Test Reference Years and Design Summer Years Explained

A Test Reference Year (TRY) and a Design Summer Year (DSY) are two different kinds of weather file used in UK building simulation, and they answer two different questions. A TRY represents a typical year of weather, assembled so that annual energy and heating performance can be assessed against ordinary conditions. A DSY represents a warm summer, assembled so that a building can be tested for summertime overheating.

The three design summer years are not three levels of warming. They are three different shapes of hot summer drawn from the historical record. DSY1 is the design summer year for the 2020s, high emissions, 50th percentile scenario¹. DSY2 is built on 2003, a year with a very intense single warm spell¹. DSY3 is built on 1976, a year with a prolonged period of sustained warmth¹. The choice between them changes the result of an overheating assessment, because a building that copes with a short sharp spell may not cope with weeks of sustained warmth.

The stakes are not academic. Increasingly hot summers could lead to a trebling of health and productivity impacts without additional adaptation². The Climate Change Committee commissioned Arup to appraise the current and future risks posed by summertime overheating to the UK housing stock at scale, looking at what factors influence risk, how homes can be adapted or upgraded to mitigate the impacts and how much that costs³. The weather file is the instrument that turns that risk into a number for a specific building.

What a test reference year is and what it is used for

A Test Reference Year is a synthetic year of weather built to represent typical conditions rather than extreme ones. It is the file used when the question is how a building performs across an ordinary year: annual heating demand, seasonal energy use, the sort of figure that feeds into a compliance calculation or an energy model.

The reason a typical year matters is that building energy performance is judged against normal conditions, not against the worst week of the decade. A heating system sized against an extreme cold snap would be oversized for almost every hour it operates. A TRY avoids that by assembling months drawn from across the record so that the resulting year is representative rather than memorable.

This is the same logic that runs through official energy modelling more widely. The Climate Change Committee's 2035 power system modelling is based on an hourly representation of demand and generation, and historical annual weather patterns to simulate future years⁴. The principle is consistent: use a weather year that stands in for the ordinary case, then test the system against it.

For a household, the TRY is the file behind the annual figures that describe how much energy a home is expected to use. It is not a forecast for a particular year and it is not a worst case. It is a yardstick, and its value lies in being the same yardstick applied consistently across buildings and assessments.

Where a project needs a single annual energy number, the TRY is the file that produces it. Where the question shifts to how a home behaves in heat, a different file is needed, and that is where the design summer year comes in.

What a design summer year is and why it exists

A Design Summer Year exists because a typical year is the wrong instrument for a heat question. If a building is being assessed for summertime overheating, a file built around average conditions will understate the problem, because the risk arises precisely in the summers that are not average.

A DSY therefore concentrates on warm weather. It is a summer-focused file, assembled so that a building can be tested against conditions that would genuinely stress it. The London guidance sets out the three variants plainly, and the descriptions are worth taking at face value: DSY1 for the 2020s, high emissions, 50th percentile scenario; DSY2 for 2003, a year with a very intense single warm spell; DSY3 for 1976, a year with a prolonged period of sustained warmth¹.

The distinction between those two historical summers is the heart of the matter. A short, very intense spell tests how quickly a building heats up and how much cooling it needs at peak. A long, sustained warm period tests whether a building can shed heat overnight and recover, or whether warmth accumulates across days and weeks. A dwelling with heavy thermal mass and good night ventilation may perform well in one and poorly in the other.

The Climate Change Committee's work on existing homes takes this seriously. Its appraisal of current and future risks posed by summertime overheating to the UK housing stock at scale considers what factors influence risk, how homes can be adapted or upgraded to mitigate the impacts and how much that costs³. The weather file is what makes that appraisal specific to a building rather than a general statement about the housing stock.

The three design summer years: which weather each captures

The three files are best understood as three different summer shapes rather than three points on a warming curve. Each was chosen because it represents a pattern of heat that a building needs to survive.

The practical consequence is that a single building can produce three different overheating results depending on which file is used. That is not a flaw in the method; it is the point of having three files. An assessment that reports only one result without saying which file produced it is reporting an incomplete picture.

The 2003 summer is remembered for the intensity of its peak, and the 1976 summer for its duration. Those are different physical stresses on a dwelling. Peak intensity drives instantaneous cooling demand and can push internal temperatures past thresholds quickly. Duration drives cumulative heat gain, and it is the pattern that punishes buildings unable to purge warmth overnight.

For a household, the relevance is that the file chosen for an assessment shapes what the assessment concludes about the home. A dwelling assessed under DSY3 is being tested against sustained warmth, which is a harder test for some constructions than a short spell. The result should always be read alongside the file that produced it.

DSY1: the high-emissions scenario behind the standard case

DSY1 is the file most often treated as the standard overheating case, and its full description explains why: the design summer year for the 2020s, high emissions, 50th percentile scenario¹. Each part of that description carries meaning.

The 2020s framing places the file in the current decade rather than in a far future. The high emissions label describes the scenario pathway the file sits within, which is the more demanding of the pathways available. The 50th percentile describes where the file sits within the spread of modelled outcomes: it is the central case, not the most extreme.

That combination is what makes DSY1 the default. It is a demanding but not worst-case summer for the present decade, which is a defensible basis for testing a building that exists now. It is not a future climate projection, and it should not be read as one. A projection models how the climate itself changes over coming decades; DSY1 is a design weather file for testing a building against warm conditions in the 2020s.

The distinction matters because the two are sometimes conflated. A building that passes under DSY1 has been tested against a central high-emissions summer for this decade. That is a meaningful result, but it is not a statement about how the building will perform in the 2050s.

The percentile framing is also worth holding onto. A 50th percentile file is the middle of a distribution, which means half of the modelled outcomes sit above it. A building tested only against the central case has been tested against the central case, and nothing more should be claimed for the result.

A single intense warm spell versus prolonged sustained warmth

The difference between DSY2 and DSY3 is the difference between a spike and a plateau, and it changes what a building is being asked to do.

DSY2, built on 2003, captures a very intense single warm spell¹. The stress is concentrated. Internal temperatures rise quickly, cooling demand peaks, and the building's ability to shed heat at speed is what determines whether it stays within limits. A dwelling with effective shading and rapid ventilation may cope well with this pattern.

DSY3, built on 1976, captures a prolonged period of sustained warmth¹. The stress is extended. There may be no single dramatic peak, but there is little opportunity to recover, and heat accumulates in the fabric across days. A dwelling that relies on overnight cooling to reset may struggle if the nights stay warm.

These are genuinely different failure modes. A lightweight building with good ventilation can perform acceptably in a short intense spell and poorly in a sustained one. A heavyweight building with high thermal mass can perform acceptably in a sustained spell, because it absorbs and delays heat, and less well in a sharp peak where the mass has not had time to buffer.

For a household, the practical implication is that an overheating assessment result is only meaningful alongside the file that produced it. Two assessments of the same home under DSY2 and DSY3 can reach different conclusions, and both can be correct for the weather they represent.

How the two datasets differ and when each applies

The two datasets differ in what they are built to represent, and that determines when each applies.

A Test Reference Year represents typical conditions across a full year. It is the file for annual energy and heating work, where the question is how a building performs across ordinary seasons. A Design Summer Year represents warm summer conditions. It is the file for summertime overheating assessment, where the question is how a building behaves when the weather is hot.

The dividing line is the question being asked, not the building being assessed. The same dwelling may need both files, each for a different purpose. An annual energy figure comes from the typical year; an overheating result comes from the summer year. Neither substitutes for the other.

There is a wider point about how official modelling handles weather. The Climate Change Committee's 2035 work uses historical annual weather patterns to simulate future years, alongside an hourly representation of demand and generation⁴. That is a typical-pattern approach suited to system-level questions. Building-level overheating work needs the summer-focused file instead, because the risk it is testing is concentrated in warm weather.

For a household, the practical reading is straightforward. If the question is what a home uses over a year, the answer comes from a typical year. If the question is whether a home gets too hot in summer, the answer comes from a design summer year, and the specific file used should be stated.

Where these weather years appear in UK building and overheating assessment

These weather files sit inside a wider UK framework of building standards, energy assessment and overheating risk work, and their role differs across the four nations.

In England, Approved Document L sets out the metrics against which new dwellings are assessed, including the target primary energy rate, expressed in kWhPE/m2 per year⁵. That is annual energy territory, which is typical-year work. Overheating assessment sits alongside it as a separate exercise, and that is where the design summer year applies.

In Scotland, the Building Standards Technical Handbook covers buildings, work, services, fittings and equipment not requiring a warrant⁶, and the Scottish Government has consulted on proposed changes to energy standards and associated topics⁷. Scotland has also developed its own heat policy framework, including the Heat in Buildings Strategy published in 2021⁸ and a Heat in Buildings Bill consultation in 2024⁸. The New Build Heat Standard was proposed for 2024⁷.

The Climate Change Committee's overheating work is UK-wide in scope. Its appraisal of current and future risks posed by summertime overheating to the UK housing stock at scale, commissioned from Arup, considers what factors influence risk, how homes can be adapted or upgraded to mitigate the impacts and how much that costs³. That work depends on weather files that represent warm summers, because the risk being appraised is a summer risk.

Energy Performance of Buildings Certificates provide the parallel data stream for England and Wales, published from 2008 to March 2018⁹. EPC data describes the stock; weather files describe the conditions the stock is tested against. The two are different instruments serving different purposes.

For a household, the weather file is the part of the assessment that is invisible but decisive. It determines what conditions the home was tested against, and therefore what the result actually means.

Where the weather files come from and how they are used

The weather files are published as part of the CIBSE weather data set, which is the recognised source for UK building simulation. The London guidance draws on these files directly, describing DSY1, DSY2 and DSY3 in the terms used throughout this page¹. The Climate Change Committee's overheating work also depends on weather data of this kind³.

Access and licensing sit with the publisher. The files are used within accredited simulation software rather than distributed informally, which is why a household will normally encounter them through an assessor or a modelling report rather than directly.

The practical consequence is that the choice of file is usually made by the assessor, and it should be stated in the report. A result that does not name the file is a result that cannot be interpreted, because DSY1, DSY2 and DSY3 test different things.

For a household commissioning an overheating assessment, the file used is a fair question to ask. It determines whether the home was tested against a central high-emissions summer for the 2020s, a short intense spell, or a prolonged warm period, and those are not interchangeable tests.

What this means for a household's energy independence

Weather files are an assessment instrument, not an energy source, and it is worth being precise about what they do and do not change for a household.

They do not reduce dependence on the grid, a supplier or a fuel. A home assessed under DSY1 or DSY3 is still connected to whatever supplies it, and the weather file changes nothing about that. What the file changes is the quality of the information available about how the home behaves, and that information is what any decision about adaptation rests on.

The dependence that remains is the same dependence that exists without any assessment: on the electricity or gas network, on a supplier, and on the building fabric itself. A weather file cannot substitute for insulation, shading or ventilation. It can only show whether they are needed.

Where the files do contribute is in making overheating risk visible before it becomes a problem. The Climate Change Committee's finding that increasingly hot summers could lead to a trebling of health and productivity impacts without additional adaptation² is a statement about what happens when the risk is not addressed. The weather file is the tool that turns that general risk into a specific result for a specific home.

For a household, the honest position is that the weather file is diagnostic. It tells you what conditions the home was tested against and what the result was. It does not change the home, and it does not change what the home depends on.