EMODnet-Biology-Zooplankton-OOPS

Workflow for the calculation of OOPS zooplankton gridded abundance maps

This is the Github repository of the EMODnet Biology product described here:

http://www.emodnet-biology.eu/blog/oops

This directory contains all the steps to create the latest gridded map products of zooplankton.

Data

The data that was used for these products are the SAHFOS/Marine Biological Association (UK) data (https://www.cprsurvey.org/) from the North Atlantic, from 1958 to 2016.

All data is available here: CPR Survey. (2018). Operational Oceanographic Products and Services (OOPS) CPR Survey Data Extract (2016) [Data set]. Marine Biological Association. https://doi.org/10.17031/W508-X849

preprocessing

The raw csv file is first preprocessed in https://github.com/EMODnet/EMODnet-Biology-Zooplankton-OOPS/blob/master/00_Preprocess_OOPS-data.R.

This script:

1. load original datafile
2. calculate for each species/variable the log(x +1)
3. create for each species a txt file: bigfile_species.txt that can be read by DIVAnd

Calculation of optimal interpolation parameters:

In the next file, the optimal interpolation parameters are calculated https://github.com/EMODnet/EMODnet-Biology-Zooplankton-OOPS/blob/master/01_OOPS_1yr_param_calculation_2000_2016.ipynb

The script:

• The script loops over the variable names in varname
  • The script calculates the optimal signal-to-noise (e) and correlation length (l) parameters for the (full) years 2000 to 2016 by cross validation.
  • These optimal parameters are stored in
    • /varname/varname_newe.txt
    • /varname/varname_newl.txt

Calculation of DIVA interpolated products:

https://github.com/EMODnet/EMODnet-Biology-Zooplankton-OOPS/blob/master/02_OOPS_diva3d_calculation_1param.ipynb

• The script loops over the variable names in varname
  • the signal-to-noise (e) and correlation length (l) parameters calculated in the previous script are loaded from
    • /varname/varname_newe.txt
    • /varname/varname_newl.txt
  • these parameters are first filtered (only l < 500 000 and e < 15 are retained) and then averaged
  • with these new averaged parameters the diva product is calculated: 1 diva interpolated map
    • per species
    • per season
    • per year
  • the output netcdf files are stored in /varname/netcdf_all/"

combining all the separate NetCDF files for one species per season

https://github.com/EMODnet/EMODnet-Biology-Zooplankton-OOPS/blob/master/03_Postprocessing_Combine_Netcdfs_per_season.ipynb

The script

• loops over the variable names in varname
• creates a new directory /varname/netcdf_season/
• and combines the data of the netcdfs in /varname/netcdf_all into 4 netcdfs (one per season): varname_StartyearEndyear_season.nc

with as season a number between 1 and 4:

1. months 1-3
2. months 4-6
3. months 7-9
4. months 10-12