[WIP] Openlem integration

[MFraters]

This is a very early state of openlem integration, and not ready for review. There is still a lot of work to be done, including (and not limited to) figuring out the conversion from the non-dimensional to dimensional case, coupling the uplift correctly, coupling the sealevel correctly, rethink the grid coupling, think how to deal with the macros, actually adding the right parameters as input files and a lot of clean up. This is mostly so that other people can already take a look at what is going on if they want to.

OpenLEM is written by Stefan Hergarten, who I am collaborating with. The source code and some documentation can be found here: http://hergarten.at/openlem/.

From a technical perspective, on a high level, openLEM can be seen as a c++ header only library. So functions can be easily called by just including it, and there are no dependencies. I also think that it will be possible to have a deep integration with openLEM, making some interesting things possible such as changing the erodibilty value based on the the rock type in aspect. I had to put it into a namespace to make sure that it doesn't overlap with aspect and dealii classes. It is heavily reliant on macros for turning on and off functionality. Like fastscape, due to the fast streampower law algorithm, this is a inherently sequential algorithm. So I put everything on mpi rank 1 and process it.

From a model perspective, it includes the shared streampower law, ice erosion, a whole orographic precipitation code and a few more other features. Currently only works in a cartesian geometry, but a spherical coordinate system is something we plan to implement soon. There is currently also no real marine sediment transport, but that is also something we are working on.

OpenLEM is gpl v3+, so I don't know if that is an issue with integrating and shipping it with aspect. I can also discuss that with him if needed.