This repo is an example of how machine learning can be used to carry out spatio-temporal data analysis of geological data along subduction zones, including mineral deposits. Modified from the workflows of Butterworth, N., D. Steinberg, R. D. Müller, S. Williams, A. S. Merdith, and S. Hardy (2016), Tectonic environments of South American porphyry copper magmatism through time revealed by spatiotemporal data mining, Tectonics, 35, 2847–2862, doi:10.1002/2016TC004289

Getting started

Step 1: Setup Runtime Environment

Use Docker or install the dependencies in your computer.

Step 2: Start Jupter Notebook

Run python3 jupyter notebook in the root folder of this repoisitory(usually the folder "spatio-temporal-exploration"). If you are using Docker, run the Docker container in the root folder of this repoisitory (See Docker section -> step2).

Step 3: Open Notebooks

Go into the "python" folder, open the notebooks and follow the instructions inside. The notebooks have been named with step numbers.

Docker

The easist way to run the workflow.ipynb is use docker.

Step 0: install Docker

go to https://docs.docker.com/install/ and follow the instructions inside.

step 1: go into docker folder and run

docker build -t my_spatio_temporal_exploration .

Alternatively, you can pull the docker container image from dockerhub.com. Run docker pull gplates/spatio-temporal-exploration and docker tag gplates/spatio-temporal-exploration my_spatio_temporal_exploration.

step 2: go back to the root folder of this repository

docker run -p 8888:8888 -it --rm -v`pwd`:/workspace my_spatio_temporal_exploration /bin/bash -c "source activate pyGEOL && jupyter notebook --allow-root --ip=0.0.0.0 --no-browser"

(IMPORTANT!!! run this command in the root directory of this repository)

step 3: in web browser, go to http://127.0.0.1:8888 and open the notebooks in the "python" folder.

Dependencies:

If you would like to setup the runtime environment on your computer, you need to install the following dependencies in Python3.

(Again, please consider using Docker if you are not too good with computers.)

For example, you can create a conda enviroment with the command below

conda create -n pyGEOL python=3.7 scipy scikit-learn matplotlib pyshp numpy jupyter cartopy pandas notebook netCDF4 opencv

pygplates -- https://www.gplates.org/download.html

scikit-learn -- https://scikit-learn.org/stable/

scipy -- https://www.scipy.org/

matplotlib -- https://matplotlib.org/

pyshp -- https://pypi.org/project/pyshp/

numpy -- https://numpy.org/

jupyter notebooks -- https://jupyter.org/

cartopy -- https://scitools.org.uk/cartopy/docs/latest/

pandas -- https://pandas.pydata.org/

netCDF4 -- https://github.com/Unidata/netcdf4-python

opencv -- https://opencv.org/

EarthByte/PlateTectonicTools -- https://github.com/EarthByte/PlateTectonicTools.git. Edit "plate_tectonic_tools_path" parameter in python/parameters.py to specify the location of PlateTectonicTools code.

Rotation Model and Age Grids:

The old AREPS plate model has some serious bugs which have been fixed in version 1.15.

The v1.15 age grids can be found here https://www.earthbyte.org/webdav/ftp/Data_Collections/Muller_etal_2016_AREPS/Muller_etal_2016_AREPS_Agegrids/Muller_etal_2016_AREPS_Agegrids_v1.15/.

The v1.15 rotation model is here https://www.earthbyte.org/webdav/ftp/Data_Collections/Muller_etal_2016_AREPS/Muller_etal_2016_AREPS_Supplement/Muller_etal_2016_AREPS_Supplement_v1.15/

FAQ

Why am I getting an error "ImportError: No module named subduction_convergence"?

Answer: The code you are trying to run depends on EarthByte/PlateTectonicTools -- https://github.com/EarthByte/PlateTectonicTools.git. You need to download the PlateTectonicTools code and edit the "plate_tectonic_tools_path" parameter in python/parameters.py to tell the code where to find the PlateTectonicTools code.

Reference

Butterworth, N., D. Steinberg, R. D. Müller, S. Williams, A. S. Merdith, and S. Hardy (2016), Tectonic environments of South American porphyry copper magmatism through time revealed by spatiotemporal data mining, Tectonics, 35, 2847–2862, doi:10.1002/2016TC004289