raphael dussin

yeah that does not work great, I wonder if it would be better with a combination of xarray.bfill and xarray.ffill along the axis perpendicular to the axis that varies most...

Hi all, I am a new MITgcm user. I have modified my optfile (darwin_amd64_gfortran) to add this : `if [ "x$NETCDF_ROOT" != x ] ; then` `INCLUDEDIR="${NETCDF_ROOT}/include"` `INCLUDES="-I${NETCDF_ROOT}/include"` `LIBDIR="${NETCDF_ROOT}/lib"` `LIBS="-L${NETCDF_ROOT}/lib...

add `export MPI_HOME=/wherever/openmpi/lib/is/installed` to your .bashrc and you're good to go! a lot of package manager properly set up pkg-config to point to where the libs are installed. too bad...

well it wasn't working for me and I needed to edit the optfile. The new solution offers the option to set explicitly the path for mpi else use pkg-config. it's...

@christophernhill I think that sounds like a plan. I got the code from Rachel Stanley and will look at the original implementation and how to fit it into MITgcm. Thanks...

is the question related to the folding of the north boundary? Which would basically be a face connection of a face onto itself with direction change from one side of...

that's exactly right, velocities reverse directions between the 2 parts of the fold. You need the information on the pivot point (index and point type). The most natural choice is...

I can give it a go, but cannot really commit to a timeline. One approach would be to create a function, say: ```python def create_north_fold(grid, pivot_index, pivot_point='corner') ``` that returns...

count me in!

same (username: raphaeldussin)