Using the New CIBSE Design Summer Years to Assess Overheating in London

This article was first published in the November issue of the CIBSE Natural Ventilation Group Newsletter.

Using the new CIBSE design summer years to assess overheating in London

Across the UK construction industry, the standard weather data used for assessing thermal comfort is normally obtained from the Chartered Institute of Building Services Engineers (CIBSE). CIBSE provide two types of weather file, known as the Test Reference Year (TRY) and the Design Summer Year (DSY), with an accompanying range of data formats and geographical locations for each.

So… What is a TRY, what is a DSY, and which one should I use?

To facilitate the creation of characteristic weather files, CIBSE assembled 20 years of real weather data (between 1983-2005) for 14 locations around the UK. These files include information about the air temperature, wind velocity and solar radiation.

The following definitions are adapted from the descriptions given on CIBSE’s website.


The Test Reference Year consists of hourly data for twelve typical (although not necessarily consecutive) months. This data was selected from the 20-year data sets, and then smoothed to provide a composite, but continuous, 1-year sequence of data.

The TRY weather files are intended to enable the likely energy consumption of buildings to be assessed by simulation under typical weather conditions.


The Design Summer Year is an actual continuous twelve month sequence of hourly data, from the 20-year data sets. It represents a median year with a reasonably warm summer. The year selected is the mid-year of the upper quartile, based on dry bulb temperatures during Apr-Sept.

The DSY enables designers to simulate the expected building performance during a year with ahot, but not extreme, summer.

In London, the middle year of the upper quartile was 1989; hence, the London DSY weather file is the actual weather recorded in London in 1989. In other UK locations, a different year is selected; for example, in Norwich the DSY year is 2004, in Cardiff it is 1999 and in Glasgow it is 1997.

TM49 and the London conundrum

Previously, DSY weather files were available for 14 locations around the UK; all of Greater London, including central London and much of Sussex & Essex, being characterised by the London (Heathrow) data.


In 2014, however, CIBSE published technical memorandum TM49: Design Summer Years for London. TM49 addressed three key questions:

  1. Is the DSY a sufficiently warm year, or should a warmer year be used?
  2. How will future climate change affect the suitability of the DSY?
  3. What is the effect of the urban heat island (UHI), and should weather data be supplied for other parts of London?

This detailed technical report concludes with some brief and fairly stark recommendations:

  1. Because … the estimated return period for the current London (Heathrow) Design Summer Year [1989] is just 1-3 years (i.e. summers as hot or hotter than 1989 are expected every 1-3 years), this weather file is not considered sufficiently extreme to provide a basis for overheating assessments in London; it is recommended that … warmer weather data should be used.
  2. Because … it is impossible to prejudge the impact of warm weather on a building in a general sense, it is recommended that … a range of weather files should be modelled, to investigate the sensitivity of the design to different weather conditions.
  3. Because … there are significant climate variations across London, associated with the urban heat island effect, it is recommended that … more specifically appropriate weather files should be used to characterise development in central, suburban and rural locations, respectively.

One major outcome of TM49 was that CIBSE released a new set of ‘Design Summer Year’ weather files for greater London. These additional data sets mean that DSY weather files are now available for three London locations: London Weather Centre (LWC), London Heathrow Airport and London Gatwick Airport, for the following baseline years:

  • 1976 – a relatively intense persistent warm spell
  • 1989 – a moderately warm summer
  • 2003 – a single intense warm spell

These weather files aim to better represent urban (LWC), semi-urban (Heathrow) and rural or peri-urban (Gatwick) locations. The report recommends that

  • London Weather Centre data be used for development in the central zone,
  • Heathrow data be used for development in urban and suburban areas, and
  • Gatwick data be used for more rural areas around the edge of Greater London.

But what does it all mean?

In a recent paper(1), published in the CIBSE journal Building Services Engineering Research & Technology, researchers from University College London and CIBSE tested how the newly-released Design Summer Year weather files impacted the assessment of overheating in a naturally ventilated office in London.

The study found that, regardless of which London weather file was employed, it was very difficult to achieve a CIBSE-compliant design without incorporating useful thermal mass/insulation and automated night cooling.

Once these strategies had been implemented, the Urban Heat Island effect was strongly observed in the form of large differences between similar buildings depending on their location. As demonstrated by Figure 1, which is taken from Virk et al.’s article, the urban heat island effect is predominantly a night-time phenomenon. With almost negligible differences observed in the mean of the daytime temperatures, but 3-4°C differences observed in the mean of the night-time minimum temperatures. Crucially, this emphasises that night-cooling will be far more effective in rural locations, such as Gatwick, because of the cooler night-time temperatures and the greater day/night diurnal variation.diurnaltemplondon

Figure 1: Mean diurnal temperature variation for during 2003 for three London locations.

Interestingly, the choice of thermal comfort metric also had a significant impact on the level of overheating between sites. Buildings in the centre of London are more likely to pass the ‘adaptive thermal comfort’ criteria of TM52, than the fixed temperature criteria used previously. This is because the adaptive model compares internal temperatures to the ‘running mean’ outdoor temperature, which is itself often higher than the static, measured temperature. And results in simulations indicating that occupant will accept higher internal temperatures if the weather has been consistently warm in preceding days.

Whether this perceived willingness to accept higher temperatures truly reflects the attitudes of occupants is probably still up for debate. In terms of compliance with guides and applying the most recent thinking on overheating, however, the adaptive thermal comfort model is definitely something designers should work with.

Owen Connick

Consulting Manager


(1)      Virk et al. Using the new CIBSE design summer years to assess overheating in London: Effect of the urban heat island on design. Building Serv. Eng. Res. Technol. 2015 Vol. 36(2) 115-128.

(2)      CIBSE Guide A: Environmental design. Chartered Institution of Building Services Engineers. 2015.

(3)      CIBSE TM52: The limits of thermal comfort: Avoiding overheating in European buildings.Chartered Institution of Building Services Engineers. 2013.

(4)      CIBSE. TM49: Probabilistic design summer years for London. Chartered Institution of Building Services Engineers. 2014.

Breathing Buildings