Parcels (Probably A Really Computationally Efficient Lagrangian Simulator) is an experimental prototype code aimed at exploring novel approaches for Lagrangian tracking of virtual ocean particles in the petascale age.

In the last two decades, Lagrangian tracking of virtual particles in Ocean General Circulation Models has vastly increased our understanding of ocean dynamics and how currents move stuff around.

However, we are now facing a situation where our Lagrangian codes severely lag the next generation of these ocean circulation models. These ocean models are so big and massively parallel, and they produce so much data, that in a few years we may face a situation where many of the Lagrangian frameworks cannot be used on the latest data anymore.

In this project, we will scope out and develop a new generic, open-source community prototype code for Lagrangian tracking of water particles through any type of ocean circulation models.

The latest version of Parcels, including tests adn examples, can be obtained directly from github via:

git clone --recursive https://github.com/OceanPARCELS/parcels.git
cd parcels; pip install -r requirements.txt
export PYTHONPATH="$PYTHONPATH:$PWD"

For a lighter checkout that does not include example data please omit the --recursive flag from the above command. As an alternatively, Parcels can also be installed directly via pip:

pip install git+https://github.com/OceanPARCELS/parcels.git

The above assumes that all dependencies are met, which can be achieved with:

curl -O https://raw.githubusercontent.com/OceanPARCELS/parcels/master/requirements.txt
pip install -r requirements.txt

In both cases a functional netCDF install is required.

A basic example of particle advection around an idealised peninsula (based on North et al., 2009, section 2.2.2) using 4th order Runge-Kutta scheme is provided in the tests directory. The necessary grid files are generated (using NEMO conventions) with:

python tests/test_peninsula.py --grid <xdim> <ydim> -p <npart>

where xdim and ydim are the numbers of grid cells in each dimension and npart is the number of evenly initialised particles. The resulting particle trajectories can be visualised using Parcel's utility plotting script:

python scripts/plotParticles.py 2d -p MyParticle.nc

An alternative execution mode that utilises SciPy's interpolator functions for spatial interpolation can be utilised with:

python tests/test_peninsula.py scipy -p <npart> --degree <deg>

where deg is the degree of spatial interpoaltion to use.