Duncan Wither

Home Notes Github LinkedIn

Basins

Written: 12-Nov-2024

Project was to look at identifying different dranage basins in scotland. The primary purpose of this was to overlay it on a map, and then see if there were some interesting mountains or ridges in it. Or to see if it made any mountains or ridges more interesting, which I think it did.

Result

We’ll skip the foreplay and head stragt to the maps:

Link to the high res version.

Colours go from yellow, to green, to blue, to black, to thick black in increasing importance. How that importance is decided is the Pfafstetter Level, but I don’t really know what it means, nor how they decdie the level for different rivers going into the sea. It doesn’t matter hugely beyond it appearing to mean bigger river -> more important.

Interesting Points

So there are a couple interesting points, but the one that jumped out to me was the triple points. There are a couple, but the one that jumped out at me was the one between the Spey, Tay and Dee: Carn an Fhidhleir. This is a munro! And it even has a wikipedia page1, even if it is a touch barren.

To the north-west lies the River Feshie, which flows into the Spey at Kincraig. To the north-east lies Geldie burn, which runs direcly into the Dee in the middle of nowhere. To the south lies Tarf Water, which feeds the River Tilt, which feeds the River Garr, to River Tumel which (finally) feeds the River Tay. So very exiting.

There are a couple other triple basin points I thought it worth highlighting, but I’ve not yet checked to see what is doing the dividing:

Note the above co-ords are from the maps, so have limited precision.

Fortunately (or otherwise) it would appear to good people at hill-bagging.co.uk have already done all of this. So if you want to see their list of triples (that you can filter on) then go here.

Regardless, I got to mess with some map data, and plot some pretty graphs, and mostly find what I was looking for.

How

So this was done - with limited intelligence - by using the HydroBASINS data from hydroSHEDS. The matlab script was highly manual (there is a reason it’s still called test.m), but is Linked. The key thing I did was to first work out which basins were inside a bounding box which equated to Scotland (and a little of Northern Ireland + England), and then plot them. It took multiple overlapping plots to do this, but given the limited scope of my interest I thought it better to go quick, dirty and highly manual than some all singing script I only used once.

Failed Attempts

So I had a couple failed attempts I wanted to touch on breifly. One was using some elevation data from here

The UK Government also puts out some good lidar data.

And OS also puts out elevation data

Plotted out nicely:

After spending a good afternoon working on wrangling the data I fond the following: https://appliedsciences.nasa.gov/sites/default/files/Riverbasin_Final.pdf Which in turn pointed me to: https://www.hydrosheds.org/products/hydrobasins Which is what I ended up using.

Some (smart) people wrote a paper doing what I thought would be an afternoons work. Well, they did a bit more, but there was more of them, and they got paid to do it3.


  1. If I link to this website as a source does that means it’s not original research?↩︎

  2. Well, the firth of forth, via Lothian’s River Almond, not Perthshire’s!↩︎

  3. Hence why I think they’re so smart.↩︎