GitLab

Allows to connect HTUs if the river has to flow up-hill by at most 1m. This is only possible if the flow accumulation and hierarchy of the river are respected. This seems to be needed because of some non-consistencies of topography within MERIT and the small size of HTUS in SAFRAN/MERIT combination.

Some corrections which improve the link-up of the HTUs on high resolution grids. The issues arose on the SAFRAN grid.
They are not all yet solved.

Adds the index of the HTU which is considered representative for the grid. It is currently the HTU with the largest upstream area. This will allow 2D diagnostics in ORCHIDEE.

Adds the bounds to the graph file for all grid points and not only the land points. This will facilitate the interpolations with CDO.

This is the end of the newdivbas branch

See merge request !2

Better treatment of dz when computing for the HTU. It is now limited to being larger or equal to zero. The limit is only applied when computing the topoindex.

The MPI exchanges are less demanding on memory and the code can now run with HydroSHEDS on large problems.

Correction when looking for the outflow position when the HTU is artificial (no Hydro data available for the grid point)