Kyle Mandli

Other than this seems to have been an issue for awhile. I am not even certain why the book would depend on [unicode-math](https://github.com/wspr/unicode-math) given what the repos states that it...

I invited @ketch, @rjleveque and @maojrs to contribute something on their [book](https://github.com/clawpack/riemann_book) for which I believe they will issue a PR for soon.

I am a bit confused now as to how this abstraction plays with the `State` object. I figured that the `FunctionSpace` object would always be present in a `State` object...

Ah, I see, I was confused by the variable naming.

I agree, it would really make this code look nicer and be easier to read.

We could switch to the same logic in AMRClaw and GeoClaw for gauges which circumvents this problem.

The primary logic starts at https://github.com/clawpack/amrclaw/blob/master/src/2d/gauges_module.f90#L398 and uses bilinear interpolation. There is a bit of complexity due to AMR that we would not need to worry about.

Reviewing, I will try to send some feedback by tonight. One thought though after reading through the PR once was that there are a couple of spots where there are...

I think this generally looks good. I would suggest that tests be added since this is a fairly complex procedure, especially in parallel. I also wonder if this functionality should...

@aymkhalil You make some good points and the core routine, `downscale_local_mean`, is really the one we would most likely want to replace. As a particular test case of what we...