What Makes a Good Sunrise?

Set the clouds, the clock and the month. On top is the view from the shore under that sky; under it, a slice through the air from the lake out toward the sunrise that shows why; and under that, how my graded mornings turned out when the sky looked like that.

The light that colors a sky before sunrise comes from below the horizon. A deck of cloud overhead sees the sun before you do, and the light that reaches it has skimmed the ground a long way out: about 357 km for cirrus 10 km up, 196 km for a deck 3 km up. Cloud in that stretch, low over eastern Georgia or the Carolinas, takes the light before it arrives, and the deck over the lake stays gray. With 30% or more low and mid cloud 100 to 300 km toward the sunrise, 13 of my 21 mornings were flat.

So two clouds decide it, and they decide different things. Ranked by the cloud toward the sunrise, a morning with color came out ahead of a flat one 7 times in 10. Ranked by the high or mid cloud over the lake, a great morning came out ahead of the rest 8 times in 10. The low cloud over the lake, the one you can see from the shore, managed 6 in 10.

With the path clear and a deck covering 90% or more of the sky overhead, 15 of 19 mornings were great; with the path clear and nothing overhead, 2 of 32. A clear sky still gives color low along the horizon, which is why most of those had some. A deck 9 km up lights 15 to 17 minutes before sunrise here, depending on the month (the sun climbs fastest around the equinoxes), and low cloud lights last, 6 to 7 minutes before.

Every graded morning sorted by the weather archive's clouds ten minutes before sunrise, the first test that fits deciding. Up to 6 of each sky's photographs are shown, shared out by grade so they look like the counts.

• The grade is mine, of one kept photograph a morning, made from contact sheets that showed only the photographs and their dates.
• The clouds are the weather archive's (Open-Meteo): a model's hourly cloud cover on a grid about 25 km across, read between the hours. They are what the archive says happened, not a forecast, and a forecast made the night before can only do worse.
• 116 of the 117 mornings are at Stone Mountain, and one is elsewhere, read at its own sunrise spot. Another lake, with hills to the east, could sort differently.
• The view from the shore is drawn, not photographed. The far shore is the real treeline, measured by the sun sweep from the boatyard shore with its heights stretched; the sky is painted with the five colors of one of my mornings under that sky, the one named on the picture; the clouds are soft shapes, because the archive says how much cloud there was, not what kind. The glow sits in the middle of the view rather than at the sunrise bearing.
• The drawing keeps it simple: straight rays (the air bends a ray that skims the ground, which lights a deck a few minutes earlier than drawn), decks at fixed heights, and colors that show the order the light goes through rather than a measurement.
• A sky with fewer than 5 mornings would say nothing here; today every sky has 13 or more.

The color index and the rest of the log's numbers are explained on How the Log Works, and the clouds against the color index month by month on the findings page.