(These setting are enabled through the application preferences.)
|Unless you are sailing in confined waters, you are encouraged to use the automatic time step method.|
When this setting is disabled, which is encouraged, the solver uses a sophisticated method of automatically choosing an appropriate isochrone time interval, based on a number of factors.
Without going into great detail, the solver starts with an initial time interval of 20 minutes. As the algorithm progresses, if the solver is encountering a lot of land intersections, it may choose to keep the time increment small. If the solver finds itself sailing in open waters then it starts to slowly increase the time step.
The maximum isochrone interval the solver chooses is 2 hours.
As the solver sees that it is starting to approach the target point, it will start reducing its time step. This has the effect of generating additional isochrone points, which generate more accurate arrival information.
If you have a preference for generating all the isochrones at a single, fixed, uniform time interval then you can enable this setting and specify the interval.
One situation where you may want to set a fixed, small time increment, is if you sail in waters closely bounded by land. The initial time step when using the automatic method is 20 minutes. If 20 minutes of travel represents a long distance for the waters you sail, you may want to set a smaller fixed time step. Note that, even in this situation, you should first try the automatic time step.
In this example open ocean passage, there are three paths generated, each with a different time step, and they are all essentially all the same. The time duration of the paths are, from left to right: 8 days 8.6 hours, 8 days 8.8 hours, 8 days 8.8 hours.
The largest difference between these three examples is the amount of time taken to compute the solution. On a 2017 13” Macbook Pro, the average of three runs for these examples is: 0.588 sec, 0.890 sec and 2.95 sec. (If you have experiences with a different weather routing system, how does its performance compare for these examples?)
For open ocean routes, using the automatic time step is highly recommended.
When weather routing on inland routes, close to land, you may need to override the automatic time step to force the solver to generate smaller steps, giving it more opportunities to get around obstructions.
Consider this example:
In this example, the router was unable to find the route north, from the entrance to the island group. Instead, it routed around a larger island and went pretty far out of the way of the more optimim route.
Look at the right hand image, which is zoomed in and showing the fleet cursor at the point where the router was expected to go north. All of the possible paths radiating out of that point either violate the upwind sailing angle limit (the yellow lines) or intersect with land (paths are not allowed to cross land.)
If the solver was using a shorter time interval, each of the lines at this point would also be shorter, allowing some of these paths to not hit land, possibly giving the solver an opportunity to find a better route.
After adjusting the solvers time step to 0.1 hr (6 minutes) the following route is generated:
By taking smaller steps, the solver generated more possible routes and found a better path north. This second route is 6 hours and 42 miles shorter.
For constrained, inland routes, using a constant short time step is recommended.