Carbon reduction in existing buildings: a transdisciplinary approach

Background

Clear policy mandates as a response to climate change are emerging in many countries around the world. The UK is one of many that are actively pursuing a national policy of substantial carbon emissions reduction. Given the large percentage of energy consumed by existing buildings (for space and water heating as well as for lighting and appliances), the existing building stock is becoming one of the key targets for public policy, and research has shown that interventions in existing building stocks can substantially reduce carbon dioxide (CO₂) emissions (e.g. Ürge-Vorsatz et al. 2007). It has become increasingly clear to policy-makers that both significant reductions in the energy demand of buildings and significant increases in their energy efficiency will be needed. Although each country has a different composition of building stock (differences in age, construction, density, scale and composition, quality, climate, etc.), there are lessons and strategic approaches that can be shared between different countries.

The 2008 Climate Change Act established legally binding targets for reductions in UK CO₂ emissions of 80% by 2050 from 1990 levels (H. M. Government 2008). The Committee on Climate Change has set a timeline and interim carbon-reduction targets, which includes an extended and stretch scenario reduction of 15–18 MtCO₂ by 2020 through energy-efficiency measures in the existing (i.e. 2005) residential stock (Committee on Climate Change 2008). This represents a cut in residential emissions of around 11.5–13.8% by 2020, with further planned reductions occurring until 2050. The achievement of even this first tranche of reductions will require the implementation of national, regional, and local strategies involving a combination of technical measures to improve the building stock itself and social interventions to influence attitudes and behaviours towards the use of energy. In pursuit of these goals, the emerging public policy, the ‘Great British Refurbishment’, was announced in February 2009. The aim is to reduce emissions from the domestic stock by over 80% by adopting a whole-house, street-by-street, and community-level approach scalable to over 1 million buildings a year by 2020 (H. M. Government 2009).

The framing of the problem of energy demand and CO₂ emissions is crucial to its eventual success. The way in which technical interventions in buildings, such as higher insulation standards, improved boiler efficiencies or integrated renewable energy technologies, can directly affect carbon emissions is in principle relatively well understood. Yet it is an unavoidable fact that, despite many technical improvements to the UK building stock, CO₂ levels continue to rise. There are many reasons for this. For example, interventions, such as insulation measures, are often poorly installed and such interventions may lead to changes in behaviour, for example more extensive heating to a higher temperature. The result is that heating energy demand reductions are often lower than expected. Some increases in emissions can be ascribed to economic growth, which leads to more or larger dwellings, which tend, over time, to contain more electrical items, and items that are also more energy intensive. Thus, there are a number of entangled and interacting economic, technical, social and behavioural factors at play. To make progress at policy, strategy and implementation levels, it is vital to investigate effects such as these, to understand how people use energy in their home, their attitudes to energy use, the effects of energy costs and income levels, and the effectiveness of building regulations, etc.

CaRB research project

This special issue features seven papers and a commentary concerned with reducing the carbon emissions of existing buildings. All but one of these articles are by UK academics who were partners in a UK research consortium called ‘Carbon Reduction in Buildings: A Socio-technical, Longitudinal Study of Carbon Use in Buildings’ (CaRB),1 which sought:

• to improve the understanding of how people actually use energy in buildings

• to formalize this understanding in models that describe the current domestic and non-domestic building stock and the patterns of energy use

• to produce tools to assist policy-makers, consultants and others in their efforts to reduce national CO₂ emissions

The project was undertaken during a period of rapidly intensifying interest and action by government and others in the area of climate change, carbon emissions and building energy use. For example, utility data at local authority level became available and there was the appearance of smaller, cheaper and more reliable monitoring equipment that enabled electrical energy demand to be recorded at short time intervals. These factors acted as a spur to particular avenues of enquiry, some of which were not envisaged at the project's initiation.

CaRB is distinguished from the other Carbon Vision Buildings consortia,2 and indeed from much previous work on energy in buildings, by its emphasis on the collection of field data of various types to understand the existing UK stock of domestic and non-domestic buildings; by longitudinal studies to understand how energy use has changed over time; and by bringing a transdisciplinary perspective to the problem. This is not to say that modelling is not important, but the research team took the view that the current technical models (which are well established and widely used for design, regulation and management) may be flawed by their poor treatment of the human dimension to the energy demand problem. Monitoring and both quantitative and qualitative household interviews were key to understanding this human dimension. Readers might like to reflect on the validity and utility of this approach when they read the papers in this issue.

To bring new perspectives to the energy demand problem, the CaRB project brought together staff from diverse backgrounds who worked closely together for a prolonged period of time, staff from engineering and building science of course, but also social scientists, mathematicians, statisticians, environmental scientists, architects and psychologists, to name but a few.3 The authorship of the papers in this issue is reflective of the cross-university transdisciplinary working within CaRB.

This CaRB project was designed to feed into and impact directly onto public policy. Results were fed through to government via three papers for the UK Foresight Project on Sustainable Energy Management and the Built Environment run by the UK Office of Science and Innovation, though briefings to the Department for Business, Enterprise & Regulatory Reform (BERR) and the Department of Energy and Climate Change (DECC) and their chief scientists, by responding to various government consultation exercises, and by contributing to the development of the forthcoming UK Building Regulations. The home energy survey instrument, which underpinned the work reported herein by Shipworth et al., was used in the Energy Demand Reduction Projects (also called the smart metering trials) run by two energy suppliers (Scottish and Southern Energy, and E.On UK) and the monitoring and data evaluation methods were used as part of the Carbon Trust Micro-CHP Accelerator project. The partners are now taking the work forward through projects funded by the Engineering and Physical Sciences Research Council (EPSRC) and others worth in excess of £20 million, which includes the UK Doctoral Training Centre for Energy Demand Reduction and the Built Environment.

The papers in this special issue are just a small sample of the CaRB work, which included 21 refereed journal articles and 37 refereed conference papers as well as bi-annual progress reports and 21 formal reports to the Carbon Trust. Recognizing the imperative of making rapid progress towards reducing emissions and thus the need for rapid dissemination of information to engage other researchers, policy-makers and stake-holders, output from the project is made available through the CaRB website (http://www.carb.org.uk).

Introducing the special issue

As an introduction to the papers in this issue, it is useful to examine the similarities in the messages that can be found in the six papers concerned with UK housing stocks. Firstly, the two papers concerned with modelling are examined. On the face of it, the empirical modelling of Summerfield, Lowe and Oreszczyn and the numerical modelling of Firth, Lomas and Wright seem to be irreconcilable, the former indicating that past energy-efficiency initiatives have not reduced the stock's energy demands, whilst the latter shows clearly the energy savings for space heating that interventions can have: but closer inspection proves insightful.

Then the papers concerned with monitoring are reviewed. That by Summerfield, Pathan, Lowe and Oreszczyn concerns the longitudinal study of energy use in 36 Milton Keynes homes, whilst the substantial paper by Shipworth, Firth, Gentry, Wright, Shipworth and Lomas examines the data from 358 homes in the CaRB nationwide energy survey to understand how occupants use their heating systems. Using the modelling results, the potential impact of the observations for national carbon emissions is explored.

The implications for energy consumption behaviours and practices and for the marketing of energy-efficiency interventions are then discussed based on the paper of Crosbie and Baker. Some initial findings from the CaRB work involving half-hourly monitoring of 310 non-domestic buildings are presented in the paper of Brown, Wright, Shukla and Stuart.

The final paper by Marsh, Larsen and Kragh, enables the similarities between the perceptions derived from the CaRB papers, to be compared with the Danish energy demand reduction context and some reflections are made thereon.

Finally, this editorial considers the commentary by Lowe and Oreszczyn and their ‘thoughts’ on the broad context of UK buildings research and how the research environment of the UK needs to evolve to address the challenges of climate change and building energy use.

Domestic energy demand modelling

The first modelling paper by Summerfield, Lowe and Oreszczyn takes a new look at the national annual energy use data of the housing stock. The annual delivered energy since 1970 and quarterly totals delivered energy since 1998, for the whole UK domestic stock, are studied. The authors, in keeping with the philosophy adopted in earlier works,4 seek to find a parsimonious model to explain the trends observed. They begin with the four-parameter empirical model presented in the UK Domestic Energy Fact File (DEFF) (Utley and Shorrock 2009), which relates delivered energy to external temperature and energy price as well as heating system efficiency and dwelling heat loss. By progressing stepwise and eliminating variables that are shown to be insignificant in their regression analyses, and introducing and then subsequently eliminating other plausible variables, they arrive at two models: the Annual Demand Energy Price and Temperature (ADEPT) model and the Seasonal Temperature, Energy and Price (STEP) model.

Both models explain the observed total delivered energy in terms only of external temperature (annual or seasonal) and energy price. They observe that when energy price was introduced as a variable, the stock's ‘heat loss coefficient and heating system efficiency were no longer significant predictors’. And yet, and importantly, the ADEPT model predicts the annual delivered energy with an R ² = 0.76, much better than the DEFF model's value of R ² = 0.48; the STEP model explains more than 99% of the variance in the quarterly energy demand data.

Investigations indicate the plausibility of the predictions of these simple models: the calculated short run energy price elasticity was –0.2 (a 10% reduction for each 50% rise in price), which is consistent with others' estimates (Hunt et al., 2003);5 the derived heat loss coefficient for the stock was 240–320 W/K, which is in line with the expected value (Utley and Shorrock 2009); and the delivered energy change with external temperature change is about half the change in delivered heating energy predicted by building physics models: i.e. a reduction in demand of about 1 MWh/year per 1°C change in ambient temperature. Taking the measured 2005 demand of 21.7 MWh (at 7°C), as given in DUKES (Department of Energy and Climate Change (DECC) 2009), this represents a reduction of 4.6% per 1°C; the authors state approximately 5%. This is, they say, to be expected since delivered heating energy, which is the temperature sensitive component, is about 60% of the total delivered energy at 7°C.

This paper has important implications for efforts to reduce the energy demand of national housing stocks. As the authors note, there is:

no evidence to date to indicate that changes in energy demand are beyond those expected due to variations in external temperature and energy price [alone].

This brings starkly into question the assumption, inherent in the DEFF model, that energy-efficiency improvements have led to, and are responsible for, the reduction in demand of the UK housing stocks. The paper's authors are therefore sceptical of estimates, based on modelling, that seek to predict the energy consumption that would now be occurring had past energy-efficiency improvements not been made (e.g. as in the BRE Domestic Energy Fact file; Utley and Shorrock 2009). The authors state:

improved energy efficiency in the domestic stock may be having an impact however [their model suggests] the impact is small compared to and/or confounded with the effects of warmer winter temperatures and occupant response to the cost of energy.

It is also possible that energy-efficiency improvements are not visible in the STEP and ADEPT models because they are collinear with one or more other variables over the period analysed (R. J. Lowe, private communication, 2009).

The two empirical models are a first attempt to develop a method for setting a target for the ‘Great British Refurbishment’ and a method for tracking, on a quarterly basis, the movement towards (or otherwise) this target. The approach enables changes due to the inherent improvement of the stock's energy efficiency to be disaggregated from any demand reduction due to higher energy costs and elevated temperatures, e.g. due to climate change.6 But the authors are cautious because the data they have since 2005 are too few to detect recent energy-efficiency improvement the stock and price sensitivity might not be sustained if, for example, the higher prices lead to demand reductions through energy-efficiency improvements rather than more transient behavioural changes. They also note that on their own the ability of the models to identify the effectiveness of specific policy measures is ‘highly limited’, but as one component of a suite of models they have a useful role to play (R.J. Lowe, private communication, 2009). They note also that real progress can only be made through large-scale representative surveys of dwellings and occupants which ‘includes energy demand and internal temperature’ data. Interestingly, this perception was a central argument for initiating the CaRB project in the first place (see above); this paper adds further weight to that argument.

The question of ‘take-back’ has become a term used as a sort of ‘catch all’ for the frequently observed higher-than-expected demand of dwellings. This phenomenon can be monitored before and after refurbishment or by comparing interventions' effects within model predictions. The classic case is that after insulation is installed dwelling occupants may heat more of their house, and to a higher temperature, because the heating system is now capable of doing this and it can be done at much lower cost than before the refurbishment. Summerfield, Lowe and Oreszczyn conclude that the delivered energy change with temperature is broadly in line with expectations based on the physics of heat loss but that ‘take-back’ may also be playing a role. Between 1970 and 2008 though, many UK homes will have been extended, loft spaces converted to living space and conservatories added. Thus, over time, smaller homes have been replaced by larger homes. This trend alone would lead to lower-than-anticipated heating energy reductions, in the warmer winters, i.e. in recent years, than might be expected, when energy use is expressed on a per-house (rather than on a per-floor area) basis. It is just possible that this is a contributory factor to the results obtained in the paper (as well as gradually increasing average internal temperatures, i.e. classical take-back), but arguably it is incorrect to call these other effects take-back (because the extensions to heated volume were not determined primarily by fuel price, insulation levels, etc.).

Continuing this theme, although the stock of UK homes has, in principle, become better insulated over time, the expected reductions in U-value and draft reduction are not achieved in practice. These factors too will contribute to the lower-than-theoretically-expected energy savings and the effect might well correlate with time (since 1970) and thus ambient temperature. Other similarly correlating effects might also be imagined: greater use of lighting and entertainment equipment when it is colder outside, for example. Neither of these factors, nor the ‘increase in space’ factor, may correlate strongly with temperature in their own right, but together they could lead to a flatter-than-expected regression line when energy demand per household is plotted against temperature. None of this alters the fact that as a nation we are failing to achieve the energy demand reductions we seek, but it is important to our understanding of why this is so.

The broader point is that there could be value in developing a framework that enables the many and diverse factors that lead to higher-than-expected (i.e. compared with theoretical models) energy use to be identified and defined, estimates placed on the magnitude of their individual effects, and their cumulative consequence calculated. Ways of reducing the largest and most pernicious ‘under-performance factors’ could then be developed.

The paper by Firth et al. might provide insight into how this could be done. It takes a conventional approach to modelling the English housing stock, using BREDEM-8 as the calculation engine and dividing the stock into 47 archetypical dwellings. Like other models before, this Community Domestic Energy Model (CDEM) takes its input data from a number of well-known sources.7 Using the model and the average climate data for the period 1971–2000, the average CO₂ emissions are calculated as 5827 kgCO₂ per dwelling. The Department for Environment, Food and Rural Affairs (DEFRA) figure for UK homes in 2004 is 6100 kgCO₂ per dwelling, which is good agreement.8

This paper seeks to determine the sensitivity of the CO₂ emissions predictions to changes in the 27 primary input parameters that ‘drive’ the BREDEM model. This has rarely been done, but by elucidating the sensitivities a number of avenues of enquiry open up. Most obviously, the parameters to which the model is most sensitive are exposed so that, when modelling the stock, attention can be focused on accurately determining these values. Conversely, those parameters to which the stock's energy demand is relatively insensitive can be estimated to lower precision. Not so surprisingly, seven of the most influential factors and their associated sensitivities9 are:

• heating demand temperature (7.4% per 1°C increase in the set point of all homes)

• length of heating period (5.7% per h longer)

• external air temperature (–6.2% per 1°C increase)

• floor area (0.4% per m² increase in average floor area)

• external wall U-value (20.1% per unit change in stock average value)

• window U-value (5.8% per unit increase in stock average value)

• infiltration (17.3% per unit increase in air change rate)

Interestingly, the value of –6.2% change in CO₂ emission per 1°C change in external temperature is reasonably close to the approximate 5% change in the national energy demand calculated, by the quite different method, in Summerfield, Lowe and Oreszczyn, as discussed above. The higher sensitivity is precisely what would be expected from a theoretical calculation of delivered heating energy that assumes no consequential occupant behaviour change (see the previous discussion on take-back etc.). The paper breaks down the sensitivities by house type and age and, not surprisingly, this illustrates that detached dwellings and those that are less well insulated are more sensitive to parameter changes.

Usefully, the paper illustrates that, for limited variations of these key parameters, their effect on CO₂ emissions is linear and that changes in one parameter have only a tiny effect on the sensitivities of others. The paper uses this to show the accumulated effect of the underperformance of elements in a national refurbishment programme might have, even without the ‘take-back factor’ coming into play.

As its authors note, in future it would be most useful to explore the range over which the linearity and superposition principle holds (or rather, since absolute precision is not required, more-or-less holds). Were it to hold over a broad range of parameter values, some useful insights could be gained. For example, it becomes possible to examine the classical take-back question. Using the sensitivities from Firth et al., if the average wall U-value of the national stock were to fall by, say, 0.6 W/m²K, i.e. from its current value of 1.3 W/m²K to 0.7 W/m²K (e.g. as a result of a national refurbishment programme), then the theoretical reduction in emissions would be: 20.1% × 0.6 = 12%. But, if this were to be associated with a 0.4°C increase in internal set-point temperature, and perhaps an increase in air change rate of 0.1 (because refurbished dwellings would be perceived as ‘too stuffy’!), the result would be an emissions reduction of 12% – 7.4% × 0.4 – 17.3%×0.1 = 7.3%, i.e. a take-back factor of (12 – 7.3)/12×100 = 39%. Thus, the ‘rule-of-thumb’ take-back factor of 50%, even when considering the entire national stock, could be entirely plausible.

Importantly this, and the forgoing, discussion show that some of the insights gained from the empirical approach of Summerfield, Lowe and Oreszczyn may be substantially the same as those that might be deduced from an appropriately applied conventional stock model. When, as here, radically different models give similar results, one gains greater confidence in their predictions and in the broad thrust of the observations and policy advice that flows from them.

Energy demand monitoring

Summerfield, Pathan, Lowe and Oreszczyn illustrate the insights that might be gained from conducting longitudinal studies of energy use. They report on 36 homes in Milton Keynes, UK, that were monitored for 18 months from January 1989 to April 1991, to capture the hourly electricity and heating energy demands, and also during the CaRB study in 2005/06 or 2007 to capture the monthly energy use. The homes, 20 detached and 14 semi-detached, were built in 1987 to high-energy standards (equivalent to the 1995 Building Regulations for England and Wales). Data on the floor area of the dwellings were collected as well as information about the household and the electrical energy equipment in the home; comparable data were not available for the 1989/91 study. The study enabled a direct measurement of how energy use has changed over the 17-year period and some observations could be made about the causes of any changes and the way these might relate to household type.

Taking the group as a whole, the small observed increase in energy demand was not significant, but significant differences were detected by dividing the 36 homes into three groups on the basis of their total energy demand in 1990. In the high-energy-using group, gas consumption had increased a little, but this could be explained by the increase in floor area, up 10% from 120 to 132 m² per dwelling. In the other groups there was no discernible change in gas consumption. It was concluded, therefore, that the energy-efficient homes of the 1990s had continued to perform in the same (energy-efficient) manner over the years. It was also evident that no proactive energy-efficiency improvements had been made over the years (though windows and boilers had been upgraded as demanded by general wear and tear).

The average electrical energy consumption of the high 1990 energy-consuming group was three times that of the twelve lowest energy consumers. More tellingly, the electrical energy demand of the high-energy users had increased significantly, by 72% over the 17 years, or 4% per annum. This change was very statistically significant even when normalized for floor area. The electrical energy demand of the other two groups was not significantly different. These observations have substantial implications for energy policy, suggesting that high-energy consumers, which consume, compared with others, even more energy over time, could usefully be the target of energy efficiency and other measures. The CDEM model also indicated that larger detached properties consume more energy. But the sensitivity study of Firth et al. also showed that they were also more ‘responsive’ to energy-efficiency measures. The two papers thus reinforce each other in suggesting that energy policy might usefully focus on ‘fuel-hungry’ households to reduce overall CO₂ emissions.

The increased electrical energy demand of the high-energy users observed in Summerfield, Pathan, Lowe and Oreszczyn was attributed to a big increase in the electrical appliances owned by the households in 2005/07 compared with 1989/91 (modems, flat screen televisions, dish washers, and tumble dryers) and to differences in lifestyle: home offices, with associated computers, fax machines, etc. had appeared since 1989/91. This theme is taken up in the paper of Crosbie and Baker (see below).

The substantial paper by Shipworth et al. is the first to explore the data collected in the national survey of home energy use conducted as part of the CaRB project. The paper seeks to understand the heating patterns (i.e. on/off times) and heating set points and the relationship of these, if any, to the house (i.e. its built form and the energy-efficiency of the fabric) and the energy controls on the system (thermostats and controls). This is an important issue because heating duration and set point are two of the most important factors in determining energy demands for space heating. Proper control of the extent and duration of home heating could be an important plank of carbon-reduction programme.

The paper focuses on the data from 358 of the 427 households interviewed that had a gas- or oil-fired central heating system. Such systems are found in 83% of the housing stock and account for 52% of CO₂ emissions from the sector. The study investigated the data from both the household interviews and from internal temperature loggers in the living rooms. Protocols were developed to detect, from the 45-minute data, the times when the heating was on (basically, when the temperature was rising) and the thermostat set point (basically, the peak temperature reached). The reported thermostat settings were collected from 172 of the households and all 358 households provided statements of the daily system on/off times, system controls (thermostatic radiator valves, room thermostat, timer, controller, etc.), house type (detached, semi-detached, etc.), roof insulation levels, window type (single or double-glazed), and extent of draught stripping.

The average thermostat set point estimated from the loggers was 21.1°C, which is comfortingly close to the BREDEM assumed value of 21°C, although the inter-house variation was ‘enormous’ having a standard deviation (SD) of 2.5°C. This variability, if it is indeed representative of the stock as a whole, introduces an uncertainty (SD) into stock CO₂ predictions of about 18% (assuming that the sensitivity from Firth et al. of 7.4% per 1°C for heating set point holds this wide range of values).

The estimated number of heating hours (from the loggers) was 8.2 (SD = 1.5 h) on weekdays and 8.4 (SD = 1.5 h) at weekends. The weekday value is close to the assumed BREDEM values of 9 h but much lower than the BREDEM assumption of 16 h heating at weekends, suggesting that weekend heating energy might be being substantially overestimated by BREDEM-based models. This would seem like an observation that demands further investigation.

An important finding was that the heating system was on for an average of about 1 h more in the detached houses monitored than in mid-terraced houses, and this difference was statistically significant. Thus, yet again, the CaRB work finds evidence that homes occupied by more affluent families10 are operated in a way that leads to higher energy demands. The paper's authors suggest that social marketing programmes might be explored to discourage the use of higher thermostat settings.

The finding that homes with double-glazing and draught stripping were warmer might indicate that they can be more easily heated to the desired temperature or, on the other hand, that some of the energy efficiency is being ‘taken back’ by the occupants as higher space temperatures. Again, the question of take-back arises.

The work found no statistical difference in the heating set points between homes with and without thermostatic temperature control and no significant difference in the heating period of homes with and without timers. (Although on average the homes with a thermostat were heated to a higher temperature and those with a timer for longer.) These results call into question the idea that improved controls will lead to more efficient use of energy in homes and reinforce the notion that understanding human behaviour and how humans interact with energy technologies is important.

These matters are considered by Crosbie and Baker, who explore the human side of energy-efficiency technologies and begin by questioning why well-known energy-saving technologies are shunned by householders. They interviewed the adult inhabitants of homes in four case study groups. Three of the studies were also the subject of monitoring within the CaRB project, and they include some of the Milton Keynes homes described above. Thus, the studies made adventitious use of the access to dwellings afforded by the technical studies and because of this it was not possible to select the case study households and so conduct a controlled study, say on the basis of household composition. Thus, as Crosbie and Baker note, their conclusions are tentative11 but their qualitative social science investigations do offer a valuable human perspective that complements the technical studies.

Importantly, this paper reinforces the point that new homes, even if they are energy efficient, are more likely to be bought because they fulfil all the customary requirements (location, size, interior layout and aesthetics) and that energy-efficient refurbishment is more likely to be undertaken to improve the quality of the home, in some sense, such as improve comfort or raising the house's value, and to save money, rather than for environmental reasons. In short, lifestyle issues are a stronger motivational factor than environmental considerations.

The paper explores the use of compact fluorescent bulbs in some depth and concludes that lifestyle factors, aesthetics and the quality of the light are the overriding concerns and that, notwithstanding the obvious cost and environmental benefits, if lighting does not meet people's expectations in terms of style and function it will not be used.

The lifestyle issue is double edged. On the one hand, measures that have positive lifestyle attributes may be readily taken up, such as energy-efficient fridges and freezers that are also stylish and which bear a trusted brand name; even when they are more expensive. On the other hand, energy-efficiency measures that have no obvious lifestyle benefit do not have the lifestyle hurdle to jump over in order to gain acceptance; this is the case with, for example, efficient boilers, cavity wall insulation, etc., which can match their less efficient counterparts in style, function and often cost. But this means that such things cannot capitalize on the strong motivation for action that lifestyle impulses bring: encouragement to adopt energy-efficiency technologies needs other instruments such as the building regulations and cost subsidies. In short, lifestyle factors may not apply to some tried-and-tested energy-efficiency technologies.

Against the backdrop of this thinking, one wonders if and how the interventions planned for the ‘Great British Refurbishment’ might capitalize positively on the lifestyle motivator. Sadly, one cannot help but think that the massive disruption and the worries about potential downstream post-installation problems when compared with the small future financial benefits (which might not materialize if energy bills do not fall following refurbishment), will significantly hinder a Great British Refurbishment.12 Are there some lifestyle benefits that could be attached to the worthy but dull energy-reduction technologies?

Many of these observations have been made by others and a limitation of the Crosbie and Baker work is that the information about the house occupiers is limited and so the way in which information, expectations, lifestyle considerations, and design and ergonomics should be tailored to different individuals and different households cannot be established. This would seem an area ripe for exploration, not in a general sense but in the very specific circumstances of the UK and its housing stocks.

The final monitoring paper by Brown, Wright, Shukla and Stuart presents some emerging findings from an analysis of half-hourly gas, electricity and water consumption data collected from non-domestic premises operated by Leicester City Council. Results are presented for a subset of data from 310 premises, which have a total of 1349 metering sites13 and which, since monitoring began in 2001, have produced over 8 million items of data. Schools, offices and libraries make up most of the sample, but some industrial units, commercial premises and shops are also included.

One of the most striking features of the study was the incidence of heating in buildings over weekends when they were unoccupied. Of the 85 sites that had been connected to the monitoring network just before 2004, and for which all three data items were available, 26 (31%) were heated during unoccupied weekend days in February 2004. Of these, 14 were continuously heated, that is, there was no time controller installed.

The paper also describes some long-term trends in the baseline electrical energy loads for 25 buildings between 2001 and 2008. Of these buildings, 16 showed an increase in base load and nine a decrease. Ignoring the two with the greatest increase and greatest decrease, the rates of change varied from –8% to 15% per annum; the mean was 5%. Of the five primary and five secondary schools in the sample, all but one secondary school showed an increased energy demand, which is consistent with the increasing use of information technology.

The authors also note buildings where the heating was on in summer, where there was no thermostatic control of temperatures and no, or faulty, boiler instrumentation such as ambient temperature compensation.

The paper's authors suggest that longitudinal empirical data from non-domestic buildings can play a vital role in energy management at the local authority level, but investment in energy managers with the resources to implement demand reduction measures is also vital. This call for monitoring mirrors that made above with regard to the domestic stock.

An overseas perspective

Much can be learned by studying work that is being undertaken in different countries, and Marsh, Larsen and Kragh provide a Danish perspective on the problem of curbing energy demands in homes. Denmark has made some impressive advances due to consistent policy over more than 30 years, a strong research base, diffusion to construction stakeholders, as well as a clear communication strategy to and acceptance by inhabitants. Since 1975, although there has been an increase in the total housing stock's floor area of 53%, the total energy demand for space heating has actually decreased by about 19%; this considerable achievement is in contrast to the findings of the Summerfield, Lowe and Oreszczyn paper cited above that can see no change in the energy used to heat a UK home.

But in Denmark, as in the UK, home electrical energy use has increased significantly since 1975, by, on average, 2.3% per annum. Marsh et al. focus on this and attribute the rise to the shift towards a knowledge-based society. One wonders, however, whether this drift is the primary cause; many of the new electrical gadgets found in homes are simply associated with entertainment, ‘lifestyle’ and convenience, e.g., larger televisions, digital reception boxes, games consoles, cordless phones, and dishwashers. The idea that a knowledge-based society will, per se, lead to increased energy use lends an air of inevitability to continued increases in electrical energy demand, a view that will require further examination.

In the context of new houses, the paper considers the value of including all energy use within the scope of building regulations. This idea, of course, is in tune with the general shift of Danish, and UK, regulations, from regulating heating energy alone, by prescribing the performance of individual fabric and system components, to a more holistic performance-based regulatory approach that covers not only heating energy demand, but also other energy demands. This represents, as the authors put it, a shift from a narrow to a newer and broader paradigm.

One area that the Danes are concerned about is the risk of increased energy use for space cooling as homes are better insulated. Thus, the 2006 energy regulations for Denmark, which are based on calculating primary energy demand using the program Be06 (which is a steady-state monthly programme like BREDEM) requires the calculation of the electrical energy required to maintain the house below 26°C. Thus, the regulations convert overheating risk into a primary energy equivalent, and then combine this with the energy demands for space heating, hot water and building services (pumps, fans, etc.) to get the primary energy demand.

The authors estimate, however, that in new housing only about 50% of the energy use is covered by this 2006 standard – the other 50%, for lights, appliances, etc. is not. They speculate therefore about whether or not lights and appliances built into the property at the time of sale should also be regulated and likewise energy-management systems, alarm systems. This begs some interesting questions: would regulation of these other items simply deter builders from installing them, because no matter how efficient, they would result in a house that is calculated to have a higher energy demand than one which omits them; and yet leaving out energy-efficient equipment opens the door for the retrofit of less efficient alternatives? And how would this shift in scope affect the potential application of regulations to a whole house refurbishment?

The paper thus opens up an interesting discussion – which aspects of new homes should and should not be regulated? One viewpoint is that some, perhaps most, electrical equipment found in homes cannot and should not be addressed by building regulations. Thus, a truly holistic paradigm for national domestic energy demand control might, in fact, only be achieved by capping or rationing domestic energy use and one way to do this, as Summerfield, Lowe and Oreszczyn's paper shows, is through price. But this is all purely speculative and reflections on the effectiveness, implementation, management, and social and economic consequence of such schemes are outside the scope of this Editorial.

The question of research capacity

Finally, this special issue concludes with a closely argued commentary by Oreszczyn and Lowe which reflects on, as they put it, ‘one of the most difficult problems facing the UK over the next 30 to 40 years: how to decarbonize the built environment’. They examine this assertion not so much in terms of the technical challenges, but, rather, by considering the capability, capacity and funding mechanisms of the UK research base. Research, they argue, will have a pivotal role to play in finding solutions, helping to ‘formulate and evaluate policy’ and in assisting ‘industry to deliver and manage progress’. In this Editorial it is only possible to provide a brief glimpse of the Commentary's contents and so readers are recommended to read and digest its content for themselves. Overseas readers may well find that it bears on their own domestic research and funding situation.

As background, the authors note that there has been a gradual decrease in the CO₂ emissions form buildings over the past 40 years of about 7.5 Mt/year (20% or so compared with the 1970 figure), which has been due to decarbonizing the fuel supply, primarily through the move to gas for heating. However, this ‘trick’ cannot easily be repeated and over the same 40 years the actual energy use of buildings has risen, in dwellings by about 33%, i.e. about 1% per year.

The authors note the need to make cuts of 80% by 2050, but they do not prescribe a route to achieving this. One way is, perhaps, to make cumulative reductions of about 1.1% per year for the next 40 years, i.e. annual reductions of about the same magnitude as the current annual increase. And as the paper by Summerfield, Pathan, Lowe and Oreszczyn (and others elsewhere) shows, electrical energy use is, in some sectors, increasing at an even greater rate than this. Reducing energy demand therefore requires both the reversing of an adverse trend and putting in place mechanisms that will lead to real emissions cuts.

The authors assert that ‘the primary function of the next ten years will be to establish effective supply chains for energy-efficiency over the following two decades’: they liken the decarbonization task to ‘a Manhattan Project for energy’.14 In stark contrast the authors note that ‘the perception in government [is] that energy-efficiency in buildings is straightforward and requires minimal investment’ and thus reducing energy demand in buildings is seen as ‘a key, relatively easy and cost-effective win’. As evidence of this the authors offer a graph from the Intergovernmental Panel on Climate Change (IPCC) synthesis report that suggests that, at a cost of about US$20/tCO₂, buildings can yield emissions cuts that are at least three times greater than those possible in any other sector, including the transport and energy-supply sectors.15 The authors think this optimistic view is misguided, and the papers in this special issue, which illustrate some of the intricate links between energy demands, energy cost, technical interventions and the building occupants, would lend support to this view.

The main thrust of Lowe and Orezczyn's paper is, though, to examine the UK's buildings research base. The authors note that research has a vital role to play in the decarbonization drive, but they see the current research base as far to small and lacking in capacity ‘more closely resembling a cottage industry than a mature academic sector’. They argue that in the face of the massive increases in funding that are likely in the coming years, there is ‘a deficit of human capital’. They also see a mismatch between the type of research required to reduce energy demand rapidly and the research expected by the important funding bodies: the former suggests that researchers should be embedded within the processes and organizations that are creating the built environment rather than standing outside of them looking in as neutral observers, and that longer and more substantial projects that hold successful transdisciplinary teams together, are needed.16 There is also a perceived need to create the mechanisms by which the models and methods developed will be more readily shared with others, currently the competitive nature of research militates against such openness.

In the UK, over the last few years there has been a general move, by research funding bodies and the EPSRC in particular, to fund larger/longer projects that enable the methods and insights from the social sciences and humanities to be brought to bear on problems that, at root, are in the engineering and scientific domain. CaRB was in the vanguard of this funding evolution. Nevertheless, as funding came to an end, the transdisciplinary team of 22 academics and research assistants and eight PhD students, established over a 54-month period, fragmented and so the momentum, common understandings and joint ways of working that were established are slowly dissipating. Some of the strongest and most productive partnership will endure though, and carry their work forward. This special issue, and the many other papers produced by the CaRB team, and the papers still to come, will hopefully be a valuable legacy.

Conclusions

Based on the CaRB projects as a whole, the partners have begun to formulate some of the key messages, and from the papers presented in this special issue, one might observe:

• that reducing energy use by implementing energy-efficiency measures is more challenging than might be expected

• that there is a shortage of information and tools by which the effectiveness of policy can be assessed

• that developing refurbishment strategies that target specific properties, such as larger detached properties, might improve cost-effectiveness

• that demand reduction initiatives might usefully address the design and the marketing of products and services

• that regulation may not be the appropriate mechanism for controlling energy use in the complex socio-technical system that is the occupied dwelling.

To these one might add:

• that demand reduction research cannot be addressed effectively by a single discipline and that transdisciplinary teams with long-term funding can make a substantial impact

• that valuable new insights can be gained by collecting hard data, i.e. measurement, monitoring, questionnaires and surveys

• that useful insights can be gained by examining activities in other countries

Finally, the CaRB project suggests that any large-scale energy demand-reduction programme for buildings should have, as a vital component, a nationwide ‘monitoring’ programme. Without such an evidence base, to guide the refurbishment programme, to assess progress and to ground models, effort may, as the history of energy demand work tells us, be ineffective or, worse, counter-productive.

Acknowledgements

This work forms part of the Carbon Reduction in Buildings (CaRB) Consortium. CaRB had five UK partners: De Montfort University, University College London, The University of Reading, the University of Manchester, and The University of Sheffield, who were joined, in the later stages, by Loughborough University. CaRB was supported by the Carbon Vision initiative, which was jointly funded by the Carbon Trust and Engineering and Physical Sciences Research Council (EPSRC), with additional support from the Economic and Social Research Council (ESRC) and the Natural Environment Research Council (NERC). The partners were assisted by a steering panel of representatives from UK industry and government. For further details, See http://www.carb.org.uk/. The CaRB project gained much from the enthusiasm and drive of its Scientific Director, Tadj Oreszczyn, whilst this Editorial benefited from comments made by a number of CaRB colleagues. This Editorial was written whilst the author was a Visiting Fellow at Clare Hall, Cambridge University, supported by a Research Fellowship from the Leverhulme Trust (Grant Number RF/0334).

Notes

The CaRB consortium was one of three funded for four-and-a-half years within the Carbon Vision Buildings (CVB) programme, funded jointly by the Engineering and Physical Sciences Research Council (EPSRC) and the Carbon Trust. The other two were Technology Assessment for Radically Improving the Built Asset Base (Tarbase), which used modelling to explore the potential carbon reductions that alternative energy-efficiency techniques and new and renewable energy systems could yield for different classes of buildings; and Building Market Transformation (BMT), which investigated how measures that could cut emissions from buildings by 50% might be taken up as widely and as quickly as possible (see http://www.carb.org.uk).

See http://www.epsrc.ac.uk/ResearchFunding/Programmes/Energy/Funding/CarbonVision/BuildingsProjects.htm/.

There was a core contingent of 22 permanent academic staff, led by the current author (CaRB Principal Investigator and Financial Director) and T. Oreszczyn (Scientific Director from UCL) and five co-investigators from De Montfort, Reading, Manchester, and Sheffield universities. They were assisted by eight PhD students and an able Project Manager. The terrific momentum generated by the project attracted the attention of eight other staff in the participating universities who were able to capitalize on the data and models that the consortium produced.

Using simple models to explore large-scale and potentially complex systems is a feature of Lowe's earlier work on building stock modelling (the 2 house model).

However, the authors would expect the long run elasticity to be of the order of 1, but whether it would be, and after how long, is uncertain (Lowe, 2003). Lowe notes that the initial response would comprise mainly coping behaviours, while longer-term responses would include investments in fabric, appliances, infrastructure and supply chains (R. J. Lowe, private communication, 2009).

Lowe also offers the observation that ‘non-linearities in underlying anthropo-physical processes, technological discontinuities and emergent social trends’ means that conclusions drawn from models, particularly very simple empirically based models like ADEPT and STEP, may appear to be true today but may not be true tomorrow. ‘What is interesting about such simple models is the fact that they may be able to serve as early indicators of such discontinuities’ (Lowe, private communication, 2009).

The Standard Assessment Procedure inputs, the tables associated with the BREDEM documentation, the English House Condition Survey, and the Market Transformation Programme.

Scaling the CDEM prediction for the period 1971–2000, during which the average temperature was 6.24°C (from Summerfield, Lowe and Oreszczyn), to a value for 2004 (an average temperature of 6.9°C), using the sensitivity produced by the model (–6.2% per 1°C change in ambient temperature), gives: 5827 – [(6.2 × 5827)/100] × (6.9 – 6.24) = 5589 kgCO₂ per dwelling, a difference from the DEFRA figure of –8.4%. However, one might expect the heating energy demand of the English stock, as predicted by CDEM, to be lower than that for the UK as a whole (DEFRA) because the latter includes dwellings in Scotland, Wales, and Northern Ireland where it is cooler.

Expressed here as percentage change in house-averaged CO₂ emissions per unit change in the parameter in question. The exact ranking of the most influential parameters depends on how the sensitivity is expressed.

Detached homes are, in general, more expensive than an equivalently sized semi-detached or terraced home.

Therefore, the paper should be read in the light of the methodological constraints and with the intent of confirming or denying, in a general sense, observations made by the studies of others.

Future ‘losses’ are a much bigger concern to people than potential future gains.

Premises may contain multiple buildings, for example a school campus, each of which is monitored or be a single building with sub-monitoring of the individual premises within it.

The Manhattan Project employed thousands of people to achieve a single objective: the development of the atomic bomb.

Even with an investment of US$100 per tonne, these two areas cannot approach the savings possible from buildings.

Rather than the usual cycle of bid, fund, hire, work, report, and then the loss of the team before, hopefully, repeating of the same cycle. Historically, government laboratories and some other organizations were to some extent freed from this cycle, but these have largely disappeared in the UK.