Daylight
The daily gain in daylight is not constant. Near the equinoxes it can exceed three minutes a day; near the solstices it approaches zero. Two facts that surprise most people follow from that.
Earth's rotational axis is tilted about 23.44 degrees from the perpendicular of its orbital plane, and that tilt stays fixed in orientation as the planet orbits. This is the entire cause of seasons and of everything below.
Day length across the year approximates a sine wave. The important property of a sine wave is that its rate of change is greatest at the midpoints and zero at the extremes.
Applied to daylight: at the equinoxes, day length changes fastest. At the solstices, it barely changes at all.
| Period | Daily change at 40°N |
|---|---|
| Around the March equinox | ~2.5–3 min gained per day |
| Mid-May | ~1 min per day |
| Around the June solstice | Under 10 seconds per day |
| Around the September equinox | ~2.5–3 min lost per day |
| Around the December solstice | Under 10 seconds per day |
At higher latitudes the amplitude is larger and the equinox rate steeper — at 55 degrees north, early spring can bring nearly four minutes of additional daylight per day, which accumulates to nearly half an hour per week.
This explains a common perception. Late February and March feel like the light is returning rapidly because it is. The week around the summer solstice feels static because day length genuinely changes by only a few seconds across it. The solstice is a turning point in name, and functionally a plateau lasting a fortnight either side.
This is the fact that reliably surprises. At mid-northern latitudes the earliest sunset falls in early to mid December, roughly a week or two before the solstice. The latest sunrise falls in early January, a week or two after. The shortest total day sits between them, on the solstice itself.
The cause is the equation of time — the difference between clock time and true solar time. Two factors produce it.
Earth's orbit is elliptical, so orbital velocity is higher near perihelion in early January. And the axial tilt means the sun's apparent motion projects unevenly onto the celestial equator across the year.
Together these shift solar noon relative to clock noon by up to about sixteen minutes over the course of a year. During December, solar noon is drifting later by roughly half a minute per day. That drift is larger than the change in day length near the solstice, so it dominates: both sunrise and sunset move later, which makes sunset time reach its minimum before the solstice and sunrise reach its maximum after.
The practical version: evenings start getting longer around 8 December, more than two weeks before the shortest day, while mornings continue getting darker until early January.
If your outdoor time happens after work, the useful light returns earlier than the calendar suggests. Evening daylight begins improving in the second week of December and by mid-January has gained a meaningful amount, even though the mornings are still deteriorating.
If your outdoor time happens before work, the opposite holds. Mornings continue getting darker through the turn of the year and do not begin improving until early January.
And in spring, plan for the light to arrive faster than intuition suggests. From the March equinox onward, a route that was marginal one week is comfortable two weeks later at mid-latitudes.
Amplitude scales strongly with latitude. At the equator, day length varies by only a few minutes across the entire year. At 40 degrees, the swing between solstices is about six hours. At 60 degrees it exceeds twelve. Above the Arctic and Antarctic circles at 66.5 degrees, the sun fails to set or rise at all for part of the year.
This is why seasonal light strategy is not transferable between locations. Advice built around a 50-degree winter is irrelevant at 30 degrees, and vice versa.
The daylight window calculator computes sunrise, sunset, solar noon and civil twilight from first principles for any date and coordinates, so you can check the actual curve at your own latitude rather than working from a general rule.