Estimate the urban heat island intensity ΔT — the extra warmth of a city centre relative to its surrounding rural area — by combining Oke's (1973) population-log formula with correction terms for albedo, impervious surface, green-space ratio and wind speed. See how cool roofs, greening and ventilation modify ΔT in real time.
Parameters
City population P
×10k
Population (×10,000). Oke's formula uses the logarithm of P
Urban albedo α_u
Asphalt/concrete: 0.10–0.20. White roofs: 0.30–0.40
Rural albedo α_r
Grass / cropland / bare soil: 0.20–0.30 typical
Impervious surface
%
Fraction of asphalt, rooftops and other surfaces that block rainfall infiltration
Green-space ratio
%
Parks, street trees and green roofs — surfaces with evapotranspiration
Wind speed U
m/s
Mean nocturnal wind. UHI almost vanishes above 5 m/s
Results
—
Oke base UHI ΔT_base (K)
—
Albedo correction (K)
—
Impervious + green (K)
—
Wind correction (K)
—
Estimated UHI ΔT (K)
—
Severity
—
Urban vs rural cross-section — heat flow schematic
Left: rural area with trees and evapotranspiration. Right: urban area with low-albedo pavement and buildings. Arrows and pulses show solar absorption → storage → nocturnal release. The urban side is shaded red as it warms; the rural side stays blue-green.
UHI vs population P (log axis)
Breakdown by correction term (base / albedo / impervious / green / wind)
P = population (thousands), α = albedo, U = wind speed. Oke's (1973) base form plus correction terms for albedo difference, impervious surface, green space and wind.
Albedo correction: +0.5 K per 0.05 of (α_r − α_u). Impervious I above 50%: +0.05 K/%. Green G: −0.05 K/%. Wind U above 1 m/s: −0.4 K·s/m.
Urban Heat Island (UHI) Intensity
🙋
In summer the news keeps saying "Tokyo will be another tropical night". Is the city really that much hotter than the countryside, or is it just an impression?
🎓
It's not just an impression. Long-term records from the Japan Meteorological Agency show Tokyo's downtown is on average 2–3 K warmer at night than its rural surroundings, and on extreme days the gap exceeds 5 K. We call that the Urban Heat Island (UHI). The drivers are basically four: concrete and asphalt store solar heat by day and release it slowly at night, dense buildings block ventilation, air conditioners and cars dump anthropogenic heat, and the lack of greenery and water removes evaporative cooling. Try moving the "City population" slider on the left — see how ΔT jumps between a 1-million city and a 10-million city.
🙋
Going from 1 million to 10 million pushed the base UHI from 1.97 K to 3.98 K — almost double! What's the formula behind that?
🎓
That's the famous Oke empirical formula: ΔT_uhi,max = 2.01·log10(P) − 4.06, with P in thousands of people. The key piece is the log10 — a 10× population jump only adds about 2 K. Oke published this in 1973 after fitting data from North American cities, and 50 years later it's still the standard starting point for citywide UHI estimates. The lower chart uses a logarithmic axis, which is why the curve flattens for very large cities — UHI saturates rather than growing without bound.
🙋
OK, so population is fixed. What can we actually do to bring ΔT down? More greenery?
🎓
There are three big levers. First, cool roofs — repainting dark roofs white or using high-reflectance coatings raises urban albedo from 0.15 to 0.30. In this model that alone cuts ΔT by about 1.5 K. Second, greening and street trees — pushing the green-space ratio from 15% to 30% lowers ΔT by 0.75 K thanks to evapotranspiration. Third, permeable pavement — dropping the impervious fraction from 70% to 50% buys you another 1.0 K through infiltration and retained-water cooling. Combine all three and a 1-million city can shed more than 3 K, which translates into a major heatstroke-risk reduction.
🙋
What about wind? People always talk about "good ventilation".
🎓
Wind is decisive. Oke observed that maximum UHI only develops on clear-sky nights with wind under 3 m/s, and largely disappears above 5 m/s. That's why the wind correction in this tool kicks in above 1 m/s with a slope of −0.4 K·s/m. The implication for planning is that securing "ventilation corridors" — wind paths through the city — is just as impactful as greening. Tokyo's bayfront skyline has actually been re-oriented in places to let sea breezes penetrate inland on summer afternoons.
🙋
One last question — what does a 3 K UHI actually mean for human health?
🎓
More than you'd think. Each 1 K of air-temperature increase raises the heat-stress index WBGT by roughly 0.7 K. WBGT 28 °C triggers a "severe warning" alert and 31 °C is "danger". Emergency-room heatstroke admissions roughly 1.5–2× for every 1 K of WBGT increase. So shaving 3 K off UHI isn't just about comfort — it's a public-health intervention on the same scale as halving heatstroke ambulance calls. That's why this tool flags ΔT > 3 K as "significant" and > 5 K as "severe".
Frequently Asked Questions
The starting point is the Oke (1973) empirical formula ΔT_uhi,max = 2.01·log10(P) − 4.06, where P is the city population in thousands. It expresses the rule of thumb that a 10× increase in population raises the maximum UHI by about 2 K. This tool adds correction terms for albedo difference, impervious-surface fraction, green-space ratio and wind speed: ΔT = ΔT_base + f(α, vegetation, U). For a 1-million city (P=1000 k), ΔT_base = 2.01×3 − 4.06 = 1.97 K. Lower urban albedo and more impervious surface push ΔT up, while greenery and wind push it down.
In this tool, increasing urban albedo α_u by 0.05 lowers ΔT by about 0.5 K. Replacing a typical dark roof (α≈0.10) with a white cool roof (α≈0.30) increases α by 0.20, so ΔT drops by roughly 2.0 K. Field experiments in Los Angeles and New York have reported 1–3 K daytime air-temperature reductions from large-scale cool-roof programs, consistent with this estimate. Note that cool roofs also raise winter heating loads (heating penalty), so a latitude-dependent cost-benefit check is necessary.
The model lowers ΔT by 0.05 K per +1% green space. Greenery cools the air by (1) increased latent-heat flux from evapotranspiration, (2) shading by tree canopy and (3) a higher effective albedo. Raising the green-space ratio from 15% to 30% therefore reduces ΔT by about 0.75 K. A single mature tree can evapotranspire 100–400 L/day on a clear day, comparable to several residential air conditioners. However, dense canopies can suppress nocturnal radiative cooling, so species selection and street-tree placement matter.
The wind correction in this tool kicks in above 1 m/s with a slope of −0.4 K/(m/s). At U = 0 the correction is zero, so ΔT_base plus the albedo, impervious and green corrections appear without any wind damping. Oke's original observations confirm that the maximum UHI develops under calm clear-sky nights (U < 3 m/s) and typically vanishes above 5 m/s. Japanese summers under the Pacific subtropical high frequently feature weak winds, which is one driver of tropical nights. Securing ventilation corridors is a central tactic in urban heat-mitigation planning.
Real-world applications
Urban planning and master plans: Major cities such as Tokyo, Osaka, Singapore and New York maintain heat-island action plans. Macro models like this one let planners convert proposed greenery targets (Tokyo aims for 30% canopy cover, Singapore for 50%), rooftop-greening mandates and white-pavement programs into expected ΔT reductions, ranked by cost. Because each 1 K cooling cuts heat-stroke ambulance calls by roughly 30%, the trade-off between mitigation investment and healthcare expenditure can be made explicit.
Net-Zero Energy Building (NZEB) design: Even a single-building energy model needs the surrounding UHI as a boundary condition. A 3 K UHI means the 30 °C design air temperature is effectively 33 °C, inflating cooling loads by 15–20%. This tool helps designers estimate the local UHI envelope around a site and back-calculate the envelope performance (green roofs, high-reflectance facades) required for LEED, CASBEE or BREEAM ratings, all of which now score UHI mitigation explicitly.
Climate-change adaptation (CCA) cost-benefit analysis: According to IPCC AR6 cities will likely warm by another 2–4 K by 2100, and the combined effect with UHI could push local summer maxima above 45 °C in some districts. While this tool is a macro empirical model, it provides quick first-cut ΔT savings for each mitigation lever, useful as a front end to scenario-based loss-estimation work by municipalities and insurers. Singapore's "Cooling Singapore" project embeds a similar formula in a planning GIS to prioritise actions.
Education and citizen science: UHI is a classic STEM topic — students measure their own neighbourhood temperatures and compare with Oke's prediction. Because this tool exposes the formula alongside the visualization, learners can experiment with "what if my school painted its roof white" and see the predicted classroom-yard temperature drop on the spot. It works as a STEAM resource that bridges geography, physics and social studies in NGO and municipal workshops.
Common misconceptions and pitfalls
The most common trap is confusing Oke's empirical value with measured air temperature. Oke's formula gives the potential maximum UHI from population alone — it's a statistical macro model, not a physical prediction for a specific night at a specific site. The actual UHI varies strongly with wind direction, terrain (cold air pools in valleys), sea-breeze penetration and the space-time distribution of anthropogenic heat. Treat this tool's number as an order-of-magnitude estimate for the city scale. Concrete heat-stress assessments need mesoscale meteorological models such as WRF together with AMeDAS or ground stations.
Next, cool roofs are not universally beneficial. High-reflectance roofs reliably cut summer cooling loads, but they also reflect solar gain in winter, raising heating loads (the "cooling penalty"). In long-winter cities like Sapporo or northern Europe the annual net energy benefit may be negative. Lawrence Berkeley National Lab studies show clear benefit south of about 35°N latitude, with case-by-case evaluation needed further north depending on insulation and heating fuel. Glare and the local thermal impact of reflected light on neighbouring buildings also deserve attention.
Finally, do not assume "more greenery is always better". Greenery cools by day through evapotranspiration, but very dense canopies suppress nocturnal radiative cooling, acting as a heat sink that keeps temperatures up overnight. Evapotranspiration also raises humidity, so part of the WBGT benefit is offset. Species choice (deciduous trees admit winter sun), placement (don't block ventilation corridors) and size (one large park beats many tiny pocket parks) all matter. A single "% green" metric can mislead. The linear correction in this tool is a first-order approximation only — detailed siting needs a GIS coupled with microclimate models.
How to Use
Enter population count (popNum) and urban extent range in kilometers to establish baseline UHI using Oke's empirical model ΔT_base = 0.41 × log₁₀(P) − 0.34 K
Input building/structure density (auNum) and albedo range (0.1–0.4 for asphalt/concrete) to calculate surface radiative forcing corrections
Specify impervious surface fraction (impNum, 0–100%) and green cover percentage (arNum) to model evaporative cooling deficit and thermal mass accumulation
Set wind speed range (auRange, 0–5 m/s) for mechanical mixing effects that suppress vertical temperature gradients
Execute calculation to obtain total ΔT estimate in Kelvin and UHI severity classification (weak/moderate/strong)
Worked Example
For Berlin (population 3.8 million, urban area 891 km²): Oke base ΔT_base = 0.41 × log₁₀(3,800,000) − 0.34 = 2.86 K. Adding 65% impervious surfaces with albedo 0.15 (bituminous roads) contributes +1.2 K. Green cover of 22% reduces this by −0.4 K via evapotranspiration. Mean wind speed 3.2 m/s decreases UHI by −0.3 K. Estimated total ΔT ≈ 3.4 K: city center reaches 24.2°C while rural surroundings measure 20.8°C on identical summer days.