Methodology
Overview
The Green City Index measures urban vegetation coverage using satellite imagery from the European Space Agency's Sentinel-2 mission. We calculate the Normalized Difference Vegetation Index (NDVI) for each city and determine what percentage of the urban area has healthy vegetation.
Data Source
Sentinel-2 Satellite Imagery
We use Sentinel-2 Level-2A (L2A) products, which are atmospherically corrected surface reflectance images. Key specifications:
- Spatial resolution: 10 meters (for red and NIR bands)
- Revisit time: 5 days (with both Sentinel-2A and 2B)
- Coverage: All land surfaces between 56°S and 84°N latitude
Data is accessed through the Copernicus Data Space Ecosystem.
Processing Pipeline
1. Granule Selection
For each city, we identify which Sentinel-2 tiles (MGRS grid) cover its boundaries. We then search for all available granules during the summer season (June-August) with less than 10% cloud coverage.
2. NDVI Calculation
The Normalized Difference Vegetation Index is calculated for each pixel:
Where NIR is the near-infrared band (B8) and Red is the visible red band (B4). NDVI values range from -1 to +1, where higher values indicate healthier vegetation.
3. Cloud Masking
We use the Scene Classification Layer (SCL) from Sentinel-2 L2A products to mask out:
- Clouds (high, medium, and low probability)
- Cloud shadows
- Water bodies
- Snow/ice
4. Temporal Compositing
To reduce noise and fill cloud gaps, we create a median composite from all valid observations during the season. This provides a representative view of vegetation during peak growing season.
5. City Clipping
The composite is clipped to official city boundaries from GADM (Global Administrative Areas Database) version 4.1.
6. Vegetation Classification
We classify pixels as "vegetated" if their NDVI value exceeds 0.40. This threshold captures healthy, actively photosynthesizing vegetation while excluding:
- Bare soil (typically NDVI 0.1-0.2)
- Built surfaces (typically NDVI < 0.1)
- Sparse or stressed vegetation (NDVI 0.2-0.4)
Metrics Calculated
| Metric | Description |
|---|---|
| Vegetation index | Percentage of city area classified as vegetated (NDVI > 0.40). This tells you how much of the city is green. |
| Average greenness | Mean NDVI value across all pixels in the city (0-1 scale). This tells you how green the city is overall, including non-vegetated areas. Higher values indicate denser or healthier vegetation. |
| Green Area (km²) | Total vegetated area in square kilometers |
| Total Area (km²) | City's total administrative area |
Example: Two cities might both have 10% vegetation index (same amount of green space), but different average greenness if one has denser parks while the other has sparse lawns.
Limitations
- Seasonal variation: Results reflect summer conditions only. Deciduous trees and seasonal vegetation may show different patterns in other seasons.
- Urban tree canopy: Trees in narrow streets may be partially obscured by buildings due to the satellite's viewing angle.
- Mixed pixels: At 10m resolution, pixels often contain a mix of vegetation and built surfaces, leading to intermediate NDVI values.
- Threshold sensitivity: The 0.40 NDVI threshold is somewhat arbitrary; different thresholds would yield different percentages.
Comparison with Morgenpost (2016)
| Aspect | Morgenpost | Green City Index |
|---|---|---|
| Satellite | Landsat 5/7/8 | Sentinel-2 |
| Resolution | 30 meters | 10 meters |
| NDVI Threshold | 0.45 | 0.40 |
| Time Period | 2005-2015 composite | Annual (2024+) |
| Coverage | 79 German cities | European cities (expanding) |
Data Attribution & License
Output Data License
All data outputs from this project (statistics, GeoJSON, maps, reports) are licensed under CC BY 4.0. You are free to share and adapt the data for any purpose, including commercial use, as long as you provide attribution:
Source Data
Satellite imagery: Contains modified Copernicus Sentinel data 2024. The Copernicus Sentinel-2 mission is operated by the European Space Agency (ESA) on behalf of the European Commission.
Administrative boundaries: GADM version 4.1 (gadm.org), licensed under CC BY 4.0.
City data: GeoNames (geonames.org), licensed under CC BY 4.0.