# The same access points. The same 300 m threshold. Two results.

In the London validation in “Auditing urban nature access claims for sustainable city governance”, the share of the population within 300 m of the same official green-space access points is **87.61% by straight line** and **50.66% on the walking network**. The difference is **36.95 percentage points**.

**3,251,463 of 8,799,776 represented residents** enter the aggregate “outside the threshold” set when measurement switches from a straight line to the walking network. The population is represented through 4,994 population-weighted centroids for London's LSOAs. Those points are not individual addresses, and the values do not describe personal routes.

[Citește în română](https://mariuscomper.uk/aceeasi-londra-doua-distante/)

## What changes

The comparison holds the OS Open Greenspace pedestrian access points, represented population and threshold constant. Only distance changes:

- **Straight line:** geometric distance from the centroid to the nearest access point.
- **Walking network:** the walking route between the node associated with the centroid and the node associated with the access point, including street detours.

The study also publishes straight-line distance to the green-space polygon. A polygon edge is not necessarily an entrance.

## Published matrix

Share of the represented population within the threshold:

| Method | 300 m | 500 m | 1,000 m |
| --- | ---: | ---: | ---: |
| Site polygon, straight line | 92.60% | 99.58% | 100.00% |
| Access point, straight line | 87.61% | 99.29% | 100.00% |
| Access point, walking network | 50.66% | 82.86% | 99.48% |

Under the same network model, the result rises by **48.82 percentage points** between 300 and 1,000 m: `99.48 - 50.66 = 48.82`. This is threshold sensitivity in the London validation, not an uncertainty interval or a universal effect.

## How the 3,251,463 count is derived

The exact count comes from the two no-access populations published in the reclassification table:

```text
4,341,752  walking network, outside threshold
-1,090,289 straight line, outside threshold
=3,251,463 enters the aggregate “outside threshold” set
```

The share check is `3,251,463 ÷ 8,799,776 × 100 = 36.949383…%`, displayed as 36.95%.

The count must not be reconstructed from the rounded percentage. `8,799,776 × 36.95% = 3,251,517`, which is 54 higher because the displayed percentage has only two decimal places.

## Represented population

The analysis covers 4,994 LSOAs and 8,799,776 Census 2021 TS001 residents. Each area's population is assigned to its population-weighted centroid. If the point crosses the threshold, the area's entire population receives the same aggregate status. Variation within the area is not recovered.

**A small aggregation difference:** the analysis uses 8,799,776 residents, 48 fewer than the regional Census 2021 total published by [ONS](https://www.ons.gov.uk/datasets/TS001/editions/2021/versions/3). This difference is compatible with ONS's “cell key” statistical protection, which can perturb published cells slightly. The comparison on this page remains unchanged.

## Limits

- Ordnance Survey warns that an access point included in OS Open Greenspace does not guarantee public access.
- Gates, opening hours, safety, usability for different mobilities, quality, crowding and realised use do not enter the percentage.
- OS Open Greenspace and OpenStreetMap are updated. This page freezes the study's published results; it does not describe current conditions at every entrance.
- The comparisons are descriptive, not causal, not a factorial experiment and do not establish the correct threshold.
- The package reproduces results from derived CSVs but omits provider caches and the complete raw GIS workflow.
- The page publishes no neighbourhood map, LSOA label, invented route or policy ranking.

## Sources

- Kexin Zhou and Liuna Geng, [“Auditing urban nature access claims for sustainable city governance”](https://doi.org/10.1038/s42949-026-00470-6), *npj Urban Sustainability*, Article in Press, 27 August 2026, CC BY 4.0.
- [Supplementary information](https://media.springernature.com/original/springer-static/esm/art%3A10.1038%2Fs42949-026-00470-6/MediaObjects/42949_2026_470_MOESM1_ESM.pdf), Tables 8, 9, 14 and 16.
- [Software package v2.1](https://media.springernature.com/original/springer-static/esm/art%3A10.1038%2Fs42949-026-00470-6/MediaObjects/42949_2026_470_MOESM2_ESM.rar), including the derived London CSVs.
- Ordnance Survey, [OS Open Greenspace: data creation](https://docs.os.uk/os-downloads/products/land-and-terrain-portfolio/os-open-greenspace/os-open-greenspace-overview/data-creation).
- ONS, [LSOA Population Weighted Centroids](https://geoportal.statistics.gov.uk/datasets/c8b3fc3c98a842708e201c8af1db15e3), [Census 2021 TS001, version 3](https://www.ons.gov.uk/datasets/TS001/editions/2021/versions/3) and [the cell-key perturbation explanation](https://www.ons.gov.uk/peoplepopulationandcommunity/populationandmigration/populationestimates/methodologies/protectingpersonaldataincensus2021results/pdf).
- [Normalised evidence and archive hashes used by local checks](https://mariuscomper.uk/aceeasi-londra-doua-distante/data/evidence.json).

Sources accessed 28 August 2026. The article was still marked Article in Press.