can sample all potential sources and run samples through a range of instruments to get data on the geochemistry of the sample, the radionuclide assemblage and mineral magnetic characteristics. Each potential source area e.g. field topsoil, channel banks should have a different set of characteristics or ‘fingerprint’. We can then compare these source area fingerprints to a sediment sample collected in-stream which is assumed to be a mixture of potential sources. This river sample can be statistically ‘un-mixed’ in order to trace back the origin of the sample. So after two years of project planning, data collection and lab processing I can finally spend some time brushing up my stats skills and work on this component.

Fortunately, an un-mixing model already exists which not only allows us to report how much sediment is coming from where, but can give an uncertainty estimation too. And that is why I’m here as it was co-written by Prof. Stewart Franks at UTAS. Over the next three months, I’ll be keeping you posted on developments. Feel free to comment, ask questions etc.