Current-informed routing plugin

[rgleason]

Woly,
We alreeady do that. Please read the manual. It is used for current gribs.
Thanks.

https://opencpn.org/wiki/dokuwiki/do...s_wind_current

It is a lot of documentation, but WxRte works closely with grib_pi, climatology_pi and ocpn_draw_pi.

Can you download a copy of this data and post it here? Or give us a link?

[rgleason]

Walt, about "nohal's tide display plugin". if this is OTcurrent, Pavel may have helped and compiled it, but I believe Mike Rossiter [Rasbats] worked on that. He would be a good one to talk with if you have questions about it.

I've enjoyed batting ideas with you.

[SerialHobbyist]

I also sail in the PNW and want pretty much the same thing! I understand it won't be "perfect", but it would be darned helpful, I think.


Right now you can reference a tidal station from a waypoint -- however it doesn't allow current stations. At least it gives you a little info in the Route Properties that way.


Trying to come up with better ways, without throwing the baby out with the bathwater, I was thinking of adding a more formal "link" to a waypoint. Perhaps a drop-down with a distance sorted list of current stations you could select from.


Then it would apply that current correction at that waypoint ETA/ETE to the default route speed for that leg. You could also add the vectors, perhaps, for a bit better accuracy.


Chas

[SerialHobbyist]

Another route planning tool I put into my own program many years ago, was to have the "ETA" field in Route Properties editable.


That way I could, for instance, set the ETA of a waypoint to the predicted slack water time for that channel. It would then calculate backwards to create a new departure time.


I found that very useful.


Chas

[seandepagnier]

Current stations are more of a guideline, so you can attempt to travel with favorable currents, and escape or reach slower currents when the tide changes.

Current stations do not consider rainfall, wind-based tide, or the effects of current ocean currents and other factors. The actual tidal current can vary widely between stations, and even very near the station. The boat speed to get to one station changes the time you arrive at the next and so the error quickly compounds.


It's just not possible to use existing tidal stations for routing in the ways discussed in this thread. You can pretend if you want but much higher resolution tidal data is needed and even then it isn't very accurate.

If you can get a grib with currents, use the weather routing plugin which handles currents already. If you want to take advantage of currents, it makes no sense to plot a route, you need to compute the best route.

[did-g]

Hi,

I have a version of OpenCPN+ weather_routing which is able to compute speed along a route, adding current stations should be doable. Of course it'd be coarse.

[SerialHobbyist]

Originally Posted by boat_alexandra

Current stations are more of a guideline, so you can attempt to travel with favorable currents, and escape or reach slower currents when the tide changes.

Yes.

Current stations do not consider rainfall, wind-based tide, or the effects of current ocean currents and other factors. The actual tidal current can vary widely between stations, and even very near the station. The boat speed to get to one station changes the time you arrive at the next and so the error quickly compounds.

That's why it is called Estimated Time of Arrival.

It's just not possible to use existing tidal stations for routing in the ways discussed in this thread. You can pretend if you want but much higher resolution tidal data is needed and even then it isn't very accurate.

I'm not looking for an estimate that needs to be correct within seconds or even minutes. But, as Walt pointed out earlier, in the PNW missing a tidal change by 30 minutes can change the total route time by 2 hours+.


Maybe we're not talking the same thing. I'm not looking for auto-routing that calculates the best route for me. I'm looking for a tool that will help me choose an appropriate departure time for a planned route.


Chas

[rgleason]

If you can get a grib with currents, use the weather routing plugin which handles currents already. If you want to take advantage of currents, it makes no sense to plot a route, you need to compute the best route.

Sean, even the RTOFS gribs which have a misson statement to include coastal tidal effects, do not even begin to get this right. The model is still best used for longer voyages and crossings.

For Example, the Gulf of Maine and Bay of Fundy (with this most remarkable configuration) tides and currents have been studied and modeled extensively to determine the effect of water turbines, salinity, sea life etc. no doubt with grib files being used.

Do you think they have taken these models and incorporated them into the model? NO they have not. Why? Maybe because it was not done under NOAA direction or something, or maybe because Trump has severely restricted NOAA funding and would dismantle it if he could. But the end result is RTOFS for Gulf of Maine is not nearly as good as it could be.

So what''s accurate?? Nothing.

That is why I wanted some tools to make my own rough grib file on the fly, that wx_routing would then use in addition to the RTOFS sea currents.

[Rasbats]

I fully appreciate the limitations of tidal current routing! But ...

I have got an alpha version of a tidal routing plugin up and running on GitHub if you want to see where I am heading with current gribs.

https://github.com/Rasbats/otidalrou...tag/v0.1_alpha

In order to allow for short route legs or a short distance between EP and waypoint the vessel is moved forward in 3 minute increments. If a 3 minute leg skips past a waypoint the short distance between the last EP and the waypoint is allowed for in the calculation of the EP after the waypoint. A particular problem was the case where there was a short route leg on starting the route.

When using the gribs you need to make sure that the route goes through areas where there are arrows at each corner of the rectangle. You can see from the attached screenshot that if the route passed closer to 'Gregory Rocks' this would not work, as the current becomes N/A.

On Linux I have not stopped the plugin dialog disappearing behind the main OpenCPN program. Need to use the toolbar to call it back into view!

An idea I have started looking at for using tidal current stations is to create an initial tidal current station table from a DR route. This table is then queried to find the tidal set and rate from the appropriate station as the vessel moves along the route.


With the present plugin it is possible to calculate routes with up to 6 departure times. Comparing these (Summary), the departure time with the shortest transit time compares well with information from Pilot Books. I looked at routes from Weymouth to Brixham and from the Lizard (Mullion Cove) to St Ives.


Mike

[rgleason]

Cool Mike. I'm going to have to try it out. Thanks.

[Rasbats]

Initial guide is here:

https://opencpn.org/wiki/dokuwiki/do...ns:otidalroute

Biggest problem is the GRIB coverage. I have had a number of failures where a part of the route is not covered by the data, although apparently correct. Probably starting a test well away from the coast is the best idea to get a feel for what is possible.

Mike

[rgleason]

Mike this is a clever plugin. I've kind of got it working but am having trouble geting it to complete.


In the USA the GFS gribs do not include current, so I also downloaded an RTOFS for the same period. Then I highlighted both file when selecting what Grib files to use (from Grib_pi). Then "current" checkbox is shown in the grib display. The current grib is a rather large interval and stays pretty much out to sea.



I created a circuitous route in the area of the current such that there were at least 4 current gribs around each route point. See attached. I tried various times and also tried moving the start point. I still get the error. Any suggestions ?


The DR seems to work, but it showing no tidal current! I wonder why. Is this all due to using two separate gribs? one for wind and one for current?

[rgleason]

I moved across the Atlantic and used current charts from CURRENTS
as you suggested in the manual. Voila, it works!

This could be a very very useful plugin for both sailors and powerboats to save fuel costs, etc.

4 screenshots -
1. Picking two files from grib_pi - a current grib file and a wind grib file. (Screenshot)
2. Create your route using Opencpn and export it.
3. Load the current and wind gribs. Set the start time.
4. Start the otidalroute. Set the start time to match grib_pi
5. Hit Calc EP, pick the the exported gpx route.
6. Wait and a message show appear (see screenshot)

7. The results are in the Route Summary file and more detailed Tidal Route Summary (screenshots)


Thank you Mike. Any idea why it is preferential to the east side of the atlantic?

[stelian]

Hi Rasbats,

I have compiled the plugin under Linux, after a few code tweaks (I'll send you some PR for those a bit later), it starts fine and works for DR route calculation, but fails for me for the ETA route calculation.

I'm attaching my very simple test route (rename from .doc to .gpx) and I've put my grib (containing both wind and current data at http://popies.net/tmp/MANW250-2257.grb.gz , it's a 70 MB file)

Is this a Linux specific problem or can you (or somebody else) reproduce this on Windows too ?

Also, can you please also include the .fbp file in the git ?

Thanks,

Stelian.

[stelian]

After some debugging, it turns out that this came from the grib format changes in O5, so I've adapted the corresponding commit from weather_routing (credits to did-g) and tidal routing compiles and runs ok for me on Linux (well, except from some bugs here and there, like clicking on cancel doesn't do what it should, deleting a computed route sometimes crashes OpenCPN etc).


PR submitted.


Stelian.