Julia Tutorials

Here, we run MITgcm, analyze ECCO solutions, and compare with Argo – all comfortably from within julia.

The included notebooks can be run on OSS, p-cluster, your own laptop, or using the ECCO-Docker configuration.

Tutorial Lineup

Tutorial	Topics
Guidelines	Install, Pkg, Julia basics
Visualise the ECCO model grid	Grids, Tiles, Maps, Sections, Projections, Polygons
Explore standard ECCO diagnostics	Zonal Means, Global Means, Time Series, Maps
Visualize and Compute ECCO transports	Transports, streamfunction, Meridional Transports, Transect Throughflows
Track the global observing system	Ocean-OPS, Argo, drifters, OceanRobots.jl
Explore Argo data and compare to ECCO	obtaining data, formatting, sampling ECCO, cost funtions
Explore SST data and compare to ECCO	OISST data, SST in ECCO, Anomaly Maps, Time Series, Marine Heat Waves
Explore SLA data and and compare to ECCO	Altimetry, PODAAC, ECCO, GRACE, cost funtions
Explore NSLCT data and compare to ECCO	Sea Level Change Team, observations, projections
Use the MITgcm.jl interface	setup, build, run, parameters, inputs, outputs
Run ECCO from Julia	configuration, HPC, diagnostics, perturbations, adjoint runs
Tracking particles in the Ocean	Pathways, Water Masses, Drifters, Floats, Plastics
Automatic Differentiation and Optimization	Adjoint, Automatic Differentiation, Optimization, Line Search
Interface to Models in Various Languages	MITgcm, Other Models, Automated Workflows

p-cluster

On the p-cluster you can use singularity instead of ECCO-Docker directly, as follows.

ssh -L 8888:localhost:8888 -i path_to_id_rsa_key_PCLUSTER -X USERNAME@34.210.1.198

source ecco-2025/book/tutorials/Julia_setup/setup_modules.csh
module add singularity-3.8.3-gcc-11.1.0-wqau5pb

singularity pull ecco-docker.sif docker://gaelforget/ecco-docker
singularity run -p 8888:8888 ecco-docker.sif