LCZ_coding

Local Climate Zone classification and temperature extraction workflows for Johannesburg, South Africa.

The repository has two halves. Python scripts at the top level pull temperature data from Google Earth Engine. An R project in johannesburg_lcz_analysis/ classifies Local Climate Zones and produces the urban heat island figures.

Temperature extraction (Python)

Script	Purpose
extract_jhb_temperature.py	ERA5-Land hourly 2 m air temperature over the Johannesburg metropolitan area
extract_jhb_modis_lst.py	MODIS land surface temperature, as an alternative to ERA5
test_gee_connection.py	Check Earth Engine authentication before a long extraction
validate_extracted_data.py	Quality checks on an extracted CSV
enable_earth_engine.sh	Enable the Earth Engine API on a Google Cloud project via gcloud

pip install -r requirements.txt
earthengine authenticate
python test_gee_connection.py
python extract_jhb_temperature.py
python validate_extracted_data.py johannesburg_temperature_data.csv

Each extraction writes a timestamped CSV, a companion _metadata.json recording the dataset, date range and sampling points, and an extraction_log.txt. Output is written next to the scripts, so no paths need editing.

SETUP_GUIDE.md covers Earth Engine registration and troubleshooting. SETUP_INSTRUCTIONS.md is the shorter walkthrough. README_GEE_EXTRACTION.md documents the extraction parameters in full, and QUICK_REFERENCE.md lists the common commands.

LCZ analysis (R)

johannesburg_lcz_analysis/ holds the classification and mapping pipeline, built on the LCZ4r package. Scripts in scripts/ run in numbered order from environment setup onwards:

Script	Purpose
01_setup_environment.R	Install R dependencies, including LCZ4r from GitHub
02_acquire_lcz_map.R	Retrieve the LCZ map for Johannesburg
03_analyze_urban_parameters.R	Urban canopy parameters and LCZ area statistics
04_climate_analysis.R	UHI intensity and thermal anomalies
05_spatial_interpolation.R	Continuous temperature surfaces
06_presentation_visuals.R	Figures for presentation

presentation/demo_workflow.R runs the sequence end to end. Outputs land in outputs/, with PNG and SVG versions of each map plus summary tables of LCZ area statistics and UHI intensity. See johannesburg_lcz_analysis/QUICKSTART.md to run it, MAPS_GUIDE.md for the map catalogue, and DATA_SOURCES.md for observational data provenance. NEXT_STEPS.md at the top level links the two halves together.

Data sources

Source	Used for	Notes
ERA5-Land Hourly (ECMWF/ERA5_LAND/HOURLY)	Hourly 2 m air temperature	~11 km grid, sampled at 10 points across the urban–rural gradient
MODIS MOD11A1 / MYD11A1 (collection 061)	Land surface temperature	1 km grid, Terra and Aqua, day and night
LCZ4r / LCZ Generator	Local Climate Zone map	Either the global dataset or a custom LCZ Generator submission
Weather station observations	Validation and interpolation	Not redistributed here; see johannesburg_lcz_analysis/DATA_SOURCES.md for how to obtain SAWS or NOAA ISD data

Extractions cover the bounding box 27.75–28.35 °E, 26.45–25.95 °S. The committed CSV (jhb_temperature_era5_20260211_121556.csv, and its copy johannesburg_temperature_data.csv) is a 30-day ERA5-Land extraction, 10 points, 7 200 hourly records; the exact parameters are in the accompanying metadata JSON.

Data policy

johannesburg_lcz_analysis/data/ and the top-level CSVs hold extracted climate data only. No health or participant data belongs in this repository.

Context

Part of heat vulnerability work at the HE²AT Center, Wits Planetary Health Research, supported by NIH award U54 TW 012083.

Citing

See CITATION.cff, or use the "Cite this repository" link on GitHub.

Licence

MIT, see LICENSE.