Examples Gallery

This section provides comprehensive examples of using marEx for various scenarios.

Jupyter Notebooks

The Jupyter notebooks in the examples/ directory demonstrate complete workflows:

Gridded Data Examples

01_preprocess_extremes.ipynb - Data preprocessing for regular grids
02_id_track_events.ipynb - Event identification and tracking
03_visualise_events.ipynb - Visualisation and analysis

Regional Data Examples

01_preprocess_extremes.ipynb - Data preprocessing for regional analysis
02_id_track_events.ipynb - Regional event identification and tracking
03_visualise_events.ipynb - Regional data visualisation

Unstructured Data Examples

01_preprocess_extremes.ipynb - Preprocessing for irregular meshes
02_id_track_events.ipynb - Tracking on unstructured grids
03_visualise_events.ipynb - Unstructured data visualisation

Batch Job Examples for HPC

marEx provides example scripts for running detection and tracking workflows on HPC systems using SLURM job schedulers. These scripts are designed to be run from a login node and will automatically launch distributed Dask clusters to process large datasets efficiently.

Available Scripts:

run_detect.py - Launches SLURM jobs for preprocessing and detecting marine extremes using marEx.preprocess_data()
run_track.py - Launches SLURM jobs for identifying and tracking extreme events using marEx.tracker()

Key Features:

Automatic SLURM cluster setup via marEx.helper.start_distributed_cluster()
Configuration through environment variables (workers, runtime, SLURM account)
Complete workflow from data loading to saving results
Designed for large-scale oceanographic datasets on HPC systems

Configuration:

Both scripts accept configuration via environment variables:

DASK_N_WORKERS: Number of Dask workers to request
DASK_WORKERS_PER_NODE: Workers per compute node
DASK_RUNTIME: Job wall-clock time in minutes
SLURM_ACCOUNT: SLURM account/project for billing

Usage:

See the HPC Cluster Setup section in the User Guide for detailed configuration instructions and usage examples.

Customisation:

These scripts are intended as templates. Copy and modify them to fit your specific:

Data sources and file paths
Preprocessing methods and parameters
Tracking configuration
HPC cluster specifications