Skip to content

Automative approaches for capturing the localised tsunami response

License

Notifications You must be signed in to change notification settings

DanGiles/Localised-Tsunami-Response

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

12 Commits
 
 
 
 
 
 
 
 

Repository files navigation

Localised Tsunami Response

The localised tsunami response can be captured using two independent methods: transfer functions and mlp. Each approach has its own folder.

=====

1. Transfer Functions: Extended Green's Law

Python code with C bindings which calculates the tsunami amplification factor (beta or alpha) for various sites. Currently set up to deal with the alpha (Lalli) formulation only. The formulations of the extended Green's Law were introduced in [1] and [2] respectively.

Compiling the C bindings

The gradient descent approach which optimises for the beta/alpha values is written in C. In order for these to work with the python code, one must first compile the C code into a shared library. Please run the following command in the src folder:

cc -fPIC -shared -o cfuncs.so gradient_descent.c

File Structures

A main.py file is located in the main folder. The various parameters associated with the optimisation are specified in the params.py file.

Workflow

  • coarse_forecast.py : Takes the maximum wave heights from the coarse forecasts and uses Green's Law to forecast to a specified highlim.

  • coarse_to_fine.py : Interpolates the maximum wave height on the coarse grid to the deepest point in a fine grid.

  • alpha_calculate.py (contained in fine_forecast.py): Optimises for the alpha parameter and forecasts for the maximum wave heights in a fine grid.

  • gauges_isobath.py : Calculates the wave heights at the gauge locations in the fine grids, calculates the relevant errors and then plots for each site.

Outputs

The code will output the optimised alpha values for each site along with the forecasted maximum wave heights.

=====

2. Multiple Layer Perceptron

A neural network approach, which utilises keras to capture the localised response is contained in ./mlp/keras_mlp.py

References

[1] Reymond, D., Okal, E. A., Hébert, H., & Bourdet, M. (2012). Rapid forecast of tsunami wave heights from a database of pre-computed simulations, and application during the 2011 Tohoku tsunami in French Polynesia. Geophysical Research Letters, 39 (11), 1-6. doi: 10.1029/2012GL051640

[2] Lalli, F., Postacchini, M., & Brocchini, M. (2019). Long waves approaching the coast: Green's law generalization. Journal of Ocean Engineering and Marine Energy. doi: 10.1007/s40722-019-00152-9

About

Automative approaches for capturing the localised tsunami response

Resources

License

Stars

Watchers

Forks

Packages

No packages published