Distance to Parks

Urbanizationv03.00 credits/per_km2

Responsible:Niccolò Calandri

Distance to Parks

Overview

Distance to parks measures the Euclidean distance (in metres) from every point in the study area to the nearest park, defined as a contiguous green area of at least 0.5 hectares (5,000 m²). This indicator provides information on accessibility to public parks and significant green spaces, following the ANGSt standard (Accessible Natural Greenspace Standard, Natural England, 2010) for the definition of "neighbourhood park".

Public parks play a fundamental role in urban quality of life: they provide spaces for physical activity, socialisation and contact with nature, contributing to stress reduction, mental health improvement and mitigation of the urban heat island effect. The 0.5-hectare threshold corresponds to the ANGSt definition of a neighbourhood park, a space large enough to offer diverse recreational and ecosystem functions.

The WHO recommends that every resident should have access to a park within 300 m of their home for significant health benefits (WHO, 2016). The ANGSt standard defines progressive thresholds: 300 m for a local park (≥ 2 ha), 500 m for a neighbourhood park.

Calculation Methodology

The indicator calculates the Euclidean distance from every point in the polygon to the nearest park:

Data source — ESA WorldCover 2021 (10 m), high-resolution global land cover map
Green classes — the following vegetation cover classes are considered:
- 10 — Tree cover (forests and tree coverage)
- 20 — Shrubland
- 30 — Grassland (meadows and pastures)
- 90 — Herbaceous wetland
- 95 — Mangroves
- 100 — Moss and lichen
Connected component analysis — scipy.ndimage.label with 8-connectivity to identify contiguous green patches
Area filtering — patches with area less than 0.5 hectares (5,000 m²) are excluded, keeping only significant parks
Distance Transform — scipy.ndimage.distance_transform_edt to compute Euclidean distance from each pixel to the nearest park
Conversion — values converted to metres based on raster resolution
Buffer — 1000 m buffer around the polygon to handle edge effects
Aggregation — calculation of mean, median and percentile distribution within the polygon

Unit: metres (m)

Data Sources

Code	Name	Provider	Resolution	Availability
`WRD_ESWCV_99`	ESA WorldCover	ESA / Vito	10 m	2020 — 2021
`WRD_IMPBS_99`	Impact Observatory	Impact Observatory	10 m	2017 — present
`WRD_S2SCL_99`	Sentinel-2 SCL	ESA/Copernicus	20 m	2017 — present
`WRD_GDYXX_99`	Google Dynamic World	Google	10 m	continuous
`WRD_OSMXX_99`	OpenStreetMap	OpenStreetMap Foundation	vector	continuous
`WRD_BIODL_99`	Biodiversa LC	Biodiversa+	10 m	2020 — present
`WRD_CLCBK_99`	CLC Backbone	EEA	100 m	2018
`WRD_CRINE_99`	Corine Land Cover	EEA	100 m	1990 — 2018
`WRD_DOMLT_99`	Dominant Leaf Type	Copernicus Land	100 m	2015 — present
`WRD_EUCRP_99`	European Crop Map	JRC	10 m	2018 — present

Quality Range and Levels

Indicator	Unit	Range	Inverted
`distance_to_parks`	m	[0, 150, 300, 500, 1000, 2000]	Yes

Inverted = Yes: a shorter distance is better, as it indicates greater proximity to parks and better access to significant green spaces.

Thresholds follow the ANGSt standard (Natural England, 2010):

Level	`distance_to_parks`	Interpretation
E	> 1000 m	Very poor park access
D	500 – 1000 m	Insufficient park access
C	300 – 500 m	Moderate park access
B	150 – 300 m	Good park access
A	0 – 150 m	Excellent park access — optimal ANGSt standard

Technical Notes

Distance to Parks maps must be generated before computation
The reader can also use the NAA layer from the database if available for the site
The 0.5-hectare threshold corresponds to the "neighbourhood park" definition per the ANGSt standard
The 1000 m buffer reduces artefacts at the polygon boundary
Green classes include wetlands and mangroves, which may not be physically accessible

References

Natural England (2010). "Nature Nearby: Accessible Natural Greenspace Guidance". Natural England Commissioned Report NECR017. (ANGSt standard)
WHO (2016). "Urban green spaces and health: A review of evidence". World Health Organization, Regional Office for Europe, Copenhagen.
Zanaga, D. et al. (2022). "ESA WorldCover 10 m 2021 v200". Zenodo. https://doi.org/10.5281/zenodo.7254221

Quality Ranges and Levels

distance_to_parks

(m)

ACritical

BLow

CModerate

DGood

EExcellent

0 → 149.99

150 → 299.99

300 → 499.99

500 → 999.99

1000 → 2000

Critical0 – 149.99

Low150 – 299.99

Moderate300 – 499.99

Good500 – 999.99

Excellent1000 – 2000

Data Sources & References

Sources

WRD_ESAXX_21WRD_IOAXX_99WRD_SCLXX_99WRD_GDYXX_99WRD_OSMXX_99WRD_BDVXX_99EUR_CBKXX_21EUR_CLCXX_18EUR_DLTXX_18EUR_CROXX_22

Methodology

The Distance to Parks map calculates the Euclidean distance from every point to the nearest park, where a "park" is defined as a contiguous green area of at least 0.5 hectares (5,000 m²). The calculation uses ESA WorldCover 2021 classification, identifying green classes: 10 (Tree cover), 20 (Shrubland), 30 (Grassland), 90 (Herbaceous wetland), 95 (Mangroves), 100 (Moss and lichen). Connected component analysis (scipy.ndimage.label) with 8-connectivity identifies contiguous green patches. Patches smaller than 0.5 hectares are filtered out. The 0.5-hectare threshold corresponds to the "neighbourhood park" definition per the ANGSt standard. Euclidean distance (scipy.ndimage.distance_transform_edt) is calculated on the resulting mask and converted to metres. A 1000 m buffer is applied to avoid edge effects. Classification into 5 levels: 1-Excellent (0-150m), 2-Good (150-300m), 3-Moderate (300-500m), 4-Bad (500-1000m), 5-Very Bad (>1000m).