Most Windows are Chronically Over-Shaded
Numerous studies have found that manually-operated window coverings are typically adjusted only rarely, with an average adjustment frequency of less than once per day. The implication is that most people don’t adjust window coverings to compensate for daily changes in the daylight level.
So why do people adjust manually-operated window coverings? There isn’t any large-scale survey data to directly answer that question, but there is plenty of circumstantial evidence that people are motivated to close window coverings far more than they are motivated to open them.
This evidence is in the form of data that shows that manually operated window coverings are likely to be mostly closed, causing windows to be over-shaded most of the time. The data comes mostly from external observations of office buildings, but there has also been at least one large-scale survey of residential window covering usage with corroborating findings. See the References section at the end of this post for specific citations.
Typical Window-Covering Settings in Office Buildings
Most of the research on window covering usage has focused on office buildings, for two reasons. First, window coverings have a greater impact on energy consumption in office buildings than in any other building type because office buildings have a greater ratio of window to wall area than other building types. Second, photographic observation of a single office building can yield data for a relatively large number of window coverings.
The bulk of the data on window covering usage in office buildings has been for horizontal venetian blinds, because they’re the most widely used (and arguably the most versatile) window coverings in both office and residential buildings. Unlike other window coverings, a horizontal blind offers two degrees of adjustment freedom: the slats can be tilted to selectively block daylight from a desired direction, and the slats can be raised or lowered to vary the window occlusion:
Since slat tilt and occlusion have different effects on the shading provided by a blind, much of the data collected on blind usage in office buildings addresses both slat tilt and occlusion.
Average Window Occlusion is High
As might be expected, studies of venetian blind usage in office buildings reveal that occlusion is correlated with variables that drive the need for shading—and that the average occlusion can be very high for sunny exposures.
For example, Inoue et al (1988) found that, for windows from which the radiosity is occasionally high enough to cause discomfort, the occlusion is correlated with the maximum solar penetration depth. And Lambeva and Mahdavi (2007) found that the average occlusion in a given month is correlated with the peak global vertical irradiance over the same month.
Accordingly, in the northern hemisphere, the lowest observed occlusions are typically for north-facing windows, while the highest occlusions are for south-facing windows—and those occlusions are typically significantly greater than 50%.
The following chart shows the average occlusions reported in four seminal papers that investigated the impact of window orientation (note that not all studies addressed all four orientations):
The values are in general agreement: occlusion on south-facing facades averages 70%, while average occlusion on the other facades is somewhat lower but still greater than 50%.
One of the sources of the data of Chart 4—Foster and Oreszczyn (1999)—is unique in that it cites a composite occlusion index that depends on the slat tilt as well as the occlusion per se (neither of which is reported separately). Each blind’s occlusion index is thus greater than the occlusion by the degree to which its slats were closed. The values plotted for Foster and Oreszczyn in Chart 4 are therefore actually the square roots of the reported occlusion indices, under the assumption that the relative slat tilt would have been roughly proportional to occlusion.
Unfortunately, only a couple of studies provide any information at all regarding the distribution of occlusion values from office to office and building to building:
- Rubin et al report occlusion distribution data for approximately 700 manually operated blinds across 6 buildings and two orientations (north and south) in 1978, well before mainstream use of video display monitors.
- Sutter et al report occlusion distribution data for 11 motorized blinds with southeast orientation in 2002, with the occupants of the subject offices spending most of their time looking at video display monitors.
The following chart shows the Cumulative Distribution Functions (CDFs) of the occlusions observed in the two studies:
Despite the substantial differences in the study parameters, the CDF for the south blinds of Rubin et al are very similar to those for the southeast blinds of Sutter et al. The interpolated distributions suggest that, for a south or southeast exposure, 50% of blinds will have an occlusion greater than ~70%, and 25% of blinds will have an occlusion greater than 80%.
Slats are Mostly Closed
Slat tilt is arguably even more important than occlusion in understanding occupants’ shading preferences and the impact of venetian blind usage on interior daylight levels.
It’s easier to adjust a blind’s slat tilt angle than to vary the occlusion by raising or lowering the slats, so slat tilt is presumably a more sensitive indicator of shading preferences.
And virtually all horizontal blinds used in office buildings raise from the bottom upwards, which means that windows are occluded from the top down. This is significant because the depth to which daylight will penetrate a room is proportional to the height of the daylight aperture above the floor. As a result, even a partially raised blind can substantially reduce the floor-averaged level of admitted daylight, depending on how its slats are tilted.
But while slat tilt statistics are arguably more significant than occlusion statistics, they’re much harder to collect. Occlusion for even a large number of blinds can be quickly and accurately measured from photographs of building facades, but slat tilt angles are difficult to determine via image analysis, even with sophisticated image processing software. For that reason, only a couple of image-based Venetian blind usage studies have considered slat tilt:
- Rubin et al were the first to photographically collect slat tilt statistics, but were able to quantize tilt to only two designations (“open” and “closed”). The report provides no information on the approximate tilt ranges spanned by each designation, and the use of a single “open” designation means that nothing can be inferred about the frequency of settings that permit a sky view versus those that permit a ground view. Finally, the study was performed before video display monitors came into general use, which could reduce the relevance of the results to today’s office environments.
- Foster and Oreszczyn combined measured slat tilt (quantized to three settings) with occlusion (quantized to five settings) to yield a composite occlusion index, but do not report slat tilt and occlusion statistics separately. Further, the quantized slat tilt settings represent just the magnitude—and not the direction—of tilt away from horizontal.
To date, only a couple of studies reporting slat tilt statistics make the distinction between tilt settings that afford a sky view and those that afford a ground view:
- Bulow-Hube (2001) manually estimated slat tilt (quantized to 15 degrees in 12 bins) via direct inspection of motorized exterior Venetian blinds on south-facing windows. However, the data were obtained as part of a controlled experiment in test rooms, not by long-term monitoring of operating offices.
- Sutter et al used a networked data collection system to collect slat tilt data from motorized interior blinds on the southeast facade of an operating office building. While the reported tilt is quantized to just three bins, separate bins are defined for positive (ground view) and negative (sky view) tilts.
Bulow-Hube and Sutter et al had different research objectives and used different experiment designs, but their findings regarding slat tilt for sunny window exposures are similar. The following chart shows the frequency of slat tilt reported by both Bulow-Hube and Sutter et al, with the 12 bins of data from the former report down-sampled to the same three bins used in the latter report:
The relatively fine quantization used by Bulow-Hube enables determination of the distribution of observed slat tilt angles. The following chart plots the cumulative of that distribution:
Note that more than 90% of the blinds had a positive slat tilt angle (which blocks the sky view), and that the median slat tilt angle (i.e. the slat tilt for CDF of 0.5) was about +40 degrees. The mostly positive tilt slat tilt angle is consistent with the hypothesis that most shading adjustments are aimed at blocking daylight glare.
The slat tilt angle of most blinds can approach ∓90 degrees when fully closed, so it might seem that a tilt of +40 degrees would still admit plenty of daylight. However, that’s not necessarily the case.
Visible Transmittance versus Slat Tilt
The visible transmittance (TV) of a blind with respect to tilt angle depends on factors such as the distance from the window to the point where the illuminance is measured, the size and height of the venetian blind, the room geometry, the reflectivity of the room surfaces, the proximity and reflectivity of exterior objects such as buildings and foliage, the window azimuth orientation, and varying weather and sky conditions. This leads to a distribution of TV versus tilt over time, and over the installed base of blinds.
The following chart plots the TV versus slat tilt angle measured in two windowed areas in the Washington DC metropolitan area (one with a western exposure and one with a northern exposure), with each curve representing a different combination of office, time of day, and sky condition, and distance to window (ranging from 0.9 to 1.8 window heights). Also plotted is a guesstimate of a median TV versus tilt curve for the entire installed base of horizontal blinds in U.S. office buildings:
Note that at the median slat tilt of +40 degrees suggested by the data of Bulow-Hube, the observed TV ranges from about 5% to 50%, with a guesstimated median of 20%.
This implies an over-shading of at least 50%—and typically 80%—due to slat tilt when there is no risk of glare.
Typical Window-Covering Settings in Residential Buildings
Perhaps the most comprehensive source of information on window covering usage in U.S. residential buildings is a 2013 report prepared by D&R International for the Department of Energy. The report provides window covering usage data collected from 2,100 U.S. households via an online survey conducted in 2012. In addition to the impressively large sample size, the survey also addressed factors such as type of window covering, window orientation, type of room (e.g. living room, bedroom, etc.), and climate zone (northern, mid-tier, and southern).
Unlike the office building research cited above, this survey didn’t attempt to resolve venetian blind settings into separate occlusion and slat tilt settings, but it did cover other window coverings in addition to horizontal blinds (specifically vertical blinds, shades, curtains/draperies, and interior shutters).
The study found that—as with office buildings—a substantial fraction of window coverings in residential buildings remain mostly closed.
Distribution of Window Covering Settings by Season
The following two charts show, for summer and winter respectively, the percentage of all the window coverings in various positions per the report:
Note that almost half of all the window coverings remain fully closed, regardless of season or day of the week, and that there is surprisingly little variation between summer and winter.
The study did find some variation in degree of closure with climate zone and window orientation, but much less than might be expected; the numbers look pretty much the same as in the two charts above.
Distribution of Window Covering Settings by Type
More interestingly, the study found a significant variation in degree of closure between the various types of window coverings. That’s shown in the following four charts for summer weekdays, summer weekends, winter weekdays, and winter weekends, respectively:
Interestingly, more horizontal blinds were closed than were any of the other window coverings (with the sole exception of interior shutters on winter afternoons). Further, the survey considered blinds with a 45-degree slat tilt to be “half open”, whereas a 45-degree tilt results in a transmittance of much less than 50% (per Figure 7 above)—so the typical amount of shading provided by the blinds in the study was even greater than suggested by the data.
Implications
Unfortunately, the cited studies don’t directly address residential users’ motivations for keeping window coverings closed, but they strongly suggest that the motivation in office buildings is to block daylight glare.
In any case, the studies show that the average interior level of glare-free daylight in both office and residential buildings is much lower than it would be if window coverings were always optimally adjusted. This highlights the need for automated shading, and specifically for a responsive daylight control capability that always maintains the desired amount of glare-free daylight under changing conditions.
References
Bickel, Stephen; Phan-Gruber; Emily; and Christie, Shannon. “Residential Window Coverings: A Detailed View of the Installed Base and User Behavior”. Report prepared by D&R International, Ltd., Silver Spring, MD, for the Building Technologies Office, Office of Energy Efficiency and Renewable Energy, U.S. Department of Energy. 2013. https://basc.pnnl.gov/library/residential-windows-and-window-coverings-detailed-view-installed-base-and-user-behavior
Bulow-Hube, Helena. “Office Worker Preferences of Exterior Shading Devices: a Pilot Study.” Article IV in Doctoral Dissertation, “Energy-Efficient Window Systems. Effects on Energy Use and Daylight in Buildings.” Department of Construction and Architecture, Lund University, Division of Energy and Building Design, Lund, Sweeden. 2001. Report No TABK-01/1022.
Foster, Michelle and Oreszczyn, Tadj. “Occupant control of passive systems: the use of Venetian blinds.” Building and Environment 36 (2001) 149-155. Accepted 10 October 1999.
Haldi, Frederic and Robinson, Darren. “A Comprehensive Stochastic Model of Blind Usage: Theory and Validation.” Proceedings of the Eleventh Annual International Building Performance Simulation Association (IBPSA) Conference, Glasgow, Scotland, 2009. https://publications.ibpsa.org/proceedings/bs/2009/papers/bs2009_0529_536.pdf
Inkarojrit, Vorapat. “Balancing Comfort: Occupants’ Control of Window Blinds in Private Offices.” Diss. University of California, Berkeley, 2005.
Inoue, T, T. Kawase, T. Ibamoto, S. Takakusa, Y. Matsuo. “The Development of an Optimal Control System for Window Shading Devices Based on Investigations in Office Buildings.” ASHRAE Transactions 94[2], 1988, 1034-1049.
Lambeva, Lyudmila and Mahdavi, Lyudmila. “User Control of Indoor-Environmental Conditions in Buildings: an Empirical Case Study.” Proceedings Building Simulation 2007.
Rea M. S. “Window Blind Occlusion: a Pilot Study.” Building and Environment 19[2], 1984, 133-137.
Rubin et al. “Window Blinds as Potential Energy Saver—a Case Study.” NSB Science Series 112, 1978, National Bureau of Standards, Washington DC.
Sutter, Yannick, Dominique Dumortier, Marc Fontoynont. “The Use of Shading Systems in VDU Task Offices: A Pilot Study”. Energy and Buildings, 38[7], 2006, 780-789.