PyPlume

A collection of notebooks and methods made unifying the process of loading two-dimensional oceanic current vector fields from models and observations, simulating trajectory models, and analyzing and visualizing particle trajectories.

The core library used in this project is OceanParcels, which is used for particle advection math.

Link to corresponding paper

Environment setup

Clone the project repository first

git clone https://github.com/jerukan/PyPlume.git
cd PyPlume

It is highly recommended to use Miniconda to set up the Python environment.

The easiest way to install the dependencies is with the environment.yml file.

conda env create -f environment.yml

This will create a Conda environment named py3-parcels.

To make the Conda environment useable in Jupyter, run these two commands

conda activate py3-parcels
python -m ipykernel install --user --name py3-parcels --display-name "py3-parcels"

Data processing

More information about how to format the data can be found in data/README.md.

Saving netcdf files from Thredds

Use the download_data.ipynb notebook to save a specified region from some online source.

Gapfilling

Use gapfilling.ipynb to gapfill a netcdf file. Instructions are in the notebook.

Running simulations

First, set up config files in configs/.

(More information on creating config files can be found in the config README)

Then run the configured simulation with the following command:

python -m pyplume simulate -c configs/name_of_config.yaml

Which outputs a new netcdf file with particle simulation data and creates a sequence of snapshots of particle movement stitch together into a gif.

Alternatively, you can choose config files in the simulation_runner.ipynb notebook for the OceanParcels simulation.

FAQ

FieldOutOfBoundError or other Field errors during simulation

If particles are intented to go out of bounds during the simulation, add pyplume.kernels.DeleteStatusOutOfBounds as a kernel to the list of kernels in the configuration. Other error codes like ErrorTimeExtrapolation can be ignored using pyplume.kernels.DeleteStatusError.

How to run simulations backward in time?

The only thing that needs to change is a negative simulation delta time, which can be set in the simulation_dt field when configuring.

Where are logs stored?

They should be stored in logs/plumelogs.log.