These data provide Normalised Difference Vegetation Index (NDVI), Enhanced Vegetation Index (EVI), and Fractional Vegetation Cover (FVC) statistics for each 2021 Lower Layer Super Output Area (LSOA) in England and Wales; 2021 Super Data Zone in Northern Ireland; and 2022 Data Zone in Scotland. These data have utility for a range of applications related to the measurement of exposure to green space, and are also a composite component of our Access to Healthy Assets and Hazards indicator.

Vegetation indices are derived from Sentinel-2 Level-2A Surface Reflectance imagery processed via Google Earth Engine. For each year (2021–2025), images are filtered to the growing season (April–October) and pre-filtered to exclude scenes with greater than 40% cloud cover. Remaining cloud and cloud shadow pixels are masked using the Scene Classification Layer (SCL), excluding classes 3 (cloud shadow), 8 (medium probability cloud), 9 (high probability cloud), and 10 (thin cirrus). A median composite is then generated from all valid pixels at 10m resolution.

Content

The data is supplied as CSV files, one for each year, attached to this record.

The following indices are computed for each pixel:

NDVI (Normalised Difference Vegetation Index): $$\text{NDVI} = \frac{\text{NIR} - \text{Red}}{\text{NIR} + \text{Red}}$$

EVI (Enhanced Vegetation Index): $$\text{EVI} = 2.5 \times \frac{\text{NIR} - \text{Red}}{\text{NIR} + 6 \times \text{Red} - 7.5 \times \text{Blue} + 1}$$

FVC (Fractional Vegetation Cover): $$\text{FVC} = \left( \frac{\text{NDVI} - \text{NDVI}{\text{soil}}}{\text{NDVI}{\text{veg}} - \text{NDVI}_{\text{soil}}} \right)^2$$

where NDVI soil = 0.2 and NDVI veg = 0.86, with output values clamped to [0, 1].

Recommendations for use:

• Use median statistics (e.g., NDVI_median) to mitigate the effects of outliers and noise
• Use FVC for determining proportional vegetation cover across zones
• Use NDVI or EVI values for characterising "greenness" intensity
• The ratio of valid_pixels_sum to total_pixels_sum provides an indication of data quality and cloud contamination for each zone

Full details of the data pipeline used to generate these statistics will be published shortly on a Github repository at https://github.com/GeographicDataService/.

Quality, Representation and Bias

These averaged measures are developed primarily as indicators of greenness for social science and public health applications. Because input data are filtered to the growing season and combined into a median composite (which may blend observations from different dates within that period), these data may not be suitable for phenology-based studies requiring precise temporal information.

The pixel count statistics are exported separately from the vegetation statistics in your code — you may want to join these before final distribution, or document that they're in separate files.

A previous "2025" version of the our Small Area UK Vegetation Indices, released in 2025 and based on summer 2023 data from EOX and ESA Sentinel-2, has now been archived and replaced by this version which includes annual files from 2021 onwards. For research reproducibility purposes you can access this previous version, along with the previous record description, at the link listed below.