Copalis Beach exercise1

Copalis Beach exercise1#

From the GeoClaw Tsunami Tutorial

See CopalisBeach examples for more about the Copalis Beach location and a list of other examples and tutorials based on this location.

The directory $GTT/CopalisBeach/exercise1 contains GeoClaw setrun and setplot functions similar to those used for example 1b in Copalis Beach example1, with the following changes:

  • There is only one Makefile that specifies OUTDIR = _output and PLOTDIR = _plots.

  • There is only one setrun.py that specifies

      clawdata.num_output_times = 6
  clawdata.tfinal = 1.0*3600.

  amrdata.amr_levels_max = 6

    So it is only refining to Level 6 (3 arcsecond) and only running out to one hour of simulated time.

To run this code#

You should first try running this code as-is by following the steps listed in:

For more information about how the code is run and the resulting plots are specified, see:

To compare to the sample results#

You can compare your results with the archived sample results if you first fetch those using:

$ python fetch_sample_results.py

Compare the plots you made in _plots to those in sample_results/_plots.

See also

  • Plot the AMR solution along a transect illustrates how to plot a single frame of the solution and extract values along a 1D transect from the full AMR solution. This is a rendered version of the notebook plot_transect.ipynb.

Exercises#

Warning

Do not modify the code in this directory. You should copy this directory to your own working directory $MYGTT, as explained in Make your own copy before running examples or notebooks, and then make modifications. Otherwise you may run into merge conflicts when you try to update the tutorial repository with a git pull!

You may want to make several copies of this directory to experiment with different modifications.

If you run into problems running the code after moving it, see the suggestions in Running/debugging the Copalis exercise1 code.

Here are some tings you might try modifying in this example:

  • After copying this example elsewhere and running it, also run “make plots” to create a webpage of results in the new directory and check that the plots look the same as those in $GTT/CopalisBeach/exercise1/sample_results/_plots.

    Then also use the script compare_gauges.py to compare the gauge results in more detail by plotting the results from your run on top of the sample results.

  • Try modifying the example to use a different set of AMR levels and refinement regions to get down to the same resolution and see how this affects the results and running time. In particular, the example is set up to use the first 6 levels from an AMR structure with resolutions

    # dx = dy = 4', 2', 24", 12", 6", 3", 1", 1/3"

    Try modifying refinement_ratios so that it uses fewer levels with:

    # dx = dy = 4', 1', 12", 3", 1", 1/3"

    Note that you will also want to modify the minlevel and maxlevel values for some flagregions in order to impose the same grid resolutions as originally in these regions.

    Modify compare_gauges.py to see how much this affects the solution at the gauges

Hint

If you run compare_gauges interactively in IPython via:

$ ipython --pylab
In [1]: run compare_gauges

then you can zoom in on the plots to see any differences more clearly.

  • Do more comparisons of different resolutions, refinement ratios, and/or refinement flagregions.

  • Add one or more new gauges in the computational domain.

  • The domain used in this example does not cover very much of the ocean, not even the entire region of seafloor deformation defined by the dtopo file. Try enlarging the computational domain to see if that changes the results. (What happens if you enlarge it beyond the region covered by the topo files being used? Remember that the kml files produced by “make data” can help you see how the domain relates to the topo and dtopo files.)

  • Try shifting this example to a different coastal region. You may need to download different topo files if you move very far.

Contents