Calvin's PhD thesis

Running OpenDrift scripts

Installing dependencies

We recommend using a virtual environment to install all dependencies for this project. miniconda is a good way to create a virtual environment with specific versions of Python, working on all major plateforms (Windows, MacOs and Linux).

After installing miniconda, you need to install OpenDrift (forked version):

git clone https://github.com/simonweppe/opendrift
cd opendrift
conda env create -f environment.yml
conda activate opendrift_simon
pip install -e .

Everytime you want to work on the project, you need to activate it before:

conda activate opendrift_simon

You should see (opendrift_simon) prepended to your command line prompt as a hint that it is now activated.

To deactivate the environment, type

conda deactivate

How to create a simulation

Create your simulation object,

o = BivalveLarvae(loglevel=0)

add readers,

data_path0 = '/nesi/nobackup/mocean02574/NZB_N50/nz5km_his_201512.nc'
reader_moana_dec15 = reader_ROMS_native_MOANA.Reader(data_path0)
reader_moana_dec15.multiprocessing_fail = True # bypass the use of multi core for coordinates conversion and seems to make the model run much faster.
o.add_reader([reader_moana_dec15])

and specify a land mask to use.

o.set_config('general:use_auto_landmask', False) # this will use the built in landmask in the Moana reader.

Seeding Particles

Define the corners of the polygon you want to seed within.

#Cape Reinga
lons = [170.75, 170.8, 170.8, 170.75]
lats = [-45.95, -45.95, -46, -46]

Define number of particles, depth, and time of release.

number = 100
z = -10
time = datetime(2015,12,1)

Seed particles

o.seed_within_polygon(lons, lats, number = number, time = time, z = z)

Set configs

There's a ton of different configurations you can add to modify particle+environmental properties

o.set_config('drift:max_age_seconds', 3600*24*30) 
o.set_config('drift:lift_to_seafloor',True)
o.set_config('drift:horizontal_diffusivity',.01) 
o.set_config('environment:fallback:ocean_vertical_diffusivity', .01) 
#etc.

Run model

Use o.run() to run the model.

o.run(stop_on_error=False, 
    end_time=reader_moana_dec15.end_time, 
    time_step=900, outfile= 'nc_out.nc', 
    time_step_output = 3600.0, 
    export_variables = [])