Saturday, April 19, 2014

Evaluating Natural Daylight Levels

After visiting one of our new early childhood development sites, we noticed the inside of the stimulation rooms (classrooms) were a little dark. They buildings are supposed to be naturally daylit, but nobody on the design team knew anything about estimating daylight or optimizing the design. Subsequently, I’ve reviewed the design and taken light levels in the field. My initial conclusion was that the light levels in the center of the rooms at the floor level often met or exceeded recommended levels (primarily because the windows extend very low to the floor below typical vision glazing and light reaches this spot from multiple directions), but at 1 meter off the floor and in many of the corners the levels were below recommended levels. The Illuminating Engineers Society (IES) recommends 50 foot candles (500 lux) on the writing surface for schools for visual comfort and productivity. Based on the 4 sites I measured, the levels with full sun are typically:
20 - 30 FC in centre of room at 1 meter
60 - 80 FC in centre of room at floor
5 - 10 FC on bench in “front” of room

With these light levels, activities low to the ground in the center of the room (building is designed for children 0 - 6) will be well lit with daylight on sunny days and most overcast days, but children in the corners of the room will have less light than ideal. While a couple of the sites have electric lights available to help alleviate this condition (two, 13 watt CFLs without a fixture), we should advise teachers in all sites to focus art projects, reading, and other visually sensitive tasks away from the bench area.
There are a number of potential ways to improve the lighting levels if doing a re-design, as well as some options to improve levels in the already constructed buildings. We decided to go with painting the interior brick white to improve reflectance. Each site has three stimulation rooms so we painted the two longest walls white from floor to ceiling in two of the rooms at one site and re-did our testing. Light levels were almost double in the painted rooms and significantly improved light levels. All sites have since been painted (some not yet to the ceiling as in the image below). While the bench area is still darker than is ideal, the rooms are much improved as a result. At other sites that are nearing the end of construction, we're going to remove some of the brick vent holes at top of the front wall and replace with a framed, translucent plastic window, which should bring the daylight levels up to recommended levels.
The lesson learned in this exercise is that daylight modeling or at least crude daylight factor calculations are critical for buildings intended to be naturally lit.
Extra info (sent to my supervisor when trying to raise the issue):
An easy way to evaluate natural lighting is daylight factor (DF), which is the ratio of outside illuminance over inside illuminance, expressed in per cent. The higher the DF, the more natural light is available in the room.
The general rule is a room needs to achieve at least 2% DF to be considered daylit, though this is still considered gloomy and electric lighting is needed most of the day. From 2 to 5% the daylighting is better, but electric lighting is still needed up until 5% for optimal visual comfort.
Using the crudest rule of thumb method of estimating daylight factor (DF = 0.1 * Glazing Area / Floor Area), it looks like we would just be above the 2% “daylit” threshold as we get 2.7% DF (13.7 m2 glazing / 50.8 m2 floor). Unfortunately, this is overly optimistic in our case for a number of reason. Daylight factor is the sum of three components: direct lighting component (DC), externally reflected lighting component (ERC), and internally reflected lighting component (IRC) such that DF = DC + ERC + IRC. The rule of thumb metrics assume typical office building values for all variables. There are a few problems with this method as we need to account for:
  • Many of the windows and one door are shaded from most direct sunlight by roofs above
  • All of the masonry vent openings are deeper (22 cm) than they are tall (8 cm) so let in no direct sunlight most of the day
  • Most interior surfaces are dark and non-reflective and standard calculations assume partially reflective white ceilings and light colored interiors
As a result, most of our windows and openings have 0 direct lighting component because of the overhangs (good for avoiding heat gain, but also less visible light), we have very little externally reflected daylight since there are no surrounding buildings other than the others we’ve built with non-reflective exterior surfaces, and we have little internally reflected lighting as our interior materials (especially the ceiling) are darker and less reflective than a typical office. We do have the benefit of very clear glazing in the windows with higher than typical visual transmittance (VT) and of course no glazing in the ventilation openings.
Multiple field measurements on the overcast day in Site A showed a range of 1% to 2% DF in the center at 1 meter and about 2.5% to 5% DF in the center on the floor. DF calculations at other sites were not possible as it needs to be overcast and low enough direct sun levels to not overload the meter, but they confirmed the Site B assessment by showing full sun measurements in line with what was expected. 
To get up to the 5% daylight factor for the entire room (suggested target), we’d ideally incorporate a combination of increased opening size, especially up high where the contribution to daylight factor is greater, and lighter and more reflective interior surfaces. Even adding the colored paint in the current rooms has already brightened the space a lot compared to the pre-painted condition. We’ll see the impact of the white walls in Site A. 


  1. Nathan,

    We can help and model your space using various software packages including Diva for example.

  2. Thanks Mike. I don't think UNICEF has any budget for this and the design has been tweaked three times now to improve daylight (interior painted white, windows placed high on front wall, and new design incorporating a clerestory) and it should be much improved. Modeling during design would definitely have been my preference, but I wasn't around during the initial design. Best regards.