How ksanjivy/swac_global_potential_dataset works
This repo is not an online service; it’s a reproducible geospatial data-processing + publication dataset for assessing SWAC (Sea Water Air Conditioning) potential globally.
Its core purpose/pipeline:
1. Start from input climatology (NetCDF seawater temperature; plus UTCI heat-stress dataset).
2. Derive isotherm depth grids for intake temperature thresholds (MATLAB .mat matrices).
3. Combine those depths with cross‑shore bathymetric profiles to compute pipeline length from coastal points to reach target isotherm depths (TXT outputs).
What the file tree tells you about architecture
Top-level directories map cleanly to pipeline stages and published artifacts:
• climatology/ — inputs (NetCDF)
• seawater_intake_depth/ — derived intermediate products (MATLAB .mat isotherm depth matrices)
• bathymetric_profile/ — inputs for distance computation (TXT profiles split in parts)
• pipeline_length/ — derived outputs per temperature (TXT per threshold)
• metrics/ — aggregations (CSV summaries)
• maps/ — visualization outputs grouped by region (world/us/europe/asia)
• scripts/ — processing & plotting implementation (MATLAB + Python), relative-path execution
This structure implies a *batch/offline* workflow with repeatable regeneration of all outputs if the large data files are present (downloaded from Zenodo).
Zenodo archive download
→ extracted into local repo directories (climatology, bathymetric_profile, etc.)
climatology/seawater_temperature_climatology_2000_2020.nc
→ (MATLAB) compute_seawater_isotherm_depth.m
→ seawater_intake_depth/iso_<T>_matrix.mat for each T
3. seawater_intake_depth/iso_<T>_matrix.mat + bathymetric_profile/bathymetric_profiles_part*.txt
→ (MATLAB) compute_swac_pipeline_length.m (+ haversine.m)
→ pipeline_length/pipeline_length_<T>.txt
pipeline_length/pipeline_length_<T>.txt (for all T)
→ (MATLAB) compute_swac_metrics.m
→ metrics/swac_*_summary.csv
→ (Python) plot_world_map.py, plot_zoom_maps.py
→ maps/<region>/*_map_<T>.png
maps/<region>/*_map_<T>.png series
→ (Python) generate_all_gifs.py
→ maps/<region>/*_map_all.gif
*parameterization dimension*:
• A loop/iteration marker: “For each intake temperature threshold T (5.0–15.0 step 0.5)”.
Architectural patterns / design principles to call out
• Offline batch pipeline (file-based interfaces between stages)
• Deterministic, reproducible processing (relative paths; fixed naming conventions)
• Artifact versioning by filename convention (<T> encoded with one decimal)
• Separation of concerns:
• MATLAB for numeric/geo computations
• Python for visualization/automation of map products
• Dataset publication workflow (raw inputs → derived products → aggregated metrics → figures)