Heat is present, but still limited.
Only a small number of dangerous heat hours occur in most neighbourhoods. The baseline establishes which places are already warm before future warming is added.
The citywide average hides the heat justice problem.
In this synthetic UDA-city case study, the future scenario turns a seasonal heat problem into a much larger neighbourhood-by-neighbourhood risk. The analysis counts hours where 2 m air temperature exceeds 35 C during the hottest months available in the simulation period.
The central finding is simple: the future gets much hotter, but not evenly. Neighbourhoods with high heat exposure also tend to have weaker cooling access, more outdoor work, and higher deprivation. That makes adaptation a question of allocation, not only ambition.
Three scenarios, one fairness test.
The future heat burden rises sharply.
Dangerous heat hours increase across the city, with the largest increases in Jade Gardens, Kampong Lama, Taman Melati, Dhobi Lines, and Fuzhou Lanes.
A constrained package is allocated where it matters most.
The scenario uses limited land conversion and cooling-access gains to reduce risk everywhere, while giving larger benefits to the most vulnerable hotspots.
Where risk concentrates.
The highest future heat burden is not randomly distributed. Kampong Lama, Dhobi Lines, Fuzhou Lanes, and Mlima Moto are identified as hotspot neighbourhoods, while Jade Gardens and Taman Melati also show very high future dangerous-heat hours.
| Neighbourhood | Type | Future T2 > 35 C hours | Main vulnerability signal | Adaptation priority |
|---|---|---|---|---|
| Jade Gardens | refuge | 169 | High future heat exposure | Physical cooling, maintain refuge role |
| Kampong Lama | hotspot | 163 | Cooling access, deprivation, outdoor work | Largest combined support |
| Taman Melati | refuge | 153 | High future heat exposure | Shade and cooling access |
| Dhobi Lines | hotspot | 141 | Cooling access, deprivation, outdoor work | Cooling access and labour protection |
| Fuzhou Lanes | hotspot | 136 | Cooling access, deprivation, outdoor work | Cooling hubs and targeted retrofits |
| Mlima Moto | hotspot | 91 | Highest deprivation and outdoor work | Strong social adaptation package |
A feasible package, not a perfect-city fantasy.
The adaptation scenario deliberately uses constraints. Across the whole city, only 20% of paved or bare-soil surfaces can be converted to evaporative vegetation, grass, or water. Cooling access can increase only by 50% in total, allocated where vulnerability is highest. Vehicle heat is treated as a policy lever, but in this SUEWS setup anthropogenic heat is set to zero, so that lever is reported as a limitation rather than counted as a physical model effect.
20%maximum citywide conversion of paved or bare-soil land
+50%maximum citywide increase in cooling access
15%vehicle-heat reduction considered, but not active in this run
What the targeted package changes.
The allocation is intentionally unequal because the baseline risk is unequal. Kampong Lama, Dhobi Lines, Fuzhou Lanes, and Mlima Moto receive the strongest social-cooling gains, while every neighbourhood still receives enough support to meet the minimum 5% risk-reduction target.
This makes the scenario more just than a flat citywide distribution: it protects everyone, but shifts the marginal adaptation effort toward neighbourhoods where exposure and vulnerability overlap.
What this does not prove.
- UDA-city is a synthetic hackathon city, so the result is a method demonstration rather than a direct policy prescription for a real municipality.
- The socio-economic vulnerability indicators are synthetic and should be replaced with local census, health, labour, and cooling-access data before real-world decision-making.
- The future scenario represents a hotter climate stress test, not a fully downscaled climate projection with uncertainty ranges.
- The land-cover simulation changes surface fractions, but the social-cooling allocation is a vulnerability-pressure calculation, not a direct SUEWS indoor-health model.
- Because anthropogenic heat is zero in the current configuration, vehicle heat reduction is discussed as an adaptation idea but not credited in the physical heat-hour results.