This page is a synopsis of my technical paper "Predicting Moonlight Brightness for Night Landscape Photography". I hope to have the paper ready for publication soon, and then find a journal that will accept it. The subject of the paper is a cross-over from astrophysics to photography, and does not fit the precise scope of many of the usual technical journals. Consequently, finding a journal that will accept it is being an issue. However, having spent thirty years on the pursuit of an excellent moonlight photography technique, I want to establish priority for my work before I give up all the details.
I took my first moonlight photograph in 1972, using my father's old Kodak Senior 620 camera and Verichrome Pan film. The results were almost invisible. It was pretty clear that I had no idea what I was doing. But I did know that the feeling of being out under the stars was one that I needed to capture somehow, and I had to find a way to go about doing it. Eventually (after several years of playing around) I figured out that bracketing the exposures could get me close to the desired result, especially if I employed a sequence starting at 15 seconds and then doubling the exposure time out to several minutes.
It wasn't long before I found out that the same exposure made on two different occasions did not give the same result. It took me about five years of stumbling around at the edge of that problem before I realized that the brightness of moonlight is extremely variable over a range of many stops, and that it was nearly impossible to figure out what the brightness of a given scene was going to be ahead of time. Well, it is now nearly thirty years later, and I am pleased to say that I've overcome that problem. The brightness of moonlight under clear skies is as easily predictable as is the brightness of sunlight under similar conditions. I have devised a technique that gives the exposure to use under such conditions. (I realize this does not work when the moon goes behind a cloud, however I am working on that problem too, and am close to a solution that can be applied whenever you can still see the moon's disk through the cloud. More on that some other day!)
There are several things that cause moonlight brightness to vary. The most obvious is the moon's phase. The brightness of moonlight varies by approximately a factor of 10 between quarter phase and full moon, based on a diffuse reflection and the geometry of the positions of the earth, sun and moon alone. This is about three and a half stops of light, which is substantial. Another factor is the distance between the earth, moon, and sun, which changes due to the earth and lunar orbits not being perfect circles. The distance from the earth to the sun varies from 0.9833 Astronomical Units at perihelion to 1.0167 Astronomical Units at aphelion. The distance from the earth to the moon varies from 356400 kilometers at extreme perigee to 407000 kilometers at extreme apogee. The amount of light that falls on a body varies with the inverse square of the distance from the light source, so the combined effect of these distance variations can be quite pronounced. The range of variation in lunar illumination is 6.9% for variation in sun distance, and 30% for variation in moon distance. This amounts to about one-third stop of brightness, which enough to change the mood of a photo when slide films are used.
A third factor in moonlight brightness is the so-called opposition effect. The surface of the moon is covered with small glassy particles that can serve as wonderful retro-reflectors. If you are within a small angle to the line between the sun and moon, the amount of light coming from the moon increases dramatically relative to that you would expect from a diffuse reflection alone. There is quite a range in the magnitude of this effect presented in scientific literature, ranging from a factor of 1.35 to a factor of 20! Whatever value you choose to use, the effect is at least one-third stop of light, which makes it significant if you are using slide film.
The final parameter that introduces variation into moonlight brightness is atmospheric attenuation, or atmospheric extinction, to use astronomer's jargon. This accounts for the amount of light that is absorbed or scattered when light from the moon passes through the earth's atmosphere. There are two factors involved: the amount of reduction per a given amount of air transited by the light (the "extinction coefficient"), and the amount of air in the light path ("air mass"). There are three principal phenomena that contribute to the extinction coefficient: molecular absorption, molecular (Rayleigh) scattering, and scattering by aerosols (particles larger than molecules). See my web page on atmospheric attenuation for more on these individual factors. Overall, an extinction coefficient value is small for dry, clear air, but can be huge for moist, dusty air. And the amount of air the light passes through can vary from one "air mass" with the moon directly above you to forty "air masses" when the moon is on the horizon. The combined effect can give an enormous variation in moonlight brightness. In clear air, the difference between having the moon directly above you and being near the horizon can be six to eight stops! So if you are looking for that wonderful grazing light that gives modeling and depth to a landscape, you really need to know what is going on here to ensure success.
The figure below shows the results of a moonlight brightness model that takes the above factors into account, but is limited to relatively clear air at an elevation of about 3000 feet above sea level. The brightness is in terms of lunar illumination in foot-candles (multiply by 10.76 if you want to use the metric units, which are lux, or meter-candelas). The parameter Z is the "zenith distance", or the angle between the moon's position in the sky and the position "at the zenith", or directly above you. The Lunar Phase Angle is a precise measure of the phase of the moon, and is the angle between the positions of the sun and earth in the sky, as seen from the moon.
Note that this figure shows only the variation due to phase and the position of the moon in the sky, and includes a correction for the opposition effect. You have to add the correction for sun-moon and moon-earth distances. The effect of the clarity of the atmosphere will have a big effect on these results.
So how bright can moonlight be? The brightest moonlight occurs with the moon at perigee and the earth at perihelion, right at full moon phase. You can never have the moon at its theoretically fullest phase, right opposite the earth from the sun, because whenever the moon goes there it enters the earth's shadow and we get a lunar eclipse! But assuming the eclipse didn't happen, we could assume the following: a phase angle of zero, very clear air with an "extinction coefficient" of 0.11 magnitudes per air mass, the moon on the zenith so that the moonlight passes through a single "air mass", the brightness would be 0.0462 foot candles (LV -2.0), neglecting "the opposition effect". If we include "the opposition effect", the brightness could be anywhere from 35% to 20 times brighter (note that the 35% is the more accepted value in the scientific literature, which would give an LV of -1.7). I once measured moonlight brightness with a Gossen Luna-Pro incident light meter in mid-winter on the Kelso Dunes at LV=-2.2, so this is a believable result.
There is a rule of thumb, sometimes called the Looney 16 Rule, which says we should treat the moon as being 250,000 times dimmer than the sun. This would have us use a shutter speed 18 stops slower than the 1/(film speed) value that is used in the "Sunny 16 Rule", which works out to about 44 minutes at f/16 under moonlight conditions with film having an ISO speed rating of 100 (uncorrected for reciprocity failure!). This is close enough to be useful under full moonlight conditions given an average earth-moon and earth-sun distance, with the moon high in the sky, and clear air. Actually, since the sun has an astronomical visual magnitude of -26.74, and the full moon an astronomical visual magnitude of -12.73, the sun is more like 402,000 times brighter than the moon, or 18.6 stops. Thus using this rule pretty much ensures a minimum of 2/3 stop underexposure - which may actually be fine, since most of the time you want an underexposure to give the impression of night in the image. However, a more precise Looney 16 Rule would give us a basic full moon exposure of 66 minutes at f/16 using ISO 100 film, though you still have to correct for reciprocity failure (do that AFTER you adjust your f-stop!) Such a rule would give a daylight-like image, and you would have to adjust the exposure to get the desired effect.
There is also something called the Looney f/4 Rule that is based on the sun being 1.4 million times brighter than the moon (about 3 times more than is actually the case). The rule says to use an exposure of (1 day)/(ISO film speed) at f/4 (1 day is 86400 seconds). Thus with ISO 100 film, you would expose for 1/100 of a day, or about a quarter hour (14 1/2 minutes). This rule seems to give a reasonable starting point under very-full moonlight conditions (LV=-2.3), assuming you are using a film who's reciprocity characteristics follow the OLD Kodak B&W film guide, and you don't add any further correction for reciprocity failure. If used with modern T-grain emulsions, you will get a severe overexposure with this method, since T-grain films show much better reciprocity response with long exposures than do older film emulsions. I suggest revising this rule to use f/11 with modern T-grain emulsions (such as Fuji Provia 100F), to take into account the better reciprocity response. Or use f/13.3 (halfway between f/11 and f/16) and then apply your own reciprocity failure correction after adjusting the f/stop. This would give a daylight-like exposure under very bright full-moon conditions. Note that this is far less time than is required by the Looney 16 Rule, since it is based on a brighter moon as well as a smaller f-stop (larger lens aperture). The two rules are not exactly equivalent.
Another method is the Rule of Three 4's that says, for a very full moon (so that you can see the full circular disk) and clear sky, use 400 speed film with an exposure of 4 minutes at f/4. I have seen numerous references to it around the web and elsewhere. This works well for older emulsions using the old Kodak reciprocity failure correction table for black and white film, with a moonlit scene brightness of around LV -3, or fairly typical full moon conditions. However, as with the Looney f/4 rule, be careful if using current T-grain emulsions, since the Rule of Three 4's will give you an overexposed image with such films. If you use a film with one of the newer emulsion types, I suggest you bracket exposure several stops toward smaller lens apertures until you understand how your film of choice responds.
But let's calculate what the lunar illumination would actually be under conditions you might expect with an "average" full moon. If we were to choose a "normal" situation, say one with a three degree phase angle, extinction coefficient value of 0.2 (clear air at 3000 feet elevation), and the moon fifteen degrees off of directly overhead, with "average" distances between the sun, moon, and earth, then the moonlight brightness would be 0.0269 foot candles (0.290 Lux, or a photographic LV=-2.8), or about two-thirds stop less than the "maximum possible" value. And a little earlier that night, with the moon at an angle of 60 degrees to the zenith, atmospheric extinction could make the light a full stop dimmer yet, and even more than that if the air is hazy!
See the table of LV values in the On Light and Camera Settings page for typical moonlight brightnesses for some phases other than full moon.
So at last we understand why it is that if you go out in "the days right around full moon" and apply a single exposure value to a given scene, your results will vary a lot over that period of time, from one full moon to the next, or even over a period of several hours in the same night. Thus it pays big dividends to understand the factors that contribute to the varying nature of moonlight, since clearly you have to know how much light you have to work with to achieve a pre-visualized result. It is, unfortunately, a very complex relationship, and precise calculation requires knowledge of celestial mechanics as well as optical physics. However, as my experience shows, mastery is possible, and the results can be well worth the effort.
In the end, unless you understand and can predict the nature of moonlight brightness, photographing by moonlight is very much like shooting at a moving target - if you don't lead the target by a sufficient amount, you won't hit it. You could use a shotgun, which is about what using the wide-bracketing exposure method does, however if you really want the results match a pre-conceived image in your mind's eye, you will have to do a lot of bracketing in small increments to get close.
I hate wasting that much film.
C. D. "Kit" Courter
Torrance, California, USA.
Allen, C. W. 1976, Astrophysical Quantities (London, Athlone)
Courter, C. D., 2003, Predicting Moonlight Brightness for Night Landscape Photography. This paper is being prepared for publication.
K. Krisciunas and Schaefer B.E., “A model of the brightness of moonlight”, Publ. Astron. Soc. Pacif. 103 (667), 1033-1039 (1991)
This web page and all contents (c)2003-2005 LunarLight Photography.
All Rights Reserved.
Contact LunarLight Photography