The compass is the backup. Look up instead.

Nautical Notes from Aaron

The most common navigation question we get is some version of "how do I use a compass?" It's a fair question with a slightly dull answer, which is that on most of the paddling you'll do in Ontario, the compass is not the thing keeping you on line. It's what you fall back on when you can't see your target: distance, fog, or dark.

The most used options are all piloting, navigating by sight. Here's a quick primer, and one that most paddlers underrate.

A few terms first, because we’ll use them below. Set is the direction the water is moving you off your line, and here it is almost always wind driven. Drift is how fast it is moving you. Half a knot sounds like nothing until you are two hours into a crossing and a nautical mile off your line.

Hand-railing

Follow a linear feature to hold direction: a shoreline, an island chain.

My favourite example is the Fox Islands, off the coast of Philip Edward Island near Killarney. It looks menacing on the chart, but if you pick a line and hand-rail the islands, the whole thing becomes much more palatable. The chain is the rail and you run along it.

In class we teach it as four decisions you make before you launch.

Pick the rail. It has to be continuous and it has to run roughly where you are going. A shore, an island chain, maybe a set of navigation buoys in a channel.

Pick your distance off. This is the one people skip. Too close and you could be in reflected water, and you cannot see past the next point. Too far and the detail you were following stops being distinguishable from everything around it. Cap Chaillon in Lake Superior Provincial Park is a good example of getting this wrong in both directions. My own preference is close enough that I can still identify the plants growing out of the rock, which is how my iNaturalist account gets updated.

Cap Chaillon, Lake Superior Provincial Park.

Hold the aspect. You are not holding a bearing, you are holding the angle the land presents to you. The shore should slide past at a steady rate and a steady angle. When that starts to open or close, you are being set in or out, and you correct it then rather than later.

Know what ends it. Pick a checkpoint and a backstop before you start. That matters most in a place like The Thirty Thousand Islands, where the rail is an island chain rather than a mainland shore and it can hand you off to a different chain than the one you meant to follow.

And take a bearing anyway. It takes five minutes on the chart and it is the thing that catches you when the rail turns out not to be the rail you thought. Consider it another backstop.

Checkpointing

Pick landmarks that confirm progress. Something with a shape that can't be confused with its neighbours: a point, a bay mouth, an island with a feature you would recognise anywhere. In Myles Bay we always use One Tree Island.

One Tree Island, Myles Bay. Hard to mistake for anything else.

You want to break the journey into more manageable chunks. Navigating out to Flowerpot Island, you might set up North Point, Middle Island, then Beachy Cove before the flowerpots themselves. It reduces error by breaking the trip into smaller, more digestible segments.

The arithmetic matters as much as the position. Off your expected time means something is pushing you, and now you know before it matters.

Counting bays is not checkpointing

What you don't want to do is count. The fourth bay, the third river mouth. It's enticing, because it seems logical but it is the quickest way to get lost.

The reason is the same rule stated backwards: if you can't identify a feature on its own merits, it isn't a checkpoint. On wilderness coast there is often no way to tell one island from the next. In Georgian Bay there are just too many islands to keep track of. A bay with a rock spur across its mouth reads as two separate places. Two shallow bays behind a low point read as one.

Pukaskwa is the example I use. The rivers there are big, for the most part, but some empty out at an angle to the coast. Whether you see one depends on how far off you're running and the line you're on, close in and parallel, an angled mouth shows you nothing but the rocks on its outer shore, and you paddle past without registering it was there. Miss one and every count downstream of it is off by one, which is worse than being lost, because each feature after it now confirms a position you don't hold.

Which is the other reason to pick a distance off and hold it. It doesn't just keep you out of the rebound, it decides what you can see.

Backstops

The feature that tells you you've gone too far. A big headland, a bay mouth, a light, or just a time limit.

The time limit is the one worth practising, because it works when nothing else will. Crossing from Kneller to Big McCoy there is no obliging headland past the island to tell you you've overshot, so the backstop is the clock.

Which raises the question I ask on the water: do you set it short, accurate, or long? None of the three is wrong as long as you adjust for the one you picked, but the answer says something about how you paddle. Short means you are prepared for what is coming. Accurate means you are analytical and you will back your numbers. Long means you are prepared to deal with a miss. Pick knowingly, and tell the group which one you picked.

What you don't do is turn one-eighty and retrace. If you've blown your time, something set you, which means the track you'd be retracing isn't the one you paddled, and your launch point is a point target you can miss a second time. Turn ninety degrees and run in to the mainland instead. A coast is a line, and you can't miss a line. Once you're on it, hand-rail home.

Pick it before you leave the beach. A backstop chosen from the cockpit at the moment you need it isn't a backstop, it's a guess. The ninety is the same, you choose its direction on the chart before you launch, based on the landmass you're most likely to hit. Not on the fly, when you're already off and the reason you're turning is that you don't know where you are.

Transits, which are the best of the lot and almost always discounted

The formal version of this is a range, and the Great Lakes are full of them. Two fixed marks, one near and low, one far and high, positioned so that a single line joining them is the safe water. Stack them and you are on that line.

They exist here because the lakes are connected by channels that are narrow lines through shoal and rock, and because the alternative was losing ships. The Munising range was built in 1905 to run vessels down the East Channel past Grand Island, where rock ledges could sink a boat and the single light that preceded it wasn't good enough. At the St. Clair Flats the river emptied into the lake through seven principal mouths, and the range existed to tell you which one you were in. At Vidal Shoals the marks were repositioned until they held the centreline of the dredged channel. A run of identical range stations went up around the western lakes between 1865 and 1877: Copper Harbor, Eagle Harbor, Baileys Harbor, Presque Isle.

Nobody builds a stone tower to mark an approximate line. That's the thing worth taking from them: a range isn't a rough guide, it's a hard edge, and you get the same precision for free off any two fixed objects you can find.

You don't need built ones. Any two fixed objects work: a rock and the headland behind it, a point of land and a bluff beyond. Hold them stacked and you're on that line. Let them open up and you're being pushed off it, and the way they open tells you which way you've gone.

Why this beats everything else in the list: it updates continuously, it needs no chart, no bearing and no looking down, it's more accurate than any handheld compass, and it works while your hands are busy and the boat is moving around. A transit will show you a drift of a few hundred metres an hour inside about thirty seconds, which is faster than you'd notice it any other way.

They exist all over the Great Lakes. The best example near us is the range marking Devil's Island, with the marks set on the Otter Islands off the Tobermory coast.

The catch is that a transit needs two things to line up, and it needs you to be able to see them. When you can't see your target, too far off, fog, or dark, you have no marks to work with, and you're on deduced reckoning: a bearing, a known speed, and a clock. That's when it's time for the compass.

Aiming off

If you paddle a bearing straight at a landing and are set off, you may not know which way to turn, especially in the dark or fog. Left and right are both plausible, and you're now searching an unfamiliar shore.

So don't aim at it. Aim a few degrees to one pre decided side, consistently, and know which side. When you reach the shore you turn one way with certainty.

The math is interesting: roughly 100 metres of error per degree per 6 km.

Russel Island across to Turning Island is about a nautical mile, and on most days you take the prevailing southwesterly on the beam for the whole of it. Over that distance a degree is worth roughly thirty metres, so a few degrees off line is the difference between arriving at the island and arriving in the open lake.

So aim off, and make the offset larger than the drift can plausibly carry you. A southwesterly sets you northeast, so aim off to the northeast by more than it can carry you. Now you know you are northeast of the island when you get there, and you turn southwest to close it without thinking about it. An aim-off smaller than the drift can carry you tells you nothing at all, because you could end up on either side.

Great Lakes set is wind, not tide

If you read up on nav, most of the good navigation reading is about tidal streams. We don't have those.

What we have is wind-driven surface current, seiche, and outflow through narrows and river mouths. Generally weaker than a tidal stream, and far less predictable, because there's no table to check the night before. Nobody publishes a stream atlas for Georgian Bay.

Which lands where we started: you can't compute a drift in the Great Lakes as easily, and you can see it on a transit in thirty seconds.

For the other half of the pre-launch call, which forecast model to trust, and which one lies to you on a single bay, that's in the weather notes.