As we close out a not-too-severe winter and "officially" head into Spring with the passage of the Vernal Equinox, slated for 8:07 pm Tuesday March 20th, we've gotten a few questions from folks who noticed that by the time we reach the equinox date, sunrise and sunset times show that we have about 7 or 8 minutes more than 12 hours of direct sunlight. The question, of course, is "if it's the Equinox, why aren't there 12 hours of sunlight and 12 hours without?"
Such a day, having sunrise and sunset 12 hours apart from each other, already occurred early in the weekend, several days before the equinox. This day we can refer to as the "Equilux," with roughly equal times having direct sunlight available and not.
The reasons for the 3-4 day discrepancy between the equinox and the day(s) nearest to 12 hours of light are twofold, one of them an artifact of how we define sunrise and sunset, the other a physical manifestation of refraction of light rays by our atmosphere.
If we lived on a smooth earth, defined sunrise and sunset according to passage of the center of the sun across the horizon, and had no atmosphere at all (which contradicts the idea of "living" here, of course) then the equinox would be the date that comes closest to 12 hours between sunrise and sunset. Instead the equinox is defined astronomically as the point in our orbit (there are two each year) at which the rotational axis of the earth tilts neither toward nor away from the sun, also the point at which the sun will be directly overhead at solar noon along the equator, and at which for most locations on earth sunrise and sunset occur almost due east and west, respectively.
While all of this works out as expected, the times of sunrise and sunset don't entirely cooperate, because
(1) those times are defined by the passage of the upper edge of the sun's disk, not its center, across the horizon. Because it takes something over two minutes for the entire solar disk to cross the horizon at our latitude, this makes our sunrise over a minute earlier and sunset similarly later than they would be using the sun's center, and adds over two minutes to the length of our sunrise to sunset time on the date of the equinox. Also,
(2) there is an even larger effect due to bending of the sun's rays by the atmosphere. In general, sunlight passing through the atmosphere is refracted downward toward the denser air near the surface. This effect is strongest near the terminator dividing light from dark, where the rays are traveling more or less parallel to the earth's surface. Since the rays are curved downward, the sun's apparent position in the sky when near the horizon is displaced upward (see the illustration above, borrowed from http://www.bedford.k12.ny.us/flhs/science/geoscience/sunrise.html). This displacement is about one full solar diameter, meaning that we are able, in a sense, to "see around the horizon" as the sun is rising and setting, so that when the sun's upper edge is actually just about to climb to the horizon at sunrise, it appears to us to already be fully risen, with its lower edge just pulling away from the horizon. This effect adds another two minutes-plus at both sunrise and sunset to the length of the day on the equinox, accounting for the remainder of the 7-8 extra minutes of equinox "daytime" I mentioned above.
Beyond all this, it's worth noting that while as a society we "officially" mark the beginning of Spring with the passage of the Vernal Equinox, meteorological and climatological records define spring as the entire months of March, April and May, partly because the weather pattern transitions that we consider "springlike" are already underway on average for several weeks before the equinox, and probably in part because it is much more convenient from a record-keeping perspective.
Finally, the same offset between equinox and equilux will occur in the fall, but this time the 12-hour day occurs 3-4 days after the autumnal equinox.