The potential environmental advantages of the compact city - transport and space efficiencies - are counterbalanced by the environmental conflicts of close living and working (for example noise and pollution).
This paradox highlights the poles of the urban debate: cramming versus sprawl.
An assessment of building energy use provides an example of the paradox of the compact city. Densification gives potential energy savings. When buildings share party walls and have a reduced surface-tovolume ratio, the result is reduced conducted heat loss. But disadvantages include increased obstructions to daylight, sunlight and natural ventilation as well as the increased likelihood of sound transfer.
The question is, what is the overall energy implication of densification? To answer this we need to consider domestic and non-domestic building types separately.
Energy demand in housing is dominated by space heating, which on average accounts for 60 per cent of the total. In dispersed developments with greater solar access, passive solar design has greater potential to reduce space heating demands, and therefore overall energy use. In cities, constraints on orientation and greater obstructions will limit this solar potential.
Heating energy for an urban house (with an average angle of elevation of 30degrees to the top of the obstruction) is about a third more than an unobstructed, greenfield, passive solar house.
However, a heating saving of 40 per cent can be reached when comparing detached housing with apartments. Therefore, the way to increase density and energy efficiency simultaneously is to increase 'compactness' of the urban fabric while maintaining a limited building depth (about 10-12m) and, where appropriate, a solar orientation to ensure access to light, sun and air.
For energy use in housing, the arguments for and against densification are finely balanced. As the average obstruction angles increase above 30degrees, the balance will begin to swing against densification.
However, an average obstruction angle of 30degrees equates to a theoretical floor area to plot ratio of up to 2.5 (assuming the terrace or courtyard form with a plan depth of 10m). This results in a density of about 200 dwellings per hectare (DPH) - well above the 50 and 75 DPH discussed by the Urban Task Force . Th is compares w ith current new housing at an average of 25 DPH. Thus relatively high housing densities can be achieved before a negative energy impact becomes significant.
Non-domestic buildings Here the picture is quite different. Let us consider office buildings - the predominant, non-domestic building type.
Unlike housing, space heating is not the main issue. The building energy breakdown for a typical air-conditioned office in the UK is 44 per cent for air-conditioning, 34 per cent for lighting and 22 per cent for heating. It is clear that avoiding or reducing air-conditioning is a primary objective for energy efficiency. The secondary approach is to reduce reliance on artificial lighting by increasing and exploiting daylight availability.
Avoiding air-conditioning and increasing daylight availability are complementary in terms of building form implications. They both point towards shallow-plan building forms to enable natural ventilation and daylight penetration. This might suggest that densities would decrease for low-energy offices compared with deep-plan buildings. This is not necessarily the case. The avoidance of air-conditioning may save up to 1m depth of services zone per floor (ie 25 per cent of the building volume assuming an initial 4m floor-to-floor height) and a further 3-5 per cent of the volume for reduced plant size.
Therefore, avoiding air-conditioning can save a maximum of about 30 per cent of the building volume. This may more than compensate for the reduced plan depths resulting in, for example, courtyard or finger-plan forms. The benefit of avoiding air-conditioning through adopting shallow plan depths is typically a reduction of up to 40 per cent of the primary energy. This is an enormous potential advantage.
Although there are limitations to using natural ventilation in the urban context, design strategies can limit the need for air-conditioning. The most obvious is that it may only be necessary to mechanically ventilate the street-facing zone (or locate less noise-sensitive accommodation there), allowing the rest of the building to be naturally ventilated from a quieter and cleaner courtyard, garden or atrium.
Having established that it is in principle possible to maintain a given density of development and significantly reduce energy use by avoiding or limiting airconditioning, we can investigate the energy impact of increasing the density of office buildings.
If we consider two possible strategies for increasing density - increasing building depths and increasing building heights - some initial analyses can be carried out.
For a low-energy, naturally-ventilated office (with mechanical fresh air supply for deep-plan areas) the total energy use doubles when the plan depth is doubled from 12m to 24m. However, for an efficient air-conditioned building the energy use increase is only 20 per cent, although the total energy is almost three times larger than the naturally ventilated option at a depth of 12m.
Therefore, increasing density by increasing building depths will inevitably result in increased energy use for office buildings, although the relative increase is less for air-conditioned buildings. Nevertheless, deep-plan, air-conditioned buildings will typically consume twice as much energy as equivalent mixed-mode buildings, suggesting that avoidance of air-conditioning by improving the urban microclimate is a key factor.
The alternative strategy for increasing density is to increase the average building height (or reduce the spacing, which will have the same consequence for sunlight and daylight availability).Lighting energy is significantly affected on all orientations, increasing by about 50 per cent for an obstruction of 30degrees.The heating load on the south facade almost doubles when the obstruction is 30degrees. Cooling is marginally reduced. Total energy use would be expected to increase with increasing densities.
One of the techniques developed by the urban environmental research team at Cambridge University's Martin Centre enables the energy assessment of real and complex urban form. This technique, 'LT Urban' , comb ines the LT energy model with image processing to extract data related to building form for large urban areas. To demonstrate the technique and its application to establishing the relationship between urban density and building energy use, a 400 by 400m part of London is used. As we are primarily interested in built form, numerous assumptions have to be made about the detailed characteristics of individual buildings, such as glazing ratio, U-values and systems.These have been standardised, based on a study of the area and making informed estimates where necessary. The base-case urban form is then altered by adjusting the building heights to produce a range of urban densities from half to double the existing.
The energy consequences of increasing the average obstruction angles are significant - for example, a 10degrees increase in obstruction results in an increase in energy use of about 10 per cent. For a built density range of gross plot ratios from 1.25:1 to 5:1 the results show that doubling the density typically increases energy consumption by about 25 per cent for this section of city.
A major effort of the Urban Task Force is to alter people's perception of urban living in the UK. By increasing urban populations through the development of brownfield sites it is imagined that there is the opportunity to increase 'the quality of life and vitality that makes urban living desirable'. A causal link is suggested but not demonstrated. Increasing the urban population is not likely in the first instance to improve the urban environment, and at worst, will increase pollution in the short term before the necessary level of investment in transport infrastructure is made. The result may therefore be increased non-domestic building energy use due to the adoption of mechanical ventilation or air-conditioning.On the other hand, urban housing can be an energy-efficient option to densification. In both cases, the link between densification and the urban microclimate continues to be an issue to be researched.
Dr Koen Steemers is a director of the Martin Centre for Architectural and Urban Studies, Cambridge University Department of Architecture, and a director of Cambridge Architectural Research Ltd, 01223 460475. He is co-author of Energy and Environment in Architecture, E&FN Spon, 2000.
Contributions by Carlo Ratti and Darren Robinson