Pete's Stuff

We, the Navigators - Part I

The day before I left Hawaii, Annie and I were browsing the shelves at Barnes & Noble before she headed to work next door. I was perusing the local interest shelves and I bought this book for the plane ride home:

Its author, David Lewis, has had years of experience sailing the world's oceans. In the late 1960s and early 1970s he undertook a study of the fast vanishing art of indigenous navigation across the open expanses of Micronesia, Melanesia, and Polynesia. He sought out native navigators across the Pacific and learned many of their techniques in the most effective way possible - by actually voyaging with them for many days at a time and essentially becoming apprenticed to them. In 1976, as part of the bicentennial celebrations in Hawaii, he was one of the crew members who sailed the 65-foot voyaging canoe Hōkūle‘a (which is the Hawai‘ian name for Arcturus) from Hawaii to Tahiti using only traditional techniques, no instruments, and no Western knowledge.

In the book, he details the essential techniques he learned from these very competent navigators, and tries to convey the completely different worldview that informed their practice. In particular, none of the navigators he worked with were ever really able to understand nautical charts, something we would consider absolutely basic to the task. Nevertheless, it was pretty amazing to find so many correspondences between the navigational techniques of the islanders and the techniques many of us use every time we run an O-course, or do an adventure race.

More detail (much more!) below the fold.

In order to highlight these similarities, I'm actually going to introduce topics in a different order than the book does - and split it up into multiple posts as well, since it's turns out to be a large topic even for a brief survey. For Pete's sake, people have written whole books about it!

A discussion of accuracy

At first glance, the problem of sailing from one tiny point island to another in the vast expanses of the Pacific almost seems intractable. But, a closer look shows that things aren't that bad. First off, even though there are many isolated atolls and islands, many do tend to occur in clusters and chains rather than being uniformly scattered. Second, a good rule of thumb is that, on average, a low island is visible on the horizon at a distance of 10 miles. Therefore, if the sailor is constantly looking for land, on the average every island might as well be 20 miles in diameter. Of course, this rough rule of thumb can vary; weather and visibility conditions can reduce it and size, or the height of an island, if it's mountainous instead of a low atoll, can increase it.

Pacific navigators developed a series of techniques to further increase this "target zone", or the effective size of the island; the three most prevalent signs used were bird life, cloud formations, and a careful observation of how ocean swells were refracted around distant landmasses. They also learned to use underwater features; particularly near land, reefs and banks were visible when closely approached and often acted as catching features.

The end result of these techniques is that navigators could assume it's possible to detect an island to within a 20-30 mile radius. If you look at a map of the Pacific islands and circumscribe each island with a 25-mile zone, the various clusters and chains coalesce into solid masses of ocean area from which land is detectable. Many of the "gaps" between islands are filled in so that there's only a small chance of sailing directly between two islands in a chain without detecting either one of them. The worst-case scenario is a perpendicular path through a linear chain of islands, as shown below, but any scattering of additional islands outside the chain, or an oblique approach, tends to make landfall of some kind more certain:

Now instead of hopelessly minute target areas, we have much larger areas that are a minimum of 40 miles in diameter, and sometimes much larger. This provides a way to estimate the azimuthal accuracy required to reach a distant target, something already we're familiar with from following compass bearings! The approximate formula is 60 degrees x target size / target distance, so a 500-mile voyage to an isolated island (effective diameter 40 miles) requires a bearing accuracy of 4.8 degrees. That's actually not too bad! Also, because the target area can now be as large as 15-20 degrees, this provides navigators the option of aiming off to help account for the uncertainties inherent in winds and currents.

Furthermore, navigators have another clue - the time spent sailing. Obviously the distance-time relationship will likewise depend a lot on the winds and currents, but an experienced sailor on a familiar craft, in familiar waters, generally has no trouble estimating distance traveled to within say, 25%. This is the equivalent of our pace counting, which also happens to be accurate to within about the same margin (once calibrated). And like orienteers, the Pacific navigators recognized the lesser accuracy of this technique, and tended to use distance judgement as more of a safety technique to avoid overrunning their target if they happened to miss land signs.

Staying on bearing

So the most important technique was staying on bearing. All across the Pacific, the same fundamental technique was used, but David Lewis learned it best from two of his tutors: Hipour, from the Caroline Islands and Tevake, from the Solomon (formerly Santa Cruz) Islands. Lewis termed it the "sidereal compass." (Warning: the following discussion assumes you have some familiarity with celestial motions.)

At its simplest, the sidereal compass involves using stars near the horizon as the mechanism to maintain a bearing. At a given latitude on the earth's surface, every non-circumpolar star in the sky will rise above the horizon at a fixed bearing every day, and likewise set below the horizon at a different, but constant bearing. The seasonal shifting of constellations mean that different stars are visible in different seasons, but they all follow this rule.

As a star rises higher in the sky, its bearing will, in general, change. The amount of change depends on both the rise/set bearing and the latitude; bearings closer to due east and west result in smaller changes, and increasing distance from the equator results in larger changes. As a result, a star can no longer function as an accurate bearing indicator after it has risen a certain distance above the horizon, although that distance is a function of the aforementioned factors. So, after a certain time, a navigator must skip from their current steering star to a newly risen star that is chosen to have the same rise bearing - and continue this process all night long. (In modern terms, this means choosing another star with the same declination.)

At least, this is the principle. In practice, many other effects intervene, perhaps the largest of which is cloud cover. If a steering star is hidden, it's still possible to use orientation relative to other, visible stars to maintain course. In fact, Lewis describes situations like this, where Hipour maintains a bearing by using a star at a fixed angle from the boat's course (for example, hidden behind one of the masts.) However, each time the steering star became visible again, he would return to using it as the primary indicator.

Each school of navigation used a generally fixed set of directional points corresponding to the rise and set points of distinctive stars. For example, in the Carolines, the rise and set bearing of the bright star Altair (α Aquilae) defined two points of the compass. In European terms, these would be approximately 81°30' and 278°30'. Other points were defined by other stars and by certain constellations. Different points were available at different times of year, depending on the stars visible in that season. In general, there were enough directional points available, and enough leeway in the methods for aligning to them, to permit keeping on a course to within 5 degrees of accuracy at all times. Of course, some care was needed to account for the angled tracks of the stars through the sky (which varied with latitude,) and to overcome weather conditions, but this was part of the reason why the teaching of navigation was traditionally a years-long process.

Next: keeping course during the day; other bearing techniques, including swells; distance estimation; and the concept called etak.