a. For CLS: 4 sets of 146 particles advected for 4 months each, re-initiated every 3 months to cover one year (2009-09 to 2010-06)
Where to download the data:
Which trajectories were computed?
- 146 particles, evenly spaced, are initiated in the MEDWEST region (see Fig.2 below for illustration) every 3 months, so that 4 sets of particles are produced (labeled exp1, exp2, exp3, exp4), initiated on 1st of july 2009, 1st of october 2009, 1st of january 2010,1st of april 2010.
- The initial locations of the particles are generated from the NEMO-eNATL60 grid, taking 1 location every 45 from the native NEMO grid. (--> it corresponds to one particle initiated every 45/60 = 0.75º)
- The lagrangian trajectories are then computed for 4 months based on the horizontal currents (u,v) at 15 m depth, and computed and saved every 1h.
- Particles are deleted from the set when they reach the regional boundaries, so after 4 months, there is 105 particles left in exp 1, 103 in exp2, 108 in exp3, and exp4 respectively). See Fig.3 for an illustration. The trajectories that are dropped before the end of the 4 months are stored in additional files apart (see below).
The netcdf files contain:
- For each experiment:
- One file contains all the trajectories that remain within the domain for the 4 months.
- Four additional files contains the trajectories that were dropped after 1, 2, 3 and 4 months. Those files are labeled with “dropped1M” “dropped2M” “dropped3M” and “dropped4M” respectively.
- The files contain variables: time, lat, lon of each trajectory, and also interpolated values of u_convert, v_convert at each particle location (other NEMO variables could be interpolated too, upon request).
- The interpolation in space is done by Ocean Parcels, which carefully takes into account the specificities of the NEMO grid (C-grid) (see OP’s doc: https://nbviewer.jupyter.org/github/OceanParcels/parcels/blob/master/parcels/examples/tutorial_nemo_curvilinear.ipynb).
- Each variable has 2 dimensions: ‘traj’ (id of each trajectory) and ‘obs’ (each obs along the trajectory, ie here every 1h for 4 month) .
- The interpolated u and v at 15m are given with the unit: m/s in the variables u_conv and v_conv in the file,and in unit: deg/s in the variables u_interp, v_interp.
Wind forcing files:
-
The surface wind data used to force the eNATL60 simulation is also provided, i.e. u10 and v10 (global) for 2009 and 2010:
-
Demo jupyter notebooks:
- How to create a particle set and advect it with Ocean Parcels notebook here,
- How to convert the sampled currents from degree/second in meters/second notebook here,
- How to plot the trajectories notebook here
Fig.1 Initial location and trajectories of 105 lagrangian particles over 4 months (2009-07 to 2009-10) computed with Ocean Parcels from the eNATL60 simulation hourly currents in the MEDWEST predefined region. The averaged SSH field is shown as background. (a) shows the entire MEDWEST domain, (b) shows a zoom on a smaller region.