Darkness for sleep has been so taken for granted that hardly anybody has investigated what happens when it is absent (even if inhuman regimes employ light at night as an efficient method for torturing their prisoners, they don't publish any research results...).
Various methods of monitoring sleep disturbances by insufficient darkness could be proposed. During our very short grant we used a method of oscillometric measurement of blood pressure by AD monitors. From 6 to 22, there were 3 measurements per hour; at night, just one per hour.
Nine people were investigated, four men and five women, average age 37 (SD 15 years). They were monitored for 48 hours of their normal life. The experimental treatment consisted in letting some light into the bedroom for the second night (illuminances of faces ranged from units to tens of lux) .
Within a period from midnight to 3 a.m., an unusual rise of systolic pressure was observed, on average by 10.95 mmHg (by 1.5 mmHg for a diastolic one). This rise was significant in both systolic and diastolic pressures at the 0.05 level.
Considering that 10 mmHg rise of systolic pressure corresponds to doubling the mortality on brain vascular accidents, the observed rise of 11 mmHg is not at all negligible.
The first sample had three hundred respondents, the other sample included one thousand people and was a representative country-wide one. In the first research we gained initial experience and got an overview, in the second one the questions targeted at several most relevant topics. The results in short: 5 % of Czech population perceives unwanted artificial light from outdoors as one of the two main causes of their sleep problems, further 7 % complains about light amounts which are not attenuated to tolerable levels, a further 20 % attenuates light pollution of their bedrooms to dark levels perceived as sufficient. However, 5 % of Czech population is unhappy to miss the full natural morning light due to barriers against the light pollution they have to use. A question remains: what are the health consequenses of reduced morning light for the remaining 20 % of population which uses barriers against light pollution?
Another result of the research is no dependence on the size of settlement of light-pollution imissions to bedrooms (so, the problem does not concern just cities).
Standard light-meters (luxmeters) are meant for rather strong light (several lux and more, seldom down to a tenth of a lux). However, much lower light levels are perceived as disturbing, by people who want to sleep (for several days around full moon, for one third of population there is not enough dark in bedrooms on clear-sky nights). As a safe limit of illuminance of the face, one millilux can be prelimimarily taken, on the basis of research done in 2003 in the Czech Republic.
A simple way of measuring illumination down to millilux level is using a solar cell and a common multimeter. The only necessity is to calibrate the system for several types of spectral composition of light (twilight, moon, high-pressure sodium, mercury, incandescent bulbs...). These calibrations and typical outdoor and indoor values for Vienna and Brno will be presented.
For studies of the influence of light at night, standard luxmeters are not the best tool. Apart from their inability to measure faint light, their spectral sensitivity matches photopic, daylight perception. Scotopic spectral response differs, and melanoptic differs a lot.
Fortunately, the blue sensors of modern digital cameras match the Brainard's melanoptic curve quite well and an even finer match can be obtained by including the other sensors as well. Taking a single image and processing it with our software (available from working directory http://amper.ped.muni.cz/light/luminance), real ``circadian'' radiance or irradiance levels can be obtained. The only need is to calibrate the camera, taking images of the solar spectrum. Melanoptic luxes and lumens will be defined and typical values for various light sources shown.
(J. Hollan thanks to Bob Mizon for correcting the English of the abstracts.)