Is the daylight factor an outdated evaluation method for modern design? John Mardaljevic argues that climate-based modelling is a far more effective approach

Climate-based daylight modelling was only developed at the turn of the millennium. This relatively new technique involves using climate-based data to predict the quantity and quality of daylight in a space. The advantage of using a climate-based lighting simulation over more traditional techniques, such as the daylight factor approach, is that it takes into account sunlight and skylight in a way that is consistent with our experience of the naturally lit environment. It will also provide data founded on absolute values for the evaluation of building designs and the calibration of daylighting schemes.

The only quantitative measure of illumination currently commonly used for the evaluation of daylight is the daylight factor (DF). The daylight factor is the ratio of illuminance at a given point in a scene (internal or external) to the unobstructed external horizontal illuminance under the Commission Internationale de l’Eclairagre (CIE) standard overcast sky. The DF is usually expressed as a percentage.

The CIE standard overcast sky has a zenith brightness three times that of the horizon, but there is no variation in brightness with compass orientation, and, of course, there is no illumination of any kind from the sun. Thus, the daylight factor solution gives the same answer regardless of the orientation of the scene, or indeed of its location. So, for a given scene, north-facing in St Petersburg will have the same solution as south-facing in Miami.

The majority of daylighting problems can be better solved through the application of climate-based techniques than through standard approaches. In addition, the practitioner armed with these techniques will discover that they are now in a position to solve many other building and design problems – for example the installation of, say, a photovoltaic array in a dense urban setting where shading is inevitably a concern. Likewise, defining the optimum balance between daylight provision and solar control in the design of low energy facades can only be achieved at the design stage if the practitioner has the facility to predict realistic measures of these quantities in the first place.

Whilst some of the applications for climate-based modelling might be considered futuristic, such as evaluating the performance of say, active shading devices or electrochromic glazing, the majority are longstanding problems encountered daily by practitioners. Six years on, climate-based daylight modelling is starting to gain acceptance through its successful application to live projects and its role as an ‘engine’ to help formulate new daylighting metrics.

The need for new daylight modelling techniques based on more meaningful metrics than the traditional daylight factor has already been recognised by some in the industry. Consultant Arup’s lighting and daylighting engineers have developed an assessment tool based on the theoretical principles formulated by the author and applied these novel techniques to a range of real design projects to provide information that could only be guessed at before. Arup have used this method for:

  • The development of a master plan to determine the impact of proposed new developments on the surrounding area and to reorganise a development for the greatest on and off-site solar benefit
  • Planning a building layout to optimise the location of an atrium in an overshadowed building
  • Optimisation of a building’s envelope to minimise summer solar loads while maximising winter solar access
  • Developing landscaping to ensure a new streetscape does not impact on existing trees
  • Designing daylight for museums to ensure sensitive exhibits do not receive too much light

To show how effective climate-based modelling can be, this article describes its practical application to two fairly typical design evaluation problems, while a third example is a theoretical case study to demonstrate the application of a new, climate-based daylighting metric called ‘useful daylight illuminance’.

Case study 1

Daylight injury assessment for the New York Arts Students League

Founded in 1875, the Arts Students League (ASL) boasts an alumni list that reads like a Who’s Who in American art, from Winslow Homer to Mark Rothko and Jackson Pollock.

ASL artists, teachers and students have all placed great value in the daylight afforded by the skylights to the top floor studios. However, the building’s precious daylight was under threat from a proposed new high-rise development immediately to the west of the building (see graphics, right, where the skylights are indicated in blue and the proposed tower in green).

The challenge was to evaluate the potential impact of this propsed scheme on the two art studios by determining a meaningful measure of the reduction in daylight levels it would cause, quantifying the impact for various alternative designs and assessing the limits of mitigation. In other words, the challenge was examining how factors such as the facade reflectance of the proposed high rise building might influence the overall level of daylight injury to ASL’s studios.

Standard evaluation methods could not fully address these concerns. For example, the skylights are north-facing and receive hardly any direct sun. This renders a shadow pattern study, which reveals direct sun exposure at a few selected times of the year fairly pointless. The daylight factor approach was also rejected here because it failed to take

into account the character of illumination in the studios as a result of sky conditions and the potential for reflected sunlight from nearby buildings.

Daylight injury

The solution I offered was an assessment of the daylight ‘injury’ using New York climate data. Total annual illumination is a measure of all visible daylight energy incident on a surface over a year. The simulations to predict this (and the other climate-based measures discussed in this article) are all founded on standard meteorological datasets that are specific to the evaluation locale.

From these standard datasets, it is possible to derive hourly-varying sky and sun conditions for use in lighting simulations. Equally, it is possible to synthesise cumulative luminance ‘maps’ for arbitrary periods (eg annual, monthly periods, etc) that contain the aggregated luminance effect of all the unique hourly sky and sun values.

For the ASL study, separate luminance maps for the annual cumulative sun and the annual cumulative sky were synthesised from the standard climate dataset for New York City.

These cumulative luminance maps were used to determine the sky and sun components of total annual illumination (TAIL) incident on the skylights of the ASL.

The simulations were carried out for the existing arrangement of buildings and with the proposed tower in place, with the tower reflectivity set first to zero (r=0) and then to 50% (r=50), as shown below.

The zero reflectance case determines the reduction of TAIL with the tower acting purely as an obstruction. In the 50% reflectance case, the tower acts both as an obstruction and a reflector of light (sky and sun). A reflectance of 50% is the highest that can be expected for any exposed vertical facade. The effect of intermediate reflectivity values for the proposed tower can be determined from a simple interpolation of the results for the zero and 50% cases.

A meaningful measure

In the results presented below (see Figure 1), the location of the skylights is outlined in the left-most image. In addition to the mean TAIL for each skylight marked on the images, the inset value shows the area-weighted mean in TAIL for both skylights. The area-weighted mean TAILs were 36,946 klux hours for the existing scenario, 23,455 klux hours with a tower of zero reflectance and 29,972 klux for a tower with 50% reflectance.

For the client, the differences in the predicted levels of total annual illumination gave a realistic evaluation of the daylight injury from the proposed tower which they could then use in negotiation with the developers. Furthermore, it was remarked that total annual illumination was a far more meaningful measure of daylight availability than ‘abstract’ quantities such as the daylight factor.

Case study 2

Daylight for a museum in St Petersburg

Climate-based lighting simulation has also been used to predict the total annual exposure to daylight for artworks in a well-known museum in St Petersburg, Russia.

Some measure of the seasonal dynamic in daylight illumination was required for this evaluation. In contrast to the ASL example outlined earlier, the museum has fixed opening hours of 10 am to 6 pm, outside of which blinds/shutters will be drawn.

The method of approach

To create a model, the hourly climate data for St Petersburg was processed to represent local time, including summer daylight savings, and 12 cumulative monthly climate files were created using the period of visiting hours only. Separate climate files were created for the sun and sky components of illumination and monthly cumulative luminance maps derived from each of the monthly climate files.

Simulations showing a hemispherical view of the inside of the room were generated for each of the 24 luminance maps (12 sky and 12 sun maps). The simulations for the sun component were carried out with inter-reflection both enabled and disabled. In this way, it was possible to determine the amount of the total illumination due to direct sun only. The method used for this evaluation readily identifies those months when exhibits are most at risk from direct illumination by the sun.

The total annual exposure is simply the sum of the 24 monthly sun and sky illuminance maps. Graphics of the 3D model are shown above.

The mean illuminance for each month was calculated by taking the sum of the cumulative sun and sky illuminances for that month divided by the number of visiting hours. The results for two points in a west-facing room are given in the plots below the 3D models (see Figure 2, above). The lines drawn between the two illumination plots below and the right-hand graphic indicate the approximate position of the wall marker points where the monthly illumination values were determined.

The values for mean illuminance are given by the total height of the bars in the graphs and the magnitude of any direct sun component is indicated by magenta shading to the lower part of the bar. The inset box in each graph gives the total annual exposure.


Climate-based modelling can supply good estimates of exposure and illumination at any stage of the design process. The cumulative monthly approach described here offers a compact form of presentation and is ideally suited to the scheduling of, say, the seasonal deployment of shading or blinds to minimise exposure to direct sun.

Case study 3 (theoretical)

Useful daylight illuminance – a replacement for daylight factors

The previous case studies simulated the illuminance effect of annual and monthly cumulative skies derived from hourly climate data. The finest level of temporal detail offered by a climate-based lighting simulation is one that predicts time-varying daylight illumination at the time-step (or shorter) of the climate data. For most climate files, this will be hourly and will result in the generation of about 4000 illuminance values (ie number of daylight hours) for every calculation point.

Useful daylight illuminance (or UDI) is a new evaluation scheme I devised to determine meaningful measures of daylight provision from the voluminous mass of illuminance data (ie hourly values at each point for the entire year).

UDI is defined as the annual occurrence of illuminances across the work plane within a range considered ‘useful’ by occupants. This range is based on a survey of occupant preferences and behaviour in day-lit offices with user-operated shading devices. This showed:

  • Daylight illuminances in the range 100-500 lux are considered effective either as the sole source of illumination or in conjunction with artificial lighting
  • Daylight illuminances in the range 500-2000 lux are often perceived either as desirable or at least tolerable
  • Daylight illuminances above 2000 lux were unacceptable – it was around this mark that blinds were drawn and/or dissatisfaction noted.

From this survey, it is possible to define UDI as the annual occurrence of daylight illuminances that are between 100 and 2000 lux.

The UDI of a scheme is applied by determining at each calculation point the occurrence of daylight illuminances that:

  • Are within the range defined as useful (100 lux to 2000 lux)
  • Fall short of the useful range (less than 100 lux)
  • Exceed the useful range (greater than 2000 lux).

Thus, only three metrics are needed to provide a compact representation of the hourly-varying daylight illuminances for an entire year at each calculation point. As such, the UDI scheme preserves much of the interpretative simplicity of the familiar daylight factor approach. Further, it seems plausible that when the UDI is exceeded, this is likely to be related to a building’s propensity for excessive solar gain.

In other words, this simple scheme can provide useful information on the intrinsic shading effectiveness of the building (including features such as facade structures, brise soleil etc) as well as on the provision of daylight.

The effectiveness of the UDI system is demonstrated here for a building with a central light well. It has standard clear double-glazing and the base case version is totally un-shaded (see Figure 3, below). Internal floor, wall and ceiling reflectances were set to typical values. Variant 1 has a 1 m shading overhang on the east, south and west facades. Variant 2 also has this, plus a lantern with shaded top over the light well.

Hourly daylight illuminances at the working plane height across the ground floor were predicted for the base case and for both shading variants using the rigorously validated daylight coefficient (DC) technique. This is an efficient means to determine any number of illuminance values (internal or external) resulting from arbitrary sky/sun conditions, but it requires only a limited number of computationally demanding calculations. London climate data was used to generate the sky and sun conditions.

The effect of shading

The results are presented in Figure 4, right. The top row of images show the percentage of the working year for which daylight illuminances were in the range 100-2000 lux (ie UDI achieved). The plots below the images show the achieved UDI along the east-west transect (the dotted line in the images above).

As well as UDI achieved, these line plots show where the UDI was exceeded and where there was a UDI shortfall. For the un-shaded base case design, the line plots show that the low occurrence of UDI achieved for the perimeter was due to the high illuminances ( above 2000 lux), which are likely to cause discomfort. The same is true for the area below the un-shaded lightwell (base case and variant 1). The UDI plots readily disclose the effect of adding shading to the base case building.

UDI is a very new idea, and there is still some research to be done to determine optimum levels of UDI achieved and, importantly, exceeded, since this is an indicator of the propensity for solar gain. Once this calibration work has been completed, the UDI scheme will, for the first time, offer practitioners the means to quantitatively assess daylight in its totality – the sun and sky together. Thus, it will help designers achieve the elusive balance between moderate levels of daylight (ie that which is beneficial for occupants and can reduce electrical lighting consumption) and excessive levels which can cause visual discomfort and increase cooling loads.

Why daylight design needs to move into the 21st century

While it hardly needs pointing out that daylight is inherently climate-dependent and time-varying, the accepted evaluation method for lighting designers, the daylight factor (DF), takes no account of this everyday fact. The DF persists as the dominant evaluation metric because of its simplicity rather than its capacity to describe reality.

It is depressingly common to see DF studies carried out routinely for almost any locale worldwide. Rarely is a scintilla of qualification given to the appropriateness of the CIE standard overcast sky conditions for the locale under evaluation. Daylight factors have also been used to analyse the performance of light-redirecting devices when it is evident that any measure of their true effectiveness must account for the capacity to redirect sunlight deep into the space (and indeed offer shading near to the window).

The shortcomings of DF-based design is particularly apparent in low energy buildings. The Leadership in Energy and Environmental Design (LEED) scheme promoted by the US Green Building Council encourages the design of low energy buildings. However, there is a growing concern in the US that the DF basis of LEED is promoting over-glazed buildings with cooling costs that are likely to outweigh any savings from daylight. In fact, providing too much daylight could well result in increased use of electric lighting as the blinds are likely to be drawn much of the time.

The example of the imaginary building used to explain the useful daylight illuminances concept (above) presents an interesting case when the LEED criteria of achieving a minimum DF of 2% in 75% of all space occupied for critical visual tasks are applied.
Only the un-shaded base-case building would attain the LEED credit for daylight: a DF of 2% is achieved across 81% of the floor area. With the addition of shading in variants 1 and 2, the 2% DF value is achieved across 72% and 64% of the floor area, respectively. In other words, the
shading needed to lessen the propensity for high illuminance (with its associated discomfort and solar gains) would, for this building, cause it to fail to achieve the LEED daylight credit.

The illuminances for the light-well example used UK climate data, whereas the LEED DF rating applies without modification to all US states – from rainy Seattle to sunny Miami. If the un-shaded base-case produces internal illuminances that are often too high for comfort under the UK climate, how much worse will the conditions be when this building is exposed to the Miami climate? Alongside a climate-based analysis, the daylight factor seems hopelessly removed from the reality of daylight illumination.

John Mardaljevic is a senior research fellow at the Institute of Energy and Sustainable Development (IESD), De Montfort University. E-mail