Randall J. LeVeque

We had to start printing `num_ghost` when we added the binary output option in amrclaw/geoclaw, since the binary dump is the entire `q` array, including ghost cells. So we need...

Binary output has been in for a long time but I recently (in the commit @ketch referenced) tried to streamline things and use the same read_t function in all repos,...

It turns out that it's actually useful to have the ghost cell values in the output for some applications, in particular if you want to compute the vorticity of the...

Many of the other things originally discussed in this issue have already been implemented, such as having a documentation branch v5.9.x and having the OSF version. So I think it's...

Agreed -- I did something along these lines recently for a project where we had a file with bathymetry values only and no topography, in which case we wanted to...

Yes, filling in would probably be better provided the coarser topo is sufficiently fine that it does something reasonable near the edges. And this would be a bit hard to...

@cjeffr: This is probably a result of the changes in #561, which was included in v5.9.1. I think this change should have fixed a bug in earlier versions where fg.x...

Maybe it could just be done in the Riemann solver along with all the other checks done there? I'm not sure how often the depth goes negative. I suspect for...

Sounds very useful as something to add to geoclaw. The one other place I know of where xarray is currently supported is in [topotools.read_netcdf](https://github.com/clawpack/geoclaw/blob/acac99becfaa10ac33a344086e00832a70455f15/src/python/geoclaw/topotools.py#L1742) where a topo DEM can be...

I suggest also adding `Z` and rearranging the arguments so `topo_type` is second to agree with e.g. `read` and `write`. (unless this will break old code where the position of...