JOSP Reservoir Land Cover Classification

GEE script + ReadTheDocs tutorial for classifying land cover around the JOSP reservoir using Sentinel-1/2 and a Random Forest classifier trained on 2025 LCRS reference points, with independent field-data QA.

Contents

• script/josp_landcover_classification.js — the full GEE script, organised into 10 numbered modules.
• docs/ — the tutorial, written for ReadTheDocs (Sphinx + MyST Markdown).

Running the script

1. Open the GEE Code Editor.
2. Create a new script and paste in the contents of script/josp_landcover_classification.js.
3. Confirm your asset paths and CONFIG.fieldDataClassProperty match your data (see docs/02_config_and_legend.md).
4. Run.

Building the docs locally

pip install -r docs/requirements.txt
sphinx-build -b html docs docs/_build/html
open docs/_build/html/index.html

Publishing to ReadTheDocs

1. Push this repository to GitHub.
2. On readthedocs.org, import the repo — it will pick up .readthedocs.yaml automatically.
3. Build. The tutorial will be live at https://<your-project-slug>.readthedocs.io.

Filling in the legend

About half the Somalia LCRS class codes carried over from the original script don't have confirmed names (marked TODO in LEGEND, Module 2 of the script). Fill these in from the official Somalia LCRS document — see docs/02_config_and_legend.md for exactly what to edit.