Dossier · Bruxelles · sector 21004E72-
Maison Rouge (Place)-Sud: canopy 26% below the city’s median, 7,180 residents
Why money for greening should come to this sector, and how the result will be checked with the same method that measured its state.
Passport
Who lives here and what is around
- residents
- 7,180
- children under 15
- 1,71625%
- aged 65 and over
- 5508%
- median taxable income, Statbel 2023
- €22,558
- tree canopy, class probability
- 0.043
- built-up and paved, class probability
- 0.667
By canopy the sector stands in the lowest 27% of the city, by greenness in the lowest 16%. Income is below the lower third’s cut (€24,154): two shortages meet here, little green and little money.
How the sector was picked. By a rule, not by hand: the lowest third by income, the lowest third by canopy, a summer surface at least 1 °C hotter than the city’s median, a commune that does not pass the Article 8(1) exemption, at least 500 residents; among those, the sector with the most children under 15. 110 sectors fit; this dossier is about the first.
State
Where the sector stands on the city’s scale
Tree canopycity median 0.059
0.043
26% below the city’s median
Greennesscity median 0.415
0.264
36% below the city’s median
Built-upcity median 0.662
0.667
at the city’s median
The tick is the city’s median, the bar the sector. Class values are probabilities, not shares of ground: compare them between sectors, but do not read them as percentages of area.
Trend
It will not improve by itself
Against twelve matched control sectors, from summers 2017–19 → 2021–24, greenness changed by +0.0132 against a significance threshold of 0.0345, and canopy by +0.0006 against 0.0144. The change did not cross the threshold: the sector is standing still.
The comparison is with a control, not with the sector’s own past. Without subtracting matched places, the city-wide shift of +0.032 makes 263 of 668 sectors (39%) look “significantly greener”, against 16 once they are subtracted.
After the baseline, in seasons 2025–2026, nothing in the sector crossed the threshold.
What this does not prove. No growth is not proof that the city did nothing: works the method cannot see may have gone on here. Restoring existing green does not show in these numbers by construction, and scattered street planting drowns in the sector’s area.
The law counts this sector. Bruxelles does not pass the Article 8(1) exemption: 35.1% green in its urban centres and clusters against a 45% threshold (the Commission’s method, CLCplus 2021). Whatever is planted or lost here enters the national count of no net loss.
Fairness
The poorest third of the city lives with half the canopy
| Group | Median income | Canopy 2024 | Greenness 2024 |
|---|---|---|---|
| Lowest third by income | €21,386 | 0.051 | 0.313 |
| Highest third | €31,926 | 0.090 | 0.494 |
| Maison Rouge (Place)-Sud | €22,558 | 0.043 | 0.264 |
Levels of summer 2024, weighted by population; income thirds by sector, Statbel 2023.
The lowest third’s canopy is 0.56 of the highest third’s. Regulation (EU) 2024/1991 requires no net loss of green and canopy by 2030 but says nothing about how they are shared between neighbourhoods; UN indicator 11.7 and the Tree Equity Score do. The sector lies below the average even for its own group.
Who lacks shade here. 1,716 children under 15 live in the sector (25% of its residents) and 550 people aged 65 and over (8%), 125 of them over 80: the groups heat endangers most. Across the city, 46% of all children live in the third of sectors with the least canopy, so the shortage of shade is not shared equally, and it falls hardest on them. This sector is in that third.
The rule and the heat meet here. By the 3-30-300 rule of urban forestry, neither measurable part of the rule is met: canopy covers 9.5% of the sector against 30%; 12% of its area lies within 300 m of a park, against half. The summer surface here is 37.1 °C, 1.5 °C above the city’s median. The less canopy, the hotter: across the city the link is almost deterministic (r = −0.89), and this sector sits at the bad end of both scales at once. Temperature here describes a state: the project does not measure cooling from an intervention, because the significance threshold for it is larger than a realistic park effect.
What to place here
Where City Planner puts the first objects
None of City Planner’s first twenty candidates of any kind falls in this sector.
What an intervention can deliver
Three measured cases instead of promises
| Intervention | Effect | Threshold | Times | Reading |
|---|---|---|---|---|
| Porte de NinoveBrussels: wasteland by the canal → park, 2.2 haAbove noise in 7 of 7 summers after it opened. | +0.246 | 0.036 | 6.8× | Above noise in 7 of 7 summers after it opened. |
| DroogdokkenparkAntwerp: old docks → a quayside parkAbove noise in 7 of 8 summers after it opened. | +0.147 | 0.046 | 3.2× | Above noise in 7 of 8 summers after it opened. |
| Park Spoor NoordAntwerp: rail yard → park, 19.5 haAbove noise in 14 of 15 summers after it opened. | +0.148 | 0.058 | 2.6× | Above noise in 14 of 15 summers after it opened. |
This is not a forecast for this sector but the range measured on comparable works: a change of cover over most of a site gives an effect three or more times the threshold, and it holds for years. The size depends on what was there before: wasteland gives more than a quayside that was already green.
What this table leaves out on purpose. Cooling: the significance threshold for surface temperature came out at 1.32 °C, more than the realistic effect of an urban park, so there are no claims about degrees here. The cost of an effect: prices compare only within one type of work, and the budgets of comparable projects are published only in part.
Commitment
What exactly we will check, and when we will refuse
- The footprint of the works is fixed before they start. The polygon, its hash and the date enter the register before the first day of works. Drawing it afterwards around whatever turned green becomes impossible.
- The control is named at the same time. 12 sectors that resemble the site in its state before the works, none closer than 1.5 km. The list is published before the measurement; if no comparable control exists, that is recorded as a refusal, not adjusted.
- The threshold comes from the city’s calibration, not chosen for the result: the null distribution is computed with the same matching and the same images as the audit itself, the way the trend’s threshold above was.
- The comparison window is built from the completion date: a baseline before the works begin, the assessment two years after they end. If the works fall inside a window, the audit declines to compute.
- The verdict takes two instruments. Greenness and land-cover classes from the same images: the first says that something changed, the second what exactly, down to “unsealed but not planted”.
- A refusal is a published outcome. Fewer than 8 images in a window, a gap to the control above 0.04, works outside the window: no verdict is given, and that is printed alongside the results.
Limits
What the method will not measure, even when everything is done right
- Cooling of air or surface: the channel’s noise is larger than a realistic effect.
- Restoration of existing green: repairing paths and ponds does not change vegetation.
- Scattered street planting measured at the level of a sector: half a hectare of new canopy on nineteen hectares is lost.
- The cause of a change: the satellite names the place, documents name the cause. Without a document a change is published as a measurement, not as anyone’s credit.
- Young plantings in their first years: saplings on a meadow read less green than the grass they replaced, and honest greening can measure as a loss.