CICE-Analysis

A Python toolkit for testing new implementations and analysing outputs from the CICE sea ice model.

Project Structure

Directory	Description
ciceplotting/	Main package — reads model outputs and generates figures
CICE-UTILS/	Saved ice_in configuration files for running experiments
Figures/	Generated figures, organised by case
Fortran/	Lightweight Fortran scripts for prototyping
Numerics/	Sparse matrix solver implementations
Rheology/	Free drift solver implementations and VP stress plotting routines

Core Package: ciceplotting

The ciceplotting package is the heart of this project. It provides everything needed to read NetCDF output files and cice.runlog to produce figures. It supports both idealised and global experiments.

How to Use ciceplotting

1. Set Up the Environment

Install the conda environment using the provided cice-env.yml file:

conda env create -f cice-env.yml

2. Configure the Namelist

Edit the appropriate namelist file to match your experiment:

Experiment Type	Namelist File
Global	namelist
Idealised	namelist_ideal

Both files share the same structure. Two separate namelists are maintained for convenience, to avoid repeatedly switching the same file back and forth.

The namelists are organised as follows:

###[Experiments]

Parameter	Description
exp	Experiment name
case	Case name — determines the subdirectory in Figures/ where figures are saved
plot_fields	1 or 0 — plot ice area and thickness fields (deprecated)
read_all	1 or 0 — read cice.runlog
plot_single	Plot a transect (deprecated)

[Dynamics]

Parameter	Description
kstrength	Ice strength parameterisation used
gridtype	Grid type used
var	NetCDF file prefix to analyse (typically iceh or iceh_inst)
dx	Horizontal resolution

[Numerical]

Parameter	Description
solver	Solver used during the run (typically fgmres)
imex	1 or 0
global	1 or 0
precond	Preconditioner used during the run

###[Figures]

Parameter	Description
datadir	Path to the input data directory
figdir	Path to the output figures directory
labels	Experiment label as it appears in figures

[Time]

Parameter	Description
dt	Time step
datestr	Date to analyse
startdate	Start date for reading data
enddate	End date for reading data
freq	Output frequency

If you want to read multiple experiments at the same time, you just need to provide multiple experiment names, solvers, preconditionners and labels. As long as they are located in the same directory and have the same outputs, the code will read and plot them all on the same figures. You might need to tweak the legend placement manually.

3. Running the analysis

To run the package, you need to use the following command:

python cice_analysis.py -c 'ideal or gx'

ideal is for idealised experiments while gx is for global.

Scripts

In the ciceplotting directory, there is also an unused (directly by cice_analsysis) folder but which contains useful bash scripts when using cice.

Architecture

- ciceplotting/
        - cice_analysis.py
        - namelists
        - setup.py
        - ciceplotting/
                - data/
                        - readdata.py
			- readnamelist.py 
		- plotting/
			- plot_cice.py
			- science.mplstyle 
		- scritps/
			- rename_files.sh 
			- run_segments.sh 
		- utils/
			- manip_arrays.py 
			- TimeUtilities.py