You are here

Focus: heat islands

The air temperature is generally higher in cities at night time compared to neighbouring rural areas, a phenomenon which is known as an 'urban heat island'.

The Belgian Royal Meteorological Institute has conducted various studies to evaluate this phenomenon in Brussels. The analysis of temperature readings demonstrates that the urban heat island effect is an existing phenomenon. Furthermore, it becomes more and more pronounced over time, in particular regarding minimum temperatures (night time temperatures). This increase is due in particular to the progressive urbanisation of the Region.

Over the period 1961-1990, a heat island in the centre of Brussels was calculated with an average value of 2.5°C, via the various simulations made, for the minimum temperature. The magnitude of the urban heat island decreases progressively towards the periphery, with a substantial positive impact in green spaces, and in particular in the Sonian forest.

Higher temperatures in city centres

The air temperature (close to the soil surface) is higher in cities than in neighbouring rural areas. This phenomenon is known as an urban heat island (UHI).

Illustration of the characteristic thermal profile of an urban heat island
Following Akbari et al. (1992). "Cooling our communities – a guidebook on tree planting and light colored surfacing.", U.S. Environmental Protection Agency, Office of Policy Analysis, Climate Change Division, Berkeley : Lawrence Berkeley Laboratory

Especially pronounced in city centres, the heat island can be associated with a very local phenomenon, for example in cases of street canyons. These are narrow streets flanked on both sides by buildings and subject to a lateral wind, which does not allow the effective dispersion of heat or pollutants.

The formation and intensity of a UHI depends on various factors, starting with meteorological conditions. The main differences in temperature between cities and the countryside actually occur in clear weather, when there is not much wind, and they are generally more pronounced at the start of the night.

These urban heat islands can be explained by the replacement of permeable soils containing vegetation with buildings and sealed coverings. So, for example:

  • the reduction in vegetation cover and the proliferation of vertical walls increase the surface absorbing solar radiation flows,
  • the use of materials in dark colours for roads and buildings (and therefore a lower albedo in urban areas) leads to a larger absorption of the incoming solar radiation,
  • the radiation entrapment effects within street canyons in city centres are the cause of an increase in temperature in these streets: given the three-dimensional structure of the street, the reflected radiation is not directly emitted into the atmosphere but remains trapped inside the street; the orientation and the gradient of the streets (and the exposure to the sun and associated winds) also influence the extent of the temperature increase,
  • the capacity of the immediate environment to lower daily temperatures by evaporation or evapotranspiration (water and plants), and by shade, is reduced.

These local temperature increases are moreover related to the human activities which are more concentrated in cities (emissions of fuel gas, emissions of warm air by air conditioning systems, warm water circulating in the drains, etc.).

The increase in temperatures related to the heat island effect is likely to lead to disruption both in terms of comfort and health (exacerbation of the effects linked to heat), and in terms of energy consumption (air conditioning) and the associated disturbances (energy consumption and the associated emissions of pollutants into the air).

Night time heat islands have a potentially greater impact on human health, to the extent that - particularly during periods of intense heat - warmer night time urban temperatures are likely to limit the relieving effect after a day of high temperatures.

What about the Brussels Region?

The Brussels Region is a city which presents a level of greenery which is significantly less in the centre than at the periphery. Moreover, it is characterised by an average rate of sealing which progressed from 26% to 47% between 1955 and 2006 (ULB-IGEAT, 2006). Just like other cities, it therefore has the conditions to develop an urban heat island.

The Belgian Royal Meteorological Institute has conducted various studies to evaluate the UHI in Brussels.

The first one was a calculation of the differences between the temperatures measured at Brussegem and Uccle, two stations of the climatological network of the RMI:

  • Uccle is a suburban station, situated in the south of the Region, 6km from the centre of Brussels.
  • The climate station at Brussegem is a rural station, situated 13km to the north-west of the centre of Brussels, in the Flemish Region.

The differences in temperature between the two stations were calculated for the minimum temperatures (night time, upper part of the chart below) and maximum temperatures (day time, lower part of the chart) of each day, during every summer over the period 1955-2006. The long-term evolution of the differences is given by the linear trend lines.

The urban effect on minimum and maximum air temperatures, as an average during the summer, between 1955 and 2006.
Source: RMI, Climate Vigilance Report, 2015

The urban heat island effect is estimated by the differences between the temperatures recorded at the station at Uccle and the rural station at Brussegem during the summer. The lack of data between 1972 and 1979 is due to a suspension of measuring at Brussegem during this period.

The results show that the urban heat island effect does indeed exist in Brussels. Moreover, it becomes more and more pronounced over time, in particular with regards to minimum temperatures (an increase which is 2.8 times faster than is the case for maximum temperatures) (Hamdi and Van de Vyver, 2011). This larger increase of the minimum temperatures in particular can be explained primarily by thermal inertia (the storage capacity of heat) which is higher in the city, combined with a lower albedo of urban surfaces. This delays the cooling of cities during the night, in comparison with neighbouring rural areas. This is coupled with a more limited evapotranspiration effect, as previously explained, (and the associated cooling by evaporation), as well as the more significant anthropogenic production of heat in the city.

Simulations of the spatial distribution of the UHI effect have also been carried out by the RMI (Hamdi et al., 2014). The various publications of the RMI indicated below will provide you with more methodological details if required.

A heat island in the centre of Brussels was calculated with an average value of 2.5°C, via the various simulations made, for the minimum temperature (night time).

Spatial distribution of the Brussels urban heat island at night, as an average over 30 years (1961-1990)
Source : Hamdi et al., 2014

Result of the simulations carried out using the ALARO operational atmospheric model of the RMI, combined with a new surface diagram, with a specific parameterisation for cities and forced by the ERA40 database. The black points indicate the location of the centre of Brussels, the station at Uccle and the station at Brussegem.

The value scale is in °C difference compared with the simulated minimum temperature over rural values (outside of the city).

The highest values are observed in the city centre. The extent of the urban heat island progressively decreases towards the periphery. A substantial positive impact can also be observed in green spaces, and in particular in the Sonian forest.

What is the effect of the urbanisation of the Brussels Region?

The simulations carried out by the RMI also had the aim of identifying the role of the progressive urbanisation in the progressive temperature increase observed at Uccle, by comparing two simulations. The first simulation takes into account the background of the evolution of the rate of soil sealing in the BCR, and the second simulation assumes a hypothetical situation without urban areas within the BCR (rural simulation). The urban effect and its evolution are estimated by subtracting the difference in the temperatures obtained from the two simulations.

The progressive urbanisation of the Region has in fact been the cause of an average rise in the temperature at Uccle of 0.09°C every 10 years.

25% of the summer warming observed at Uccle between 1960 and 1999 is therefore due to an intensification of the urban heat island effect related to progressive urbanisation, rather than local or regional climate change (RMI, 2015; estimate based on the difference between temperatures observed at Uccle and Brussegem). 

What should we expect in the future?

Forward-looking simulations have also been carried out by the RMI (Hamdi et al., 2014 and 2015).

They demonstrate that:

  • Climate change has a limited impact on the intensity of the urban heat island as an annual average, with a night-time increase during the winter and a daytime decrease during the summer;
  • The increase of the intensity of the night-time urban heat island during the winter is related to a forecasted wind reduction by 2050, according to climate simulations;
  • The daytime decrease during the summer is related to dryer soils (which is the cause among other things of an increase in rural temperatures), given the reduction of summer rainfall according to climate simulations;
  • The impact of climate change on the Brussels urban climate will be greater during heatwaves, in combination with the future development of the city. Given that the urban heat island intensifies during a heatwave, the urban population will be more exposed to the urban effect during the summer, since the climate models forecast an increase in the frequency of heatwaves in the future.
Date de mise à jour: 06/12/2017

Etudes et rapports

Hamdi R., Deckmyn A., Termonia P., Demarée G. R., Baguis P., Vanhuysse S. and Wolff E., october 2009, "Effects of Historical Urbanization in the Brussels Capital Region on Surface Air Temperature Time Series: A Model Study", Journal of Applied Meteorology and Climatology, volume 48 issue 10, pp. 2181-2196 (.pdf)

Hamdi R. and Van de Vyver H., February 2011, "Estimating urban heat island effects on near-surface air temperature records of Uccle (Brussels, Belgium): an observational and modeling study", Advances in Science & Research, volume 6 issue 1, pp. 27-34 (.pdf)

Hamdi R., Van de Vyver H., De Troch R. and Termonia P., 2013, "Assessment of three dynamical urban climate downscaling methods: Brussels’s future urban heat island under an A1B emission scenario", International Journal of Climatology, Published online (.pdf)

Hamdi R., 2014, "Impact des changements climatiques dans les villes: Contraste entre stress thermique urbain et rural", Summary of the presentation given for the "Stakeholdersmeeting : 5ème rapport d’évaluation du GIEC – partie 2: ‘Changements climatiques : impacts, adaptation et vulnérabilité’ - Conclusions et actions d’adaptation en Belgique" that has been organised on the 6th of mei 2014 by BELSPO and SPF Santé publique, Sécurité de la chaine alimentaire et Environnement, 3 pages (.pdf, in French only)

Hamdi R., Giot O., De Troch R., Deckmyn A. and Termonia P., 2015, "Future climate of Brussels and Paris for the 2050s under the A1B scenario", Urban Climate, vol. 12, pp. 160-182. (Summary, .pdf)

RMI, 2015, "Vigilance climatique", 86 pages (.pdf, in French and Dutch only)

Vanhuysse S., Depireux J., Wolff E., 2006. "Etude de l’évolution de l’imperméabilisation du sol en Région de Bruxelles-Capitale", ULB/IGEAT on behalf of the Ministère de la Région de Bruxelles-Capitale, Administration de l’Equipement et des Déplacements/Direction de l’Eau, 60 pages (summaried within the chapter "Prévention et gestion des inondations dues aux pluies d'orage estivales " of the 2003-2006 report). (.pdf, in French and Dutch only)