Testing different spatial grid resolutions

[glugeorge]

I will conduct 3 series of tests on spatial resolution:

1. Changing coarse grid resolution
2. Changing fine grid resolution
3. Changing the point where the fine grid meets the coarse grid

I am unsure if it is that valid to change 3 much, especially because it seems to be involved in properly resolving the boundary layer. I will need to do some additional calculations to figure out where this transition occurs.

[glugeorge]

From experiment 1: Changing the coarse grid resolution has some impact on the results, shown below:

In the big picture, there isn't much difference, but if we zoom in, the differences are on the order of a couple hundred meters:

Also I find that it's not necessarily a monotonic increase in accuracy (the order of the colors top down are not the same order as the resolutions in the legend)

[glugeorge]

On the other hand, the fine resolution has a huge impact, notably on the grounding line position:

Here the changes are monotonic, and appear to converge as the resolutions get higher:

[jkingslake]

This all looks good. A plot of grid spacing against the final GL position would show the convergence nicely.

[jkingslake]

I find it a little confusing that the right hand plots are labelled 'initial steady state'. They aren't the initial conditions of the simulations are they? They are the final state right?

[jkingslake]

Also maybe you can move your search to larger grid spacing for the upstream grid to see if there is the usual pattern of convergence in that part of the parameter space.