The Navigatrix has been updated. The new website can be found at navigatrix.net.




Post new topic Reply to topic  [ 5 posts ] 
Author Message
 Post subject: Slocum's Route Forecaster - Asking For Your Input
Site Admin

Joined: 20 Mar 2012, 13:32
Posts: 116
I'm working on a little piece of software that would allow you to view GRIB wind forecast data along a route in OpenCPN - kind of a "Saildocs moving spot forecast on steroids". Let's call it "RteFcst" for this discussion.

Example: Say you have a route defined in OpenCPN like the one below (we do follow a strange course sometimes...).
Attachment:
route.png
route.png [ 26.75 KiB | Viewed 13772 times ]

Assume you also have GRIB file with wind forecast data that covers/overlaps the route in location and time. Then RteFcst would combine route and forecast (plus info about you average boat speed and departure time), and provide you with a gpx file that would show up like this when loaded as a temporary layer in OpenCPN:
Attachment:
rtefcst.png
rtefcst.png [ 57.94 KiB | Viewed 13772 times ]

The little arrows show true wind direction with a "N-NNE-NE-ENE..." granularity. The label at each arrow provides apparent wind angle and speed, and the time for which this data is valid (which is when you would be at that position based your average boat speed and departure time). For example, "29R@21 00Z03Feb" forecasts the apparent wind to be 29° off the port bow with 21kn at 00:00utc on February 3 - starting to get a bit too close for comfort... "1G@24 06Z03Feb" means apparent wind 1° off the starboard bow ('Green') six hours later. Alternatively you can have the labels with true wind info and ommit the time/date portion if it gets too cluttered.

So here's my first question for you: do you find this useful? No need to be polite - 'no' is a fair answer. If you do find it useful - are there any tweaks that would make it better?

Now the second part: How about you don't have to get a separate GRIB forecast but could simply copy and paste the OpenCPN route into an email request? In response you would get the same gpx file as above but at a fraction of the file size of a GRIB forecast for an area that covers your route (think SSB/Pactor or sat phone connections).

This could work like this: In the "Route Properties" window in OpenCPN you enter your average speed and your expected departure time. Then you right-click anywhere in the waypoint list and select "Copy all as text".
Attachment:
route_properties.png
route_properties.png [ 125.36 KiB | Viewed 13772 times ]

Then you simply open an email and paste the the whole lot into the mail.
Attachment:
email_claws.png
email_claws.png [ 119.07 KiB | Viewed 13772 times ]

Send off the message to a yet undisclosed email address, and back comes a gpx waypoint file with wind data along the route that can be displayed in a temporary layer in OpenCPN as shown above.

OK - question number two: Would this type of email request interface be useful? Can you think of easier/better ways to encode the route information into a request email? Any problems you can see with this approach?

If you do think this "route forecast request" would be useful, I have a favor to ask of you:
Could you please follow the procedure above for any route you may have in OpenCPN ("Copy all as text" in Route Properties window and paste into email) and send the resulting email message to rtefcst [at] namaniatsea.org

There will be no response to those messages and they will not go beyond my inbox. I'm simply trying to get an idea of how these requests end up looking for different locales and OpenCPN configurations. Also different email clients and email providers might impact the final format and content (automatically appended footers, etc). If we end up going this route this will help to put together a robust request parser.

Many thanks in advance for your help!

Now - what has all this to do with Slocum (see title of this post)?? RteFcst is a small extension to Alex's Slocum project. Alex has found a very ingenious way to reduce the size of GRIB files by a factor of 4-5 by 'simply' by cutting down the precision of wind speeds from '15.34776534223kn' (current GRIB format) to 'Force 4'. From Alex's project page on GitHub:
Quote:
Why Slocum?
-----------

In "Sailing Alone Around the World" he talks about his tin clock:

"My tin clock and only timepiece had by this time lost its minute-hand,
but after I boiled her she told the hours, and that was near enough on
a long stretch."

If Slocum was able to successfully circumnavigate via celestial navigation
using a clock with no minute hand, then modern day sailors should be able
to get by using weather forecasts without floating point precision.

PS: The code for RteFcst can be found at https://github.com/ms8r/slocum/blob/wpfcst/python/sl/lib/rtefcst.py - work in progress though...


Top
   
 
 Post subject: Re: Slocum's Route Forecaster - Asking For Your Input

Joined: 21 Feb 2013, 12:46
Posts: 76
Well,

think I´ts a good idea.
The problem I had crossing from Panama to Australia was that if I copied the gribfiles from Airmail into Open CpN the area covered was usually way to big. on the other side, I´ d like to see waves ond Isobars as well.
Opening Gribfiles from Airmail into CPN is pretty basic anyway..., especially if you have to use a superslow windows system.
Hope that is a helpful input.

Atmo


Top
   
 
 Post subject: Re: Slocum's Route Forecaster - Asking For Your Input
Site Admin

Joined: 20 Mar 2012, 13:32
Posts: 116
Thanks Atmo. I also like to see isobars over a bigger area. The problem with wind data is that it tremendously inflates GRIB file size because wind strength is stored with full floating point precision and two vector components for each grid point - no compression will be able to squeeze that out. A real pain at data transmission speeds of < 1Kb/sec... So what we typically do is getting two files: One with pressure and rain data only, across a large area (synoptic scale, 24hr intervals, 7 days out - the entire SW Pacific is still less than 10kB in Airmail), and another with spot wind data along the route (2 days out).

What I'm trying to do with Rtefcst is simply to have spot data along the route overlaid with synoptic scale isobars (which I can still display with OpenCPN's GRIB overlay). Also, I'd like to have a bit more flexibility than the Saildocs built-in moving spot forecast provides (e.g. not sailing a straight line, specifying a starting time for the forecast, and showing apparent wind speed and direction).


Top
   
 
 Post subject: Re: Slocum's Route Forecaster - Asking For Your Input

Joined: 18 Sep 2013, 13:02
Posts: 13
It may be useful to have windspeeds shown with fletched arrows so one can see them at a glance (as well as numerically?), and perhaps even higher windpeed arrows shown in red?

cheers Jules


Top
   
 
 Post subject: Re: Slocum's Route Forecaster - Asking For Your Input
Site Admin

Joined: 20 Mar 2012, 13:32
Posts: 116
Thanks Jules, I like the color idea. Because I'd like to be able to use OpenCPN's GRIB display to overlay other forecast data (say isobars), I'm using OpenCPN user icons for the route point windarrows. Barbed/fletched arrows with the usual 5kn resolution would multiply the number of icons required by 7+, but a simple "red-yellow-green" coding would probably already help? It would also be more in line with Slocum's (the SW program) overall philosophy of a data resolution that's "good enough" for cruising (i.e. Beaufort scale).


Top
   
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 5 posts ] 


Search for:
cron

Credits © 2010 - 2024