Site change brief: Medebridge Road - Solar Farm
Summary
Satellite measurements show a strong-confidence change in surface cover over 23.9 ha (39% of the site footprint) between April 2023–September 2023 and April 2026–September 2026 (6 vs 6 usable captures). Likely vegetation removal with exposed ground (consistent with earthworks/site clearance, but see alternatives). Alternative explanations are listed below and have not been excluded by satellite data alone.
Requested scope
| Question | Has the surface cover of this site changed between the two comparison periods, and by how much? |
|---|---|
| Before period | 2023-04-01 to 2023-09-30 |
| After period | 2026-04-01 to 2026-09-30 |
| Scope / exclusions | 1 site >= ~5 ha; two comparison periods within Sentinel-2 archive (2017+); pre-existing imagery only. No guarantee of usable cloud-free imagery; cannot resolve objects < ~20 m; not a survey, planning, environmental or legal certification; no inference of revenue/throughput. |
Site identity and footprint provenance
| Site | Medebridge Road - Solar Farm — operator/applicant as recorded in source register: Medebridge Solar |
|---|---|
| Asset class | solar_pv, Eastern (GB) |
| Centroid (WGS84) | 51.53327, 0.32343 |
| Footprint area | 61.5 ha polygon source |
| Footprint source | Footprint from Overture Maps 6f4e6ed8-902d-3f85-9386-32e63df0eb64 (source: OpenStreetMap; licence ODbL via OSM). Matched to register point: {"distance_m": 2773, "name_similarity": 0.94, "area_ratio_vs_register": 0.68, "score": 0.63, "candidates": 11}. Mapped polygon, not an official boundary. |
| Official identifiers | repd_ref: 10369; planning_authority: Thurrock; planning_ref: 21/02159/FUL |
| Register source | Overture Maps feature 6f4e6ed8-902d-3f85-9386-32e63df0eb64 — Overture Maps Foundation (OpenStreetMap-derived) (accessed 2026-11-08) |
| Independent record | REPD status 'Operational'; under construction 2024-04-01; operational 2025-09-01; REPD record updated 2025-11-11 |
Imagery used
Sensor: Copernicus Sentinel-2 MultiSpectral Instrument, Level-2A surface reflectance (10 m: blue, green, red, NIR; 20 m: SWIR-1 and scene classification, expanded to 10 m by nearest neighbour). Pixels read as bounded windows from the AWS Open Data cloud-optimised archive (Earth Search STAC); each scene's ESA product identity cross-checked against the official Copernicus Data Space Ecosystem catalogue where available.
109 captures were assessed over the site footprint; 12 were used (best usable capture per month). Excluded: 65 × cloud/shadow over AOI.
| Capture (UTC) | Satellite | ESA product | Scene cloud | Footprint clear | CDSE check |
|---|---|---|---|---|---|
| 2023-04-13 11:06 UTC | sentinel-2b | S2B_MSIL2A_20230413T105619_N0509_R094_T31UCT_20230413T135300.SAFE | 26% | 100.0% | ✓ verified |
| 2023-05-26 11:16 UTC | sentinel-2b | S2B_MSIL2A_20230526T110629_N0509_R137_T31UCT_20230526T124002.SAFE | 3% | 100.0% | ✓ verified |
| 2023-06-25 11:16 UTC | sentinel-2b | S2B_MSIL2A_20230625T110619_N0509_R137_T31UCT_20230625T123508.SAFE | 0% | 100.0% | ✓ verified |
| 2023-07-07 11:06 UTC | sentinel-2a | S2A_MSIL2A_20230707T105621_N0509_R094_T31UCT_20230707T171859.SAFE | 0% | 100.0% | ✓ verified |
| 2023-08-21 11:06 UTC | sentinel-2b | S2B_MSIL2A_20230821T105629_N0509_R094_T31UCT_20230821T140250.SAFE | 36% | 100.0% | ✓ verified |
| 2023-09-15 11:06 UTC | sentinel-2a | S2A_MSIL2A_20230915T105701_N0509_R094_T31UCT_20230915T152500.SAFE | 6% | 100.0% | ✓ verified |
| 2026-04-07 11:06 UTC | sentinel-2b | S2B_MSIL2A_20260407T105619_N0512_R094_T31UCT_20260407T133134.SAFE | 0% | 100.0% | ✓ verified |
| 2026-05-25 11:16 UTC | sentinel-2c | S2C_MSIL2A_20260525T110621_N0512_R137_T31UCT_20260525T144513.SAFE | 2% | 100.0% | ✓ verified |
| 2026-06-24 11:16 UTC | sentinel-2c | S2C_MSIL2A_20260624T110621_N0512_R137_T31UCT_20260624T144908.SAFE | 0% | 100.0% | ✓ verified |
| 2026-07-29 11:16 UTC | sentinel-2b | S2B_MSIL2A_20260729T110619_N0512_R137_T31UCT_20260729T133901.SAFE | 0% | 100.0% | ✓ verified |
| 2026-08-13 11:16 UTC | sentinel-2c | S2C_MSIL2A_20260813T110621_N0512_R137_T31UCT_20260813T144414.SAFE | 0% | 100.0% | ✓ verified |
| 2026-09-22 11:16 UTC | sentinel-2c | S2C_MSIL2A_20260922T110751_N0513_R137_T31UCT_20260922T144312.SAFE | 35% | 100.0% | ✓ verified |
Visual evidence
Measurements observed
Directly measured from satellite data
| Quantity | Before | After | Change | Unit |
|---|---|---|---|---|
| Mean NDVI (vegetation index) | 0.424 | 0.299 | -0.125 | index (−1 to 1) |
| Mean bare-soil index | 0.091 | 0.238 | 0.147 | index (−1 to 1) |
| Mean NDBI (built-up index) | 0.038 | 0.298 | 0.260 | index (−1 to 1) |
| Mean NDWI (water index) | -0.505 | -0.299 | 0.206 | index (−1 to 1) |
| Area flagged as changed | — | — | 23.92 | hectares |
| Share of valid footprint flagged | — | — | 38.9% | % of 6151 valid 10 m pixels |
| Usable captures in composite | 6 | 6 | — | count |
| Before-period NDVI variability (median temporal s.d.) | 0.222 | index | ||
Interpretation inferred
Interpretation — not directly observed
Vegetation loss exposed ground (confidence 0.76 on a 0–1 scale). Likely vegetation removal with exposed ground (consistent with earthworks/site clearance, but see alternatives).
Alternative explanations not excluded by the satellite data
- Agricultural harvest, ploughing or fallow land (seasonal, not development)
- Seasonal phenology difference if before/after months differ
- Drought stress or grass die-back
- Residual undetected cloud shadow or haze
Method
- Every Sentinel-2 L2A capture in each comparison period over the footprint was assessed using ESA's scene classification layer; captures with <90% clear pixels inside the footprint were excluded.
- For the best usable capture in each month, reflectance bands were read for the footprint window only, and NDVI, NDWI, NDBI and a bare-soil index computed per 10 m pixel.
- A multi-temporal test removed pixels that were transiently much brighter than the same pixel's period median (probable cloud missed by the classification); captures with >3% such pixels were excluded.
- Per-pixel median composites were compared between the seasonally matched periods using the decision rule above.
Limitations
- Sentinel-2 pixels are 10 m (some bands 20 m). Objects or changes smaller than roughly 20-30 m across, vehicles, small structures and building interiors cannot be resolved.
- Images are occasional snapshots (nominal revisit 2-5 days, but cloud frequently prevents usable views in the UK). The record is not continuous and changes between usable captures are not observed.
- Cloud, haze and shadow masking uses ESA's Scene Classification plus a multi-temporal bright-outlier test; residual undetected cloud or haze can bias individual measurements.
- Spectral change indicates a change in surface cover. It does not by itself establish the cause, ownership, legality, planning status, safety, value, throughput or contractual compliance.
- Seasonal and agricultural cycles (harvest, ploughing, drought) can produce changes similar to development; comparisons use the same months in each year to reduce, not eliminate, this effect.
- Where the site boundary is a mapped or estimated footprint (see Site identity), measurements outside the true boundary may be included or parts of the site excluded.
- Historical imagery cannot guarantee present or future conditions. This report is observational analysis, not a survey, valuation, environmental, planning or legal certification.
Quality control
| real data only | True |
|---|---|
| licences cleared | True |
| change event human qa | approved |
| images present | True |
| scenes cdse verified | 12/12 |
| company identity complete | False |
Licences and attribution
- Contains modified Copernicus Sentinel data 2023–2026
- Contains public sector information licensed under the Open Government Licence v3.0 (Renewable Energy Planning Database, DESNZ).
- Site footprint © OpenStreetMap contributors (ODbL), via Overture Maps.