|This article was written for
SailNet by David Brayshaw, who develops and sells navigational software under the
corporate name Local Knowlege Marine Software.
The key element in any long-distance race or cruise is routing, which essentially means
picking the most favorable course from starting point to destination. This requires you to
estimate wind, and often current, along the way and weigh these effects to take best
advantage of your boat's sailing characteristics (or what performance sailors refer to as
the boat's polars). What makes this tricky, of course, is that the wind is changing not
only from point to point, but over time, so it makes a great deal of difference when you
actually arrive at any given spot. What looks like a fast route on the first day or two
may deliver you right into a "hole," with very little wind for days thereafter,
whereas a slower initial route, perhaps one where you sail more miles, may enable you to
avoid this and continue your progress without interruption. And strong currents such as
the Gulf Stream, which can flow up to five knots, will often dominate routing. Add to that
the fact that sometimes the wind may be so adverse that following an otherwise beneficial
current doesnt pay and you can see how complex this subject can become. But with a
little help you can learn to balance the competing factors to maximize favorable weather
and current over the entire route.
In the last few years, marine weather and wind forecasts have become much more
accurate, and the Internet now provides access to numerous public websites where you can
obtain high-quality, up-to-date information, much of it for free. Sailing instructions are
being relaxed to make this practice legal while racing, and the requisite onboard hardware
for obtaining this information has dropped sharply in price. Services such as my product
called "Local Knowledge" (available at www.goflow.com), can provide detailed
ocean current forecasts, derived by analyzing satellite data. So, the information is
available to plan your route. What is also available, and relatively new, is "routing
software," which can help you put it all together.
PC programs such as my own Force 4, or Deckman, Raytech, and MaxSea use information
regarding wind, current, and your boats polars to calculate the most effective route
for a given destination. Basically, the software does in seconds what would take an
ordinary mariner weeks, if not months to do by hand: it tries out every possible route,
adds up the time, and tells you which route is the quickest. The result is displayed
visually, on the same electronic chart that shows your position, and can also be broken
down into a series of waypoints, headings, etc. The truth is, software like this can put
your navigation on a par with that of topflight professionals.
The 2002 Newport to Bermuda Race is an excellent case-in-point. I sailed the race as
the navigator aboard Morning Glory, an 80-foot maxi sloop owned by Hasso Plattner,
and had a chance to compare the Force 4 predictions with actual conditions over some 635
miles. This race offered a full spectrum of conditions, starting in strong easterly winds
off Newport, RI, on June 14 and ending in even stronger southwesterly winds
approaching Bermuda a few days later. In between, an approaching low-pressure system from
the southwest signaled the transition from one wind regime to the other, with a dangerous
area of weak winds near its center. The first challenge in routing was to minimize contact
with the low. The second was to take advantage of perhaps the most unusual and dramatic
Gulf Stream pattern in the history of this race. Almost invariably, the Stream flows from
west to east, pretty much at right angles to the rhumb line from Newport to Bermuda.
However, this year the stream took a sharp right turn west of the rhumb line, and headed
south for almost 160 nm before making a hairpin left turn and coming back north to resume
its normal easterly flow. That formed a big U of rapidly flowing water, which reached five
knots at its peak. (See Figure 1.)
This Gulf Stream "meander" meant a huge opportunity for all the racers,
offering the prospect of a four to five-knot push in the right direction for more than 150
nm. The opportunity came at a cost, however, with the usual buildup of large, choppy waves
in the Stream when wind and current direction oppose each other. This situation is common
in races up from Miami, such as this springs Leg Six of the Volvo Ocean Race to
Baltimore, but unprecedented in the Bermuda Race. It also meant that the boats had to work
their way well west of the rhumb line to line up with it, bringing them closer to the
approaching low and its central windless hole. This presented a classic routing decision:
how to get to and use the Stream without paying too much of a price in light air.
Routing programs are designed to take wind and current information from digital files
(a file of numbers, rather than a screen image). There is a standard "grib"
format for weather files, and a grib file can contain a number of variables, from wind
predictions to pressure, current, and sea height, etc. While the router requires both wind
and current, there are no public sources of accurate current data available on the web,
and so current is obtained as a separate file and cannot be updated during a race. This is
the second Bermuda race for which I built a current-prediction model from data provided by
oceanographer Jennifer Clark. Her analysis is based on temperature/wave height differences
(altimetry), as measured by satellite photography, and years of experience. My PC model
was developed the day before the race and distributed to racers via e-mail. It cannot
capture changes in the ocean current during the race, but fortunately these changes are
small over a matter of a few days. In the 2002 race (and in 2000 race), the "Local
Knowledge" current model was described as "spot-on" and "astonishingly
accurate" by some top racers, and I too found it right on the money.
Wind prediction is a different matter because its much harder to forecast days in
advance, and different methods (or "models") yield somewhat different results.
Top boats often purchase custom forecasts from commercial sources, but I have been quite
impressed by the free sources on the web. For this race, I relied primarily on the
five-day forecast distributed on MaxSea's navcenter.com website. The grib file contains
prediction times three hours apart, with a spacing of about one degree, within which the
software interpolates values. This is relatively high resolution, even for commercial
sources, and I have had good experience with it over the past year. That said, it is
prudent to obtain several sources, if possible, and route with each to detect any
discrepancies that might affect the result.
Morning Glory started in the ninth group, at 14:30 EDT, and I computed a routing
solution shortly after the start, choosing as a destination point a mark at the northern
approach to Bermuda, where the shoals and reefs form a barrier that one must keep to
starboard, rounding to the east in the final miles to the finish. The routing solution
identifies both an optimal (quickest) route, and several alternative routes for
comparison. This feature is unique to Force 4, and can be very helpful to sort out just
how much faster the optimal route is, and whether other considerations might dictate some
adjustment to it. You can step along the alternative routes in synch, or specify a common
intermediate time, and see how predicted boat position will correspond to the wind and
current expected at that time and location.
The accompanying Force 4 image above (Figure 2) shows the initial computation, stepped
forward to midnight on June 14, with the approaching low very evident in the lower left
quadrant (higher winds are shown in bolder colors, to approximately 25 knots for the red
vectors in the extreme north). The red line shows the optimal route, with the black dot on
each colored route showing predicted boat position at that common time. This told us that
we could get past the low in adequate winds and still keep westward enough to hit the Gulf
Stream at the right spot. As it happened, the forecast was quite accurate and we did just
that. The trailing boats were not so fortunate, and some found themselves parked for four
to five hours. Figure 3 (below) shows the situation at the Saturday morning position
report (at 0800 EDT). By then the low had moved east of our prior track, presenting a
considerable obstacle to slower boats taking the same line.
Figure 3 also shows predicted times for the routing solutions (color-coded to match the
line) in hours and minutes, and distance from the black dot on each route to the
destination point in the routing (northern approach to Bermuda). Predicted positions on
the optimal red and (time-equivalent) magenta routes are slightly beyond, but very close
to the actual reported positions of leading boats in IMS Class 9. At the point in the race
when this image was frozen, Morning Glory (MG) and Boomerang (BO) were tied
for first in terms of distance to the destination, with Pyewacket (PY) some eight
nm further away. Boomerang, however, was closer to her optimal route, and was in a
slightly better position, the equivalent of 10 minutes or so ahead. All these boats,
however, had achieved their objective of getting past the low with good speed and in good
position to maximize use of the Gulf Stream.
Figure 1 above shows the Gulf Stream current situation (red vectors), with eastward
flow of the main stream coming in below the fleet, dipping sharply south, turning briefly
east and then north again. Our group of boats caught some boost (about one to two knots)
from the southerly flow at the eastern edge of a warm eddy in the early morning hours, but
we got a really big push as we followed the red route down through the stream (at about
four to five knots) and exited the Stream as it turned to the east. This was a wild ride,
with driving rain, thunderstorms, and violent wave action, described by many as their
roughest Bermuda Race ever. But it was fast. On board Morning Glory we logged some
13 knots through the water, plus four or so more on the speed-over-ground measurement, and
emerged from the Stream on a rhumb line reach toward Bermuda with an apparent lead and
good boat speed. Naturally, this is when the mainsail halyard broke, just at dark, and
shortly afterward Pyewacket came up from behind and sailed on by.
We ended up finishing third, hours behind the leaders, but the routing software had put
us in a position to win, and at least on even terms with some of the best navigators in
the world, including Stan Honey on board Pyewacket (line honors and new record) and
Volvo navigators Steve Hayles (on Boomerang in second) and Dee Smith (Blue
Yankee, which corrected out to first overall). The information is there, and the
technology available, to make any sailor a first-class navigator, and you don't have to be
a computer whiz to work with it.
David Brayshaw is the proprietor of Local Knowledge Marine
Software (the purveyor of GoFlow Current Systems and Force 4 Offshore Routing software).
For additional details on the race, see www.goflow.com/nb02race.htm.