Passage planning is a crucial aspect of any sailing journey, requiring careful consideration of weather conditions, route optimization, and boat performance. In this blog post, I will share my approach to weather passage planning using the Professional version of PredictWind, which has become an integral part of my workflow. While I am not a meteorologist and have no formal training beyond books and personal experience, I have developed a method that has proven effective over numerous voyages. This method combines pre-departure planning, real-time adjustments, and post-passage analysis to continually refine my technique. I have no affiliation with PredictWind, it is just the best tool I have found.
Weather Models
One misunderstanding I see many sailors make is the belief that one app or website provides a better forecast than another. This belief is usually untrue since most weather apps use the same weather data provided by the ECMWF (European Centre for Medium-Range Weather Forecasts) or the United States National Weather Service (GFS). The difference between the apps is only how the data is presented. Saying one app has a better forecast than another is like saying Taylor Swift’s Shake It Off is better on Apple Music than Spotify when it’s the same song provided by two different services. The weather apps I prefer to use are Windy and PredictWind. I like Windy’s user interface and how the weather data is organized. I use Windy daily to check the weather and to look at the weather on a macro scale to see the movement of high and low-pressure systems, fronts, etc. On the other hand, the Professional version of PredictWind is the app I use for passage planning. In my opinion, PredictWind is hard to beat when determining when to depart, the route to take, and the conditions expected along the way. Since this post is about passage planning and not an overview of the weather, I’ll talk about how I use PredictWind and show screenshots from their website.
Setting Up PredictWind
One of the powerful features of PredictWind is that it suggests a route optimized for your boat. I have heard people complain that PredictWind’s weather routing doesn’t work well, but that is usually because they don’t have their boat polars set correctly in the application. I admit this is frustrating at first and takes some time to get right, but it’s critical for the application to produce an optimized route. As I mentioned, I save my initial routing information and reflect on it at the end of the passage. This is when I end up tweaking my polars so future routing will match more closely to what I sailed. For example, we sail to about 80% of the factory polar provided by PredictWind for the Lagoon 450. We are a bit faster than the 80% numbers, but I want to error on the side of being conservative. It also doesn’t consider that we have a Code 0 and a spinnaker, so our light air performance is much better than they predicted. We rarely follow their suggested route exactly, choosing to sail dead downwind instead of gybing back and forth. We also motor less than they predict over long distances and motor more over short distances. With all this said, PredictWind’s routing gets me about 85% of the way there, and I use my expert knowledge about how we sail compared to their recommended route to fill in the rest. The takeaway is to invest some time into getting the boat’s polar data close to reality. Otherwise, PredictWind’s recommended route will be way off.
(Side note: I’m currently logging all our NMEA 2000 passage data to a database with the plan to build a custom boat polar instead of using the predefined polar in PredictWind. PredictWind has a product called the DataHub, which is supposed to do this for you, but I’m skeptical it will work well.)
Departure Planning
For this blog post, I will use a real-world passage we completed when we sailed from Green Turtle Cay in the Abacos to the Chesapeake. We planned to haul out Coronado at Zimmerman’s in Matthews, VA, at the end of May. We planned to be able to leave from The Bahamas to the Chesapeake starting at the beginning of May. That gave us four weeks to find a suitable weather window. Experience has taught us that the higher the latitude, the harder it is to find good weather for a time period long enough to make a journey more than a few days. We wanted to make the 750 nautical mile trip in one shot but knew we would likely have to sail north part of the way, anchor for a while, and wait for good weather to finish the trip. To spoil the end of the story, in this particular passage, we made it to Mathews, Virginia, on the Chesapeake Bay in one trip.
Once we arrived in the Abacos, I started looking for a weather window using the Departure Planning feature in PredictWind. This feature runs four iterations of the passage over a user-specified period of time. When I started looking for a suitable date to depart, I set the time period to be 24 hours, producing four passages separated by one day. I can compare the days to see if day one, two, three or four is preferrable for departing. Remember that I rarely find a perfect departure day, especially the longer the distance. It’s always a compromise between wind speed, wind angle, sea state, and the general state of the weather (usually thunderstorms). Experience has taught us to take the first decent weather window. If you wait for the weather to be perfect, you risk wasting a lot of time. We typically don’t compromise on rough seas on the quarter, sailing upwind (unless the air is light) and if there are severe thunderstorms. High wind and large following seas aren’t our favorite sailing conditions, but they’re manageable. Anna gets seasick when the seas hit us on the quarter, and the boat sloshes around like a washing machine. It’s worse when it occurs on the first or second day before she has her sea legs. Time and experience will help you understand which conditions you and your crew can tolerate and which you can’t.
Don’t pick a start date too far into the future when using Departure Planning. The weather forecast is usually only accurate for about 72 hours from the last forecast; after that, it becomes a guess. When looking at forecasts, know that weather models are better at predicting how an already established pressure system or front will evolve than predicting where and when a new pressure system or front will form. The further into the future you look at the weather, the more likely you’re looking at the latter prediction than the former, which becomes more of an issue when you’re looking at longer passages. The weather at the end of your passage is a big unknown if you’re looking at departing four days from today to do a five-day passage. The weather on the final day of the passage is now nine days into the future! This is why even when it appears I have a great weather window; I keep running the Departure Planning models every day to see the trend. There have been numerous times I’ve told Anna, “We’re leaving on Friday, be prepared…” only to say on Thursday night, “I’ve changed my mind, we’re leaving on Sunday instead…” This brings up an important point: a good sailor will change their mind on the departure date if the forecast starts trending negatively; a bad sailor will leave anyway just because they’re committed.
When you run Departure Planning, you must select four weather models out of six to calculate the route. The models are ECMWF, GFS, Spire, UKMO, PWE and PWG. (For more information on each weather model, click here.) I find the ECMWF to be the best all-around and consistently good. If I’m sailing short distances near land, I use the PWE and PWG models since they include the effects of land on wind speed and direction. Offshore, I find Spire to be good, and GFS is excellent at predicting tropical storms, so I’ll include them instead of PWE and PWG. If I see one model way off compared to the other three, I’ll likely swap it out for another. Keep in mind that the four model limitation is only for Departure Planning. The Weather Routing feature uses all six models.
Looking at my output from Departure Planning for the Abacos to the Chesapeake Bay, Departure 2 is a good day to leave. I’m making this conclusion based on several factors. The Motoring Time is only 1.5 days (on average). About 94% of the time, the wind will be greater than 8 knots. The CAPE (Convective Available Potential Energy) value measures the energy available for convection, indicating the potential severity of thunderstorms. The CAPE value is a bit high per the GFS model, which is typically pretty good at predicting thunderstorms. You usually need a high CAPE value (over 1,000) with rain to have thunderstorms. I would keep my eye on this one since it’s getting close.
The waves will be less than 2 meters, which is very important to us, but more important is the effect of the sea state on the boat. That’s where the four lower sections in the screenshot come into play. This is where PredictWind excels, in my opinion. The amount of roll will be less than 4° for the entire passage. This is important for us because anything over that is where Anna gets seasick. Vertical Acceleration is caused by the boat bashing into waves, which is usually over 0.2g’s when sailing upwind in rough seas. The final section has a lot packed into one number: the sea state’s impact on the boat’s performance. Less than 100% of boat polar means the seas are rough enough to slow the boat. Overall, Departure 2 looks reasonable.
Once I narrow down the day I want to leave, I rerun the Departure Planning to see if there is a preferred time to go. In this case, we’re sailing such a long distance that our departure time doesn’t matter much. But if it was a shorter journey, I may have better weather if I leave in the afternoon versus the morning. Of course, with shorter sailing distances, we also have to consider our arrival time, so we may not pick the most optimal departure time because it would cause us to arrive in the middle of the night (something we avoid).
Weather Routing
Now that I have my departure date, I’ll switch to the Weather Routing feature in PredictWind. In this case, I chose to leave at 1100 to time our departure with slack tide at the Nunjack Channel, the point where we plan to exit the Abacos into the Atlantic Ocean. The following image shows the suggested routes based on each model’s prediction of the weather (using my boat’s polar as explained above). All six models suggest we head northwest and follow the wind as it veers around.
The combined wave and swell look good to us. Since we’re sailing offshore, I would consider this sea state fine.
The Roll and Vertical Acceleration are approaching the uncomfortable zone, but when I reconciled this with the route, it was apparent this is caused by the Gulf Stream. We expect the ride to be rough getting into and out of the Gulf Stream. We also downloaded the forecast for ocean currents and tracked our location relative to the forecasted location of the Gulf Stream. The PredictWind routing model considers the effects of currents on your route. If there is a benefit to sailing into a favorable current, it will route you in that direction.
I mentioned before that the high CAPE value caught my eye. We might have some thunderstorm activity on May 5th and 7th. We need to continue monitoring this, especially in lighter air, where we don’t want to get caught with our spinnaker or Code 0 up in a squall. You cannot do much to avoid lightning, but we try.
Once we are on our way, I will download the weather twice daily using the PredictWind Offshore app. This version of the app saves all the data offline, which is handy for analysts after the passage and for safety in case we lose communications while en route. Since the weather is constantly changing, so will your route, so it’s essential to rerun the weather router based on your current position. I typically calculate a new route in the morning and the evening before sunset, which coincides with when all six models PredictWind offers are updated. It also allows us to discuss our strategy for night sailing before we start our sleep rotation.
The Passage to the Chesapeake
Since I took these screenshots before we completed the passage, you may be curious how it matched reality. We departed Green Turtle Cay at our planned time of 1100, getting through Nunjack Channel around 1200. Once out of the channel, the wind was light(ish), around 10 knots, so we decided to sail with the Code 0, which put us on a route similar to the red line. As the wind speed picked up and veered around, we furled in the Code 0 and pointed higher with the genoa, heading towards our rhumb line. By the middle of the next day, we had to turn on one engine and motor-sail for part of the day. By the end of that day, we had the engine off and were sailing again. The sea state during this time was on the beam, but not uncomfortable. Sunday (day three) we encounter squalls with a few thunderstorms. None of the squalls hit us, but flashes of lightning in the distance certainly kept us worried. The thunderstorms were to our west, along the coast of Georgia, and only made their way east after we had sailed past. Approaching the Gulf Stream, the seas were rough. It was rougher than both the data and I had anticipated. However, once we were firmly in the Gulf Stream, the seas flattened out, and the 2 to 3 knots of current put a smile on our faces. It wasn’t until the fourth day, May 7th, when we were past Cape Hatteras, sailing along the Virginia coast, that we experienced the second round of thunderstorms. This time, the rain and lightning were offshore, to our east, and once again, didn’t impact us other than make us worry.
The PredictWind forecasted route and sea state were about 90% accurate in my estimation. The rough sea state approaching the Gulf Stream was worse than they predicted, but not a surprise. The two times we encountered thunderstorms were in the data, but only the GFS accurately predicted it. Our conservative boat polar predicted us to arrive Tuesday night, but we dropped the anchor Tuesday afternoon. This was expected since we sailed faster. We also motored around 20 hours, less than the 38 hours they predicted. Faster sailing and less motoring are shortcomings caused by the boat polar used by the PredictWind models, and we are aware of this and will consider this on any passage. The wind speed and direction were pretty close to what was forecasted.
Conclusion
In summary, successful passage planning requires a combination of reliable tools, practical experience, and continuous learning. PredictWind has proven invaluable in optimizing our routes and ensuring safer, more efficient passages. However, it is essential to recognize that no tool is perfect. The key to maximizing the benefits of PredictWind lies in understanding its capabilities and limitations, accurately configuring your boat’s polar data, and interpreting the forecast data with a critical eye.
By regularly reviewing past voyages and adjusting our strategies based on what we learn, we continue to refine our approach and improve our passage planning skills. Remember, while technology can significantly aid in planning, the skipper’s judgment and experience are irreplaceable. With a thoughtful, informed approach, you can make the most of the resources available and enhance the safety and enjoyment of your sailing adventures.
Sailing Coronado 