adepoplan_backend

This program creates a deposition simulation from a json configuration file. The output consists of many files, including figures, tables and a formatted report which presents the results of the simulation.

Installation

First, install conda and create an environment using the in-repo definition file conda.yaml:

conda env create -f conda.yaml

Then, install the package using pip:

pip install git+https://github.com/Fiskaaling/adepoplan_backend.git

Usage

The program initiates a build chain, where each build step creates new files that are created on top of existing files in the build directory. Initially, the build directory should contain three files:

• adepoplan.json: Configuration parameters specified by user interface
• cages.geojson: A geojson file with one polygon for each fish farm cage
• feed.csv: A table of feeding data

The program is executed from python as

python -m adepoplan_backend adepoplan.json

where the location of adepoplan.json (which may be specified as a full path) also defines the build directory.

The program will generate a particle tracking simulation, aggregate the results into a concentration map, generate some figures and a simple report.

The structure of the input files are documented as comments in the example files, which are located in tests/end_to_end/ex1.

Future work TODO/wish list

Issues:

• indexing error in concentration, where geojson cannot have poly_id that is not referenced in the production_data file
• something strange happens when particles_per_day = 100
  • this might be a me (aka Birgitta) problem

TODO:

• using pultiple forcing files to cover larger area
• set up preselected boundary boxes for the different fjords/sounds

Nice to have:

• write relevant information into the attributes of the .nc files, such as:
  • forcing used
  • particle liifespan
  • taucrit
  • vertical mixing
  • numerics setup
• easy way to separate the steps
  • e.g.
    • one: release particles from all cages
    • one: ladim for all cages
    • multiple: crecon
      • test out different production data and thus weights
  • e.g.
    • one release
    • multiple ladim, testing different taucrit
    • one crecon pr. ladim output