Spaced-OoooO-Out ::

With the paper CLustre: semi-automated lineament clustering for palaeo-glacial reconstruction now accepted for publication, we have released the sourcecode for CLustre at Github.

When undertaking a palaeo ice sheet reconstruction using geomorphological evidence, it is standard to map the ridge crests (and often outlines) of glacial landforms and in particular drumlins. These form coherent “flow sets” that indicate discreet periods of ice flow and are manually grouped together based upon spatial homogeneity. My PhD thesis reviews this process: Chapter 2 - Overview Of Ice Sheet Reconstruction From Geomorphological Evidence.

CLustre undertakes an objective and repeatable approach to performing a classification and grouping of drumlins, using similar heuristics to manual mappers. The paper presents the algorithm, uses a Monte Carlo method to test it on a simulated dataset before applying it to two case studies. The code is based upon the open source Python and can be run directly from the command line or within a console…… a sample dataset from Stokes and Clark (2003) is also provided for testing (and should be cited if used).

Please feed back any comments!