Asking Polaris for a Second Opinion on North

Polaris is about 0.7 degrees off true north. That’s the number I keep turning over while I drive out to the airfield at 10 PM with a camera bag, a tripod, and a Thermos of coffee that’s already cooling.

The idea arrived sideways, the way most of my hobbies do. I’ve been flying out of this strip for years, swinging the compass on the painted rose like every other pilot, trusting the deviation card taped inside the panel. But last week I caught myself staring at a star-trail image and thinking: the sky already knows where north is. Why am I trusting paint and magnets when I could trust geometry?

Stacking the Evidence

The setup is almost embarrassingly simple. Tripod on a crack in the apron—same crack every time, because repeatability matters more than perfection. Camera pointed toward Polaris. Intervalometer set for 10-second exposures, 180 frames, about 30 minutes of Earth rotation caught in small pieces.

I learned stacking discipline from A Knight’s Walk Across a Programmable Night, where I tiled a knight’s tour across the sky and let the stitcher find the seams. Same principle here: many short frames beat one long exposure. You can reject the headlights, the satellites, the airplane that blinks through at frame 47. The stack forgives interruption. A single 30-minute bulb exposure forgives nothing.

The air is -14°C and still. My toes are gone by frame 60. I stand there anyway, watching the camera click, running 29,000-something chess games through my head to stay warm. The sky turns. The stars draw arcs I can’t see yet.

What the Script Actually Does

Back inside, the laptop warm on my knees, I stack the frames and watch the trails emerge. Polaris traces a tiny circle—not a point, but an arc, because it’s not quite at the pole. The script fits a circle to that arc, finds the centre, and that centre is true north.

pole_center = fit_arc_center(trail_points)
ref_line = runway_heading_deg  # known from survey
true_north_offset = angle_between(pole_center, ref_line)

The math is high-school geometry dressed up in numpy. The satisfaction is something else entirely.

Magnetic declination in central Alberta is about 14° east right now, drifting roughly 0.1° per year. Deviation—the aircraft-specific error caused by alternators and radios and the steel in the firewall—is a separate beast. The calibration card has to separate these two. Beginners confuse them constantly. I know this because I confused them constantly, approximately two hours ago, when I parked the tripod too close to a chain-link fence and spent twenty minutes blaming the math for what the fence did.

The Card

The output is a wallet card. True north bearing from my reference crack. Magnetic north after declination. Date stamped, because a calibration without a date is just a historical artefact.

I print it on cardstock, trim it, and slip it into my kneeboard next to the deviation card that came with the aircraft. Now I have two sources of truth: one from paint and magnets, one from 180 frames of starlight.

They agree within half a degree.

I don’t know why this pleases me so much. Maybe it’s the inversion—using the sky to check the ground equipment instead of the other way around. Maybe it’s the quiet fact that Polaris won’t always be the pole star. Axial precession means the pole traces a 26,000-year circle through the heavens. Around 2102, Polaris will be at its closest approach, about 0.45° off. In 5,000 years it’ll be something else. My little card is rooted in a slow drift that doesn’t care about my preflight checklist.

What Didn’t Work

The first stack was too short. Fifteen minutes of frames made the arc fit noisy, the centre unstable. I needed more rotation to see the geometry clearly.

I also forgot the sidereal day. The sky rotates in 23 hours, 56 minutes, 4 seconds—not 24 hours. For a single night it doesn’t matter much, but if I start comparing stacks across weeks, that four-minute difference will compound.

And the fence. The fence was a mistake I should have seen coming.

Packing Up

The airfield is quiet again. The wind sock hangs limp. I have a card in my pocket that says true north is 7.2° left of the crack in the apron, and I believe it because I watched the stars say so.

There’s a ritual in aviation called compass swinging—parking on the painted rose, comparing your compass to known headings, logging the deviation at each cardinal point. Pilots have been doing it for a century. What I’ve built tonight is a DIY compass rose made of starlight, and the only paint involved is the fading white line I’ve been using as a reference.

The coffee is cold. The card is warm from the printer. That seems like enough.