🪂 Club XC plannerJoinSign in

Where do Millington's thermals actually start?

Every pilot has a theory. The thermal-trigger map lets the flights decide.

Tracing a climb back to the ground

A thermal leans with the wind. By the time you're circling in it at 400 metres, it may have left the ground a kilometre upwind. For every climb, the planner measures the drift from the circles themselves (or uses the wind from a Vector vario), estimates how long the air took to rise, and traces the climb back to the patch of ground it probably came from. Then it looks at that ground: which way it faces, how steep it is, whether the sun was on it, and whether it's a hollow, a spur or open field.

Many flights pin it down

One climb only tells you the source is somewhere back along its drift line. But lines from different pilots, days and winds cross where the real trigger is. The overlap builds up into numbered spots, which you can rank by how many climbs point there or how strong they were.

The 3D flythrough at Millington: two pilots on different days (March and September), launches lined up, with their climbs drawn back to a shared source (column 1). Right: the side-on profile along the wind, with each climb's bar and a dashed line down to its estimated trigger.
The 3D flythrough at Millington: two pilots on different days (March and September), launches lined up, with their climbs drawn back to a shared source (column 1). Right: the side-on profile along the wind, with each climb's bar and a dashed line down to its estimated trigger.

The 3D flythrough replays the flights over satellite imagery and the planner's terrain. Each climb is drawn as a "snake" that starts at its estimated trigger on the ground, leans with the wind and follows where the glider actually circled.

A first look at Millington

With only 11 climbs from 2 flights it's a rough first look, but a pattern is already hinting at itself. Compared with how common each kind of ground is in the area:

GroundShare of climbsShare of area
NW-facing slopes36%4%
W-facing slopes27%6%
Hollows and valleys (40 m+ below surroundings)27%3%
Spurs and hilltops (40 m+ above)18%4%
Flat ground36%69%
"What the data says": each kind of ground's share of climbs against its share of the area. Above 1.5× suggests that ground triggers more than chance.
"What the data says": each kind of ground's share of climbs against its share of the area. Above 1.5× suggests that ground triggers more than chance.
The same climbs in pictures: by hour, wind direction, month, height gained, strength and sun height. Almost all of them came in north-westerlies, so treat the pattern as "Millington in a NW", not Millington in general.
The same climbs in pictures: by hour, wind direction, month, height gained, strength and sun height. Almost all of them came in north-westerlies, so treat the pattern as "Millington in a NW", not Millington in general.

It also lists places worth checking: ground that looks like the trigger spots so far but has no climbs pointing at it yet.

Candidate spots with their position, slope, height relative to surroundings and sun on the ground.
Candidate spots with their position, slope, height relative to surroundings and sun on the ground.
This shows where we have found lift, not everywhere lift exists: pilots fly the same lines, so the sample is biased. The terrain is bare earth (trees add 10 to 25 m), and it needs around 30 climbs per site before the patterns mean much. More flights make it better, which is the best reason to add yours.
Want to try it? Message me and I'll give you a demo on one of your own flights. You choose who sees each flight you upload: the club, the club without your name, or only you.