Possible problem with fgmax bilinear interpolation?

[rjleveque]

When updating $CLAW/apps/tsunami/bowl_radial_fgmax I noticed that setting interp_method = 1 gives some garbage around the edges of the fgmax grids. (https://github.com/clawpack/apps/pull/99)

Need to investigate.

Note that setting interp_method = 0 for pw constant interpolation from FV cell centers to fgmax points is generally recommended and seems to work ok. Bilinear interpolation has problems near the shoreline in general if the depth and topography are interpolated separately. But the new issue see is away from the shore.