Disk Partitions and Directory Structure

As discussed in Chapter 6.1.1, the SURF-VM environment utilizes two main disk partitions:

• the disk /dev/sda, "mounted" at the root directory /
• the disk /dev/sdb, "mounted" at the /scratch directory.

This section focuses on the /scratch partition, outlining its structure and contents.

SURF Directory Structure Overview

The /scratch directory hosts all SURF platform components and follows a structured directory layout, separating code, input data, and experiment outputs into well-defined folders. This organization is illustrated in Figure B.1 and summarized in the table below.

Directory	Description
from_GUI/	Stores user-defined configuration files for each experiment, created via the GUI.
surf_install/	Contains installation scripts and utilities for managing all SURF components.
surf_datasets/	Stores static datasets required for simulations, such as bathymetry, remapping weights.
surf_nemo/	Contains the source code of the SURF-NEMO package.
experiments/	Includes folders for each downscaling experiment executed by the user.

Below is a detailed description of the contents of these directories.

Fig. B.1 SURF package directories tree.

1. from_GUI/ Directory

Path: /scratch/surf/from_GUI/

This directory contains user-defined configuration files named setParFree.json. Each file specifies the input parameters for a specific downscaling experiment and is stored in a subdirectory named after the experiment ID (e.g., /scratch/surf/from_GUI/expID/setParFree.json).

At present, these files may need to be created or managed manually or by scripts. In future releases, they are intended to be generated directly via the SURF graphical user interface (GUI).

2. surf_install/ Directory

Path: /scratch/surf/surf_install/

This directory houses the surf_install package which provides essential tools for installing and managing SURF platform components. The main subdirectory surf_install_1.00/ includes:

• scripts/: Bash scripts for installation (install.sh) and release creation (dorelease.sh).
• ChangeLog.txt: A log of changes and updates to the surf_install package.
• ReadMe.txt: Overview and usage notes for the package.
• version.sh: Script that displays the package version on the VM desktop.

3. surf_nemo/ Directory

Path: /scratch/surf/surf_nemo/

This directory houses the surf_nemo package (see Section 6.2). The surf_nemo_1.01/ folder includes:

• nemo/: The NEMO ocean model source code (v3.6).
• scripts/: Scripts for pre- and post-processing required to run simulations.
• utilities/: Helper scripts containing utility functions for specific pre-/post-processing tasks.
• setParFree.json: A template configuration file for case-study experiments.
• ChangeLog.txt: Summary of changes to the SURF-NEMO package.
• ReadMe.txt: Describes the package contents and usage.
• Licence.txt: Licensing information.
• version.sh: Displays the current version on the VM desktop.

4. surf_datasets/ Directory

Path: /scratch/surf/surf_datasets/

This directory houses the surf_datasets package which contains essential input datasets used during model execution (see Section 6.2). The surf_datasets_1.01/ folder contains:

• bathymetry/: GEBCO-2014 bathymetric data (30 arc-second resolution).
• coastline/: GSHHG coastline datasets from NOAA’s National Geophysical Data Center (NGDC).
• meshmask/: to holds meshmask files for the parent ocean model and atmospheric sources.
• experiments_regrid/: to holds weight files for remapping ocean and atmospheric input data, along with meshmask and bathymetry files remapped onto the child grid (used in operational runs).
• ChangeLog.txt: Log of changes for the dataset package.
• ReadMe.txt: Documentation and usage notes.
• version.sh: Displays dataset package version on the VM desktop.

5. experiments Directory

Path: /scratch/surf/experiments/

This directory stores all executed downscaling experiments. Each experiment is located in its own folder named after the experiment ID (e.g., expID) and includes:

• setParFree.json: A copy of the configuration file used for the experiment (copied from surf/from_GUI/expID/).
• code/: The source code version used (copied from surf_nemo/current/).
• data/: All datasets used during the experiment:
  • indata/: Source input files.
  • extrapoldata/: Extrapolated input data.
  • regriddata/: Regridded data.
  • outdata/: Simulation outputs.
  The input datasets are downloaded from local or web ftp repositories for the selected simulation period.
• figure/: Visual outputs from the simulation:
  • indata/: Plots of input data.
  • extrapoldata/: Plots of extrapolated data.
  • regriddata/: Plots of regridded data.
  • outdata/: Plots of output data.
  • Comparative plots between parent and child data.