Photo by Group 5 during the site visit on 28 August 2026.
TOTAL BUILDING PERFORMANCE AND DIAGNOSTICS
The campus was planned as a rainforest. The architects organised it into four layers, and each one was given the job its equivalent does in a forest.
The forest floor is the park like ground plane, described in the design material as fenceless, seamless and car less. The understory is the podium, where the green fingers between blocks are there to bring in daylight and ventilation. The canopy holds the ward towers, and the design material calls it the most ventilated and day lit layer of the campus. The emergent layer is rooftop farming and gardens.
That is a useful place to start, because it means the wards were never intended to be sealed. The ventilation strategy for the top of this building is a stated design ambition, written into the concept before any equipment was chosen. What follows is what happens when a building holds that ambition and a fully conditioned hospital at the same time.
What the rainforest does for the air
It decides where people breathe outdoor air and where they breathe filtered air. Indoor air can be managed in three ways, in order of preference: control the source, dilute with outdoor air, then filter what remains (Hartkopf et al., 1997). The rainforest uses the first two at the scale of the whole campus. Only the sealed side gets the third.
Forest floor
Where: Car-less Level 1 ground plane. Drop-off, taxi and ambulance lanes at B1; car parks at B3 to B4.
Source controlVehicle exhaust is released away from where people walk, wait and enter.
On 28 August: Regional haze dominated. The lawn read 112.8 µg/m³ against 99.6 at the road frontage, so the ground plane cannot lower outdoor particulate. Traffic gases were not measured.
Understory
Where: The podium and the green fingers between blocks.
DilutionThe gaps are wind and daylight corridors that bring outdoor air into semi-outdoor space.
On 28 August: The B1 semi-outdoor corridor held 0.82 of outdoor PM2.5. Open space tracks the outdoor air, good or bad.
Canopy
Where: Ward towers. Class B and C wards, about 23,982 m², naturally ventilated.
Dilution, no filterCross ventilation from the prevailing monsoon, with openings on two sides.
On 28 August: Air speed on the open lawn averaged 0.15 m/s. The ward side is estimated at 66 to 83 µg/m³, not measured.
Emergent
Where: Rooftop farming and gardens.
No direct roleThe project material describes no link to the air of occupied rooms.
On 28 August: Not measured.
The rainforest sets which route each space takes. The two routes below are what that choice looks like on a haze day.
Layers from the campus design summary, Woodlands Health Campus and SAA Architects. Readings: Team's own, 28 August 2026.
Two ventilation strategies in one building
Woodlands Hospital does not have a ventilation system. It has two, running side by side under one roof and which one you breathe depends on which ward class you are admitted to.
About 23,982 m2 of the assessed floor area, 13 percent of it, is not air conditioned. These are the class B and C wards. Natural ventilation here means no air conditioning at all. It is the windows, and the north to south orientation was chosen, in the project team's own words, because of the circulation. The submitted design documents put it more formally: wards are designed with effective openings and layout to induce prevailing winds. This side works by dilution. Outdoor air comes in continuously and takes contaminants out with it.
Ward block facade with casement windows open, beneath projecting louvred sunshades. Photograph: Team's own, 28 August 2026.
Photo by Group 5 during the site visit on 28 August 2026.
Neither approach is better than the other. They fail differently and they fail at different times. On an ordinary day both work. They diverge when the air outside gets bad.
One outdoor air stream, two routes
Outdoor air, 28 August 2026
100.8 µg/m³
regional haze alert active
If you are in a clinic, lobby or ward on the air conditioned side
Filter at the air handling unit
What reaches you
2.0 to 7.0 µg/m³
About 98 percent of the particulate did not get through
If you are in a class B or C ward
Open window, no gate
Estimate
66 to 83 µg/m³
0.65 to 0.82 of outdoor air
Estimated, not measured. Ward access was not granted, so nobody has published a figure for this side.
Same outdoor air, two doors, and only one of them has a filter behind it.
Diagram: Team's own. Particle counts are illustrative and the figures at each end are measured values.
Air quality connection to other performances
Only the links below are carried, because only these have a measurement or a document behind them. Hover, focus or tap a feature to read what it does and the evidence behind it.
Green arrows mark features that contribute to indoor air quality. Red arrows mark features that draw from it.
Three of these five share one shape. A decision was taken for another reason and produced an air quality outcome that was not the reason.
Quantitative evaluation
Seven points were measured through the Medical Centre on 28 August 2026, from the MRT connection at basement two up to the Level 2 clinic lobby, with two outdoor references. The survey fell inside an active regional haze episode. The day before, the ASEAN Specialised Meteorological Centre raised its highest transboundary haze alert for southern ASEAN. Outdoor readings of 85 to 119 ug/m3 are external conditions, not instrument error, and the air speed data agrees: outdoor air at the open lawn averaged 0.15 m/s, which is close to still.
That makes envelope performance the finding rather than the outdoor number. The building was filtering a haze loaded air stream on the day it was measured.
Apparent PM2.5 reduction
98%
At the Level 2 lobby against the outdoor reference
Highest CO₂ above outdoor
112 ppm
Against a criterion of 700
Canteen PM2.5
39.7 µg/m³
About twenty times the Level 2 lobby on the same air system
Indoor points exceeding a criterion
1 of 4
Humidity at the canteen
Measured points, Medical Centre, 28 August 2026
| Point | Level | Space | Ventilation | People | CO₂ ppm | Above outdoor | Temp °C | RH % | PM2.5 µg/m³ | Indoor to outdoor |
|---|---|---|---|---|---|---|---|---|---|---|
| L1 outdoor | Bus stop frontage | Outdoor | 5 to 6 | 442 | 6 | 31.7 | 64.9 | 99.6 | 0.99 | |
| L1 outdoor | Flower Garden open lawn | Outdoor | About 10 | 431 | Below reference | 30.3 | 73.5 | 112.8 | 1.12 | |
| B1 | Semi-outdoor corridor | Semi outdoor | 10 to 15 | 480 | 44* | 28.4 | 77.9 | 83.0 | 0.82 | |
| L1 | Medical Centre lobby | Air conditioned | 2 to 3 | 535 | 99 | 25.8 | 58.8 | 7.0 | 0.07 | |
| L2 | Level 2 clinic lobby | Air conditioned | 2 to 3 | 542 | 106 | 24.2 | 61.4 | 2.0 | 0.02 | |
| B1 | Dining and retail, the canteen | Air conditioned | 40 to 50 | 545 | 109 | 25.7 | 68.1 | 39.7 | 0.39 | |
| B2M | MRT connection corridor | Air conditioned | 10 to 15 | 548 | 112 | 26.0 | 58.2 | 18.7 | 0.19 |
Carbon dioxide, air temperature and relative humidity are five second logged series. Particulate is a spot reading matched to each point by time window. Outdoor references are 436.5 ppm and 100.8 µg/m³. The asterisk on P10 means it was logged at 07:56 in a separate early session, so its differential is approximate and its three parameters are not simultaneous with one another.
Data source: measured on site by Group 5 on 28 August 2026, 9am to 11am.
Where the points sit
The air quality points are the same points as the acoustic survey, measured the same morning. Select a point in the table above to see it on the plan.
Fine particulate (PM2.5 µg/m³) comparison
Data source: measured on site by Group 5 on 28 August 2026, 9am to 11am.
Screening against the governing criteria
Singapore has no general indoor air quality standard that covers this building. SS 554 and SS 553 both exclude hospitals from their own scope, in every edition including the current 2026 ones. The instrument written for this building type is the BCA Green Mark for Healthcare Facilities scheme, under which the campus holds Platinum, reference GM2661/4/2017, certificate issued 31 January 2020. That scheme reaches SS 554 by reference rather than the other way round, adopting its Table 1 for carbon dioxide and its filtration clauses. All four indoor points are public areas, so the public area criteria apply.
| Parameter | Criterion | Measured | Result |
|---|---|---|---|
| Operative temperature | 24.2 to 26.0 °C | Conforms at every indoor point | |
| Relative humidity | 58.2 to 68.1 percent | Exceeded at the canteen only | |
| CO₂ above outdoor | 99 to 112 ppm | Conforms at about one sixth of the criterion | |
| CO₂ control method | Held within 13 ppm across a fifteenfold occupancy range | Not demonstrated by the measurement | |
| Particulate matter | 2.0 to 39.7 µg/m³ | Pending the numeric limit was not read from the source document | |
| Filtration at air handling units | Not inspected | Out of scope and scored zero in the as built submission |
Select a criterion to see the basis for its value.
Data source: measured on site by Group 5 on 28 August 2026, 9am to 11am.
Peaks matter more than means
A carbon dioxide trace in a circulation space is event driven. Each peak is a group of people passing the probe, and peak carbon dioxide does not follow occupancy. The canteen held forty to fifty people and never passed 563 ppm. The Level 2 lobby held two or three and touched 713. The difference is volume and air change, not headcount, and for a hospital the peak is the more useful reading, because a short sharp excursion in a small enclosed lobby is the exposure pattern that matters and a mean averages it away.
| Point | Space | Baseline CO₂ ppm | Peak CO₂ ppm | Excursion CO₂ ppm |
|---|---|---|---|---|
| PO3 | Level 2 lobby | 505 | 713 | 208 |
| PO3 | Level 2 lobby, second event | 533 | 675 | 142 |
| PO8 | Flower Garden outdoors | 411 | 533 | 122 |
| PO1 | Medical Centre lobby | 519 | 606 | 87 |
| PO6 | MRT connection corridor | 528 | 596 | 68 |
| PO5 | Dining and retail, the canteen* | 493 | 563 | 70 |
No return to baseline was observed at PO5. The rise is slow and monotonic, which indicates steady occupancy rather than discrete events.
Data source: measured on site by Group 5 on 28 August 2026, 9am to 11am.
Method and limits
Seven points measured between 07:56 and 10:43 on 28 August 2026, with two outdoor references. Carbon dioxide, air temperature and relative humidity logged at five second intervals on testo 440 dP instruments. Particulate recorded as spot readings on a TSI DustTrak DRX and transcribed by hand. The inpatient wards were not entered, because access was not granted, so every statement about the naturally ventilated route describes design intent rather than delivered performance. Air speed was recorded at three points only, the instrument having stopped before four points were reached. Coarse particulate was recorded but is not analysed here. Sound levels are held in the acoustics section and are not repeated.
One thing follows from the certification dates. The certificate was applied for in April 2017, issued in January 2020 and expired in January 2023. The first patients arrived in December 2023. The award therefore certifies design and construction intent, assessed on a building that was never occupied while the certificate was live, and no in operation assessment of this building appears on the public record.
Air quality here is not resolved by a system. It is negotiated between two of them. The sealed route is measurably good, and it is good because of decisions taken for heat and for noise. The open route is a stated design ambition, written into the rainforest concept at the top of this page, and on the morning we measured it had no filtration stage on a day when outdoor air was roughly ten times its baseline.
Singapore's own standards now concede the tension. The current revision of SS 554 introduces a resilience mode, setting out readiness plans for periods of poor outdoor air quality, and the revision of SS 553 adds recommendations for air conditioning operation during haze. Neither standard covers hospitals. So the question this study leaves open is a practical one: what happens in the naturally ventilated wards when the pollutant index rises, and what replaces the ventilation the windows were providing.
One asymmetry belongs beside it. Class B and C are the subsidised classes, in a catchment the hospital itself describes as lower socio economic status.