Back to KPI list
Urban Heat Island
Urban Heat Island
Urban Heat Island
Urban Heat Island
Urban Heat Island
Risksvv0
Responsible:Niccolò Calandri
Layer Urban Heat Island on real sites
Real Estate UK
Catania
Berlino Quarter
Mina el Cerrejon
Urban Heat Island
Overview
The Urban Heat Island (UHI) effect occurs when built-up areas record significantly higher average temperatures than the surrounding rural zones. The phenomenon is primarily caused by the reduction of vegetated surfaces, the high heat absorption of construction materials (concrete, asphalt), and anthropogenic heat production. The KPI shows, for the months from April to August, the distribution of UHI intensity levels, calculated through two indicators: the UTFVI (Urban Thermal Field Variance Index) and the night-time land surface temperature (Night LST).
UHI has direct impacts on human health (heat waves, summer mortality), energy consumption for cooling, urban biodiversity, and air quality. The severity of the phenomenon is directly correlated with the density and type of urban land cover (Oke, 1982). The KPI is inverted: higher UHI intensity values indicate more critical conditions.
Calculated Indicators
1. UTFVI (Urban Thermal Field Variance Index)
The UTFVI quantifies the UHI effect intensity through the variance of the urban thermal field. It is calculated as:
UTFVI = (LST_pixel - LST_mean) / LST_mean
where LST_pixel is the land surface temperature of the pixel and LST_mean is the mean temperature of the analysed area. Positive values indicate pixels warmer than average (UHI hotspots); negative values indicate cooler pixels (mitigation zones).
2. Night LST (Night-time Land Surface Temperature)
The night-time LST directly measures surface temperature during night-time hours (typically 01:30 local time for MODIS Aqua). Night-time temperatures are particularly significant for UHI since urban surfaces retain heat accumulated during the day.
Calculation Methodology
The KPI uses MODIS LST data at 1 km resolution, downscaled to 25 m via a machine learning model exploiting:
- Land Cover (ESA WorldCover, 10 m) — land cover classification
- NDVI (Sentinel-2, 10 m) — vegetation index to estimate evapotranspirative cooling capacity
- Altitude (NASA SRTM, 30 m) — correction for altitudinal thermal gradient
Calculation pipeline:
- LST data acquisition — MODIS Terra and Aqua data (day and night) for April–August months
- Quality filtering — exclusion of cloud-contaminated pixels using MODIS QC bands
- Downscaling — upscaling LST from 1 km to 25 m with ML model (features: ESA WorldCover, Sentinel-2 NDVI, SRTM)
- UTFVI calculation — application of the UTFVI formula pixel by pixel
- Night LST calculation — extraction of mean night-time LST for the April–August period
- Polygon aggregation — calculation of UHI intensity level distribution within the polygon
Analysis period: April — August (most representative months for the UHI phenomenon)
Units:
- UTFVI: dimensionless index
- Night LST: degrees Celsius (°C)
Data Sources
| Code | Name | Provider | Resolution | Availability |
|---|---|---|---|---|
WRD_MAQLT_99 | MODIS Aqua LST | NASA / USGS | 1000 m | continuous |
WRD_MTRLT_99 | MODIS Terra LST | NASA / USGS | 1000 m | continuous |
WRD_ESAXX_99 | ESA WorldCover | ESA / Impact Observatory | 10 m | 2020, 2021 |
WRD_S2XXX_99 | Sentinel-2 | ESA/Copernicus | 10 m | 2017 — present |
WRD_SRTMX_99 | NASA SRTM | NASA | 30 m | static |
Quality Range and Levels
UTFVI
| Indicator | Unit | Range | Inverted |
|---|---|---|---|
utfvi | — | [0, 0.2, 0.4, 0.6, 0.8, 1] | Yes |
Inverted = Yes: a lower value indicates lower UHI intensity and better conditions.
| Level | UTFVI | Interpretation |
|---|---|---|
| A (Excellent) | 0 – 0.2 | UHI effect absent or negligible; area at or below average temperature |
| B (Good) | 0.2 – 0.4 | Slight UHI effect; mild warming above average |
| C (Moderate) | 0.4 – 0.6 | Moderate UHI effect; impact on thermal comfort conditions |
| D (Poor) | 0.6 – 0.8 | Significant UHI effect; health risk during heat waves |
| E (Critical) | > 0.8 | Severe UHI effect; strong warming; critical health risk areas |
Night LST
| Indicator | Unit | Range | Inverted |
|---|---|---|---|
night_lst | °C | [-10, 20, 25, 30, 35, 45] | Yes |
Inverted = Yes: lower night-time temperatures indicate better conditions and lower nocturnal UHI intensity.
| Level | Night LST | Interpretation |
|---|---|---|
| A (Excellent) | < 20°C | Cool night-time temperatures; excellent thermal mitigation |
| B (Good) | 20 – 25°C | Comfortable temperatures; slight nocturnal warming |
| C (Moderate) | 25 – 30°C | Mild temperatures; moderate nocturnal UHI impact |
| D (Poor) | 30 – 35°C | High temperatures; warm nights; thermal stress |
| E (Critical) | > 35°C | Very high temperatures; tropical nights; elevated health risk |
Technical Notes
- Analysis limited to April–August months to capture the summer period, most representative for UHI
- Downscaling from 1 km to 25 m via ML significantly improves spatial resolution
- Night-time Land Surface Temperature (Night LST) maps need to be generated
- MODIS data have a 1–2 day revisit time, ensuring good temporal coverage
- Cloud filtering reduces data availability in areas with high summer cloud cover
References
- Oke, T.R. (1982). "The energetic basis of the urban heat island". Quarterly Journal of the Royal Meteorological Society, 108(455), 1–24.
- Voogt, J.A., Oke, T.R. (2003). "Thermal remote sensing of urban climates". Remote Sensing of Environment, 86(3), 370–384.
- Zhou, D., Xiao, J., Bonafoni, S., Berger, C., Deilami, K., Zhou, Y., ..., Sobrino, J.A. (2019). "Satellite Remote Sensing of Surface Urban Heat Islands: Progress, Challenges, and Perspectives". Remote Sensing, 11(1), 48.
Quality Ranges and Levels
utfvi
ACritical
BLow
CModerate
DGood
EExcellent
0 → 0.19
0.2 → 0.39
0.4 → 0.59
0.6 → 0.79
0.8 → 1
A
Critical0 – 0.19B
Low0.2 – 0.39C
Moderate0.4 – 0.59D
Good0.6 – 0.79E
Excellent0.8 – 1night_lst
(°C)ACritical
BLow
CModerate
DGood
EExcellent
-10 → 19.99
20 → 24.99
25 → 29.99
30 → 34.99
35 → 45
A
Critical-10 – 19.99B
Low20 – 24.99C
Moderate25 – 29.99D
Good30 – 34.99E
Excellent35 – 45Data Sources & References
Sources
WRD_MAQLT_99WRD_MTRLT_99WRD_ESAXX_99WRD_S2XXX_99WRD_SRTMX_99
Methodology
MODIS LST data (Terra: WRD_MTRLT_99, Aqua: WRD_MAQLT_99) at 1 km resolution, downscaled to 25 m via ML model. Downscaling features: Land Cover (ESA WorldCover, WRD_ESAXX_99), NDVI (Sentinel-2, WRD_S2XXX_99), Altitude (NASA SRTM, WRD_SRTMX_99). Analysis period: April–August. UTFVI = (LST_pixel - LST_mean) / LST_mean. Night LST: mean night-time surface temperature for April–August period. Cloud filtering via MODIS QC bands. Two inverted indicators: utfvi [0, 0.2, 0.4, 0.6, 0.8, 1] and night_lst [-10, 20, 25, 30, 35, 45].