New Grib Presentation

[Ptizef]

Originally Posted by rgleason

Awesome.
Humm, I changed the lat lon and got a box that did not move. I must have done it wrong and will try again.
It would be nice to eliminate that email that says there is an error. I wonder if I should test this further then contact saildocs somehow about their server requirements?

Do you think this is worthwhile implementing as a grib request in grib_pi? --Provided the feature is going to stick around?

Thanks, Rick

I have tested too
From Saildocs, it runs.
In O it 'almost' works although there are some minor problems to solve.
Also needs to add the possibility in the request dialog.
That is included in my to do list but I'm far from my tools at the moment.
I'll seeit mid February
Jean Pierre

[HUR450502]

Originally Posted by Ptizef

I have tested too
From Saildocs, it runs.
In O it 'almost' works although there are some minor problems to solve.
Also needs to add the possibility in the request dialog.
That is included in my to do list but I'm far from my tools at the moment.
I'll seeit mid February
Jean Pierre

... as requested by rgleason, I included in my mail the Saildocs request - repeat it here:

send GFS:40N,38N,004W,002W|0.5,0.5|0,3..180,192..384|PR MSL,WIND,SFCTMP,WAVES,RAIN|5.0,100,2014012715

and attached the resulting file.

Best regards, Uwe

[HUR450502]

.... attachment seems not to work, try again changed filetype to .grb.mpg - you need to remove the .mpg.......

[HUR450502]

Originally Posted by HUR450502

.... attachment seems not to work, try again changed filetype to .grb.mpg - you need to remove the .mpg.......

... do it again, change filetype to .grb.gpx - the .gpx needs to be removed.... Let's see, whether or not it works....

[HUR450502]

ok, the .grb filetype cannot be uploaded. If you need it, please send me an e-mail, and I will send it there, o.k.??

Best regards, Uwe

[cagney]

Originally Posted by rgleason

Awesome.
Humm, I changed the lat lon and got a box that did not move. I must have done it wrong and will try again.
It would be nice to eliminate that email that says there is an error. I wonder if I should test this further then contact saildocs somehow about their server requirements?

Do you think this is worthwhile implementing as a grib request in grib_pi? --Provided the feature is going to stick around?

Thanks, Rick

To get rid of the error mail just skip "send" in the beginning of the sentence.

 GFS:42N,34N,030W,022W|0.5,0.5|0,6..144|PRMSL,WIND,SFCTMP,WAVES,RAIN|7.0,130,2014012718

works fine. The last field is not strictly necessary, but keeps the size down if you're not interested in the full grib period.

Thomas

[rgleason]

Thanks Hur and Thomas,

Okay it works for me now. I changed the area around Horta and changed the Movement Vector to 7 knots 58 degrees, towards England

Email: query@saildocs.com
Subject: gribauto
First Line:

GFS:42N,37N,030W,022W|0.5,0.5|0,6..144|WIND,WAVES,RAIN|7.0,58,2014012718

Here is the grib that was received if you want to try, take the .doc off please.

[rgleason]

Thomas

This is a pretty nice feature. I don't quite get what the 2014012718 does to make it smaller. It does not seem to make much of a difference in these files. How does it work and how do you know that is what it is for??

Thanks.
Rick

[HUR450502]

Dear all,

don't know, whether or not this is of interest, but I generated an entry for 1... - 16 days.

The last field simply is the date and hour (UTC) when the request was created. Format is YYYYMMDDHH. What Saildocs does with it, I do not know ....

So here are all the entries, generated by Airmail:

send GFS:40N,38N,004W,002W|0.5,0.5|0,3..24|PRMSL,WIND,S FCTMP,WAVES,RAIN|5.0,100,2014012721
send GFS:40N,38N,004W,002W|0.5,0.5|0,3..48|PRMSL,WIND,S FCTMP,WAVES,RAIN|5.0,100,2014012721
send GFS:40N,38N,004W,002W|0.5,0.5|0,3..72|PRMSL,WIND,S FCTMP,WAVES,RAIN|5.0,100,2014012721
send GFS:40N,38N,004W,002W|0.5,0.5|0,3..96|PRMSL,WIND,S FCTMP,WAVES,RAIN|5.0,100,2014012721
send GFS:40N,38N,004W,002W|0.5,0.5|0,3..120|PRMSL,WIND, SFCTMP,WAVES,RAIN|5.0,100,2014012722
send GFS:40N,38N,004W,002W|0.5,0.5|0,3..144|PRMSL,WIND, SFCTMP,WAVES,RAIN|5.0,100,2014012722
send GFS:40N,38N,004W,002W|0.5,0.5|0,3..168|PRMSL,WIND, SFCTMP,WAVES,RAIN|5.0,100,2014012722
send GFS:40N,38N,004W,002W|0.5,0.5|0,3..180,192|PRMSL,W IND,SFCTMP,WAVES,RAIN|5.0,100,2014012722
send GFS:40N,38N,004W,002W|0.5,0.5|0,3..180,192..216|PR MSL,WIND,SFCTMP,WAVES,RAIN|5.0,100,2014012722
send GFS:40N,38N,004W,002W|0.5,0.5|0,3..180,192..240|PR MSL,WIND,SFCTMP,WAVES,RAIN|5.0,100,2014012722
send GFS:40N,38N,004W,002W|0.5,0.5|0,3..180,192..264|PR MSL,WIND,SFCTMP,WAVES,RAIN|5.0,100,2014012722
send GFS:40N,38N,004W,002W|0.5,0.5|0,3..180,192..288|PR MSL,WIND,SFCTMP,WAVES,RAIN|5.0,100,2014012722
send GFS:40N,38N,004W,002W|0.5,0.5|0,3..180,192..312|PR MSL,WIND,SFCTMP,WAVES,RAIN|5.0,100,2014012722
send GFS:40N,38N,004W,002W|0.5,0.5|0,3..180,192..336|PR MSL,WIND,SFCTMP,WAVES,RAIN|5.0,100,2014012722
send GFS:40N,38N,004W,002W|0.5,0.5|0,3..180,192..360|PR MSL,WIND,SFCTMP,WAVES,RAIN|5.0,100,2014012722
send GFS:40N,38N,004W,002W|0.5,0.5|0,3..180,192..384|PR MSL,WIND,SFCTMP,WAVES,RAIN|5.0,100,2014012722

Does this give you any further clue??

BTW- from day 8. on, there are only entries every 6 hours, which makes sense, as the prediction is by no means close to precise by then...

Best regards, Uwe

[rgleason]

Very helpful HUR.
I found a website which details Moving Gribs with the movement added.
Tips For Using GRIB Files To Receive Weather Forecasts For Boat Voyage


S/V Morgans Cloud

However, there are quite a few features of the SailDocs server that, when using email to access it—by far the most efficient method over sat phone—are only available by writing the email request directly, so that’s what I do.

Let’s assume that we are about a third of the way from Bermuda to the Azores in Morgan’s Cloud at position 33°58’ North and 054°42’ West.
Every day I would email the following script to query@saildocs.com
send GFS:36N,32N,052W,058W|0.5,0.5|6,9,12,15,18,24..72, 84..168|WIND|7.0,080
This looks intimidating, but with a bit of practice and reference to the instructions, you can easily decode it to mean:

• send me the GFS model forecast for wind only;
• for two degrees of latitude to the north and south and three degrees of longitude to the west and east of our position;
• advance that area assuming that our boat will travel at 7 knots on a heading of 80 degrees true;
• provide data on a half a degree grid (the most detail that the GFS model provides);
• starting six hours after the model run time and then for every three hours to 18 hours and then every six hours out to three days and then every 12 hours for seven days.

This script seems to yield the maximum amount of useful information about the wind over the next week for the minimum amount of satellite phone time. I start at six hours because the models are usually delayed by that much anyway, and reduce the forecast frequency over time since GRIB accuracy drops substantially after 72 hours, so there is not a lot of point in getting the data for every three hours after the first day. Having said that, I look out seven days because it’s good to get a warning if the model thinks my ears are going to get blown off, even if it does not happen.
.

So it looks to me like these techniques are going to stay around!!!

[HUR450502]

Hi Rick et al,

... yes, I believe, they definitely will stay around, as they will limit the amount of data significantly for such long period. This is especially useful, when you are requesting this via HF with limited bandwidth.

Let me "dream" for a moment:

OpenCPN "knows" the current ships position. What, if Open CPN could build such a request, by "asking" the user how many miles "around" the current position you are interested in the GRIB data and then estimating course and speed (to be entered by the user) and the wanted weather data and then "builds" this request, which then can simply be copied and pasted into the request message to be sent to Saildocs?? In a later version OpenCPN could directly send this via the apporopriate method to Saildocs?

The latter is not so necessary, as I believe there is a lot of coding effort necessary, and building a request and copying into an request message is sufficient in my opinion.

With the "Airmail" program you have three possibilities to send such request 1. Via Telnet (when you have an internet connection, 2. via HF (using e.g. Pactor or Winmor etc. ...) or via Satphone (Airmail povides for appropriate configuration).

If, however, an interface to "Airmail" could be built, this would even be more convenient (the transmission method then is to be selected by the "Airmail" user).

On the other hand, "Airmail" provides a graphical interface, where you can draw a square around your position (which "Airmail" knows as it can interface to the GPS and the position is shown in the graphics window) and then just enter your request data. The only disadvantage is, that you have to stop OpenCPN, wen you want to use "Airmail" as both are "competing" for the GPS usage....

However, this is probably a loooong way ahead and it would be most sufficient already, to just be able to display these "stepped" GRIBs in OpenCPN (which worked in former releases...).

Best regards,

Uwe

[rgleason]

I'll try again:

Email: query@saildocs.com
Subject: gribauto
First Line:
Code:
GFS:42N,37N,030W,022W|0.5,0.5|0,6..144|WIND,WAVES, RAIN|7.0,58,2014012718
Here is the grib that was received if you want to try, take the .doc off please.

25kb Size Pdf Doc seem to work. change end to grb.

[rgleason]

I like HUR's suggestions and suggest additions to the Saildocs grib request box:
Morgan's Cloud Example:
GFS:42N,37N,030W,022W|0.5,0.5|0,6..144|WIND,WAVES, RAIN|7.0,58,2014012718

Taking in pieces:
GFS:42N,37N,030W,022W (Do requests have to be in even degrees?)
Use Ship's position: (but Editable overide)
Box width x height: 4 (degrees) 6 (degrees)

|0.5,0.5
Resolution as in settings box dropdown, fine.

|0,6..144
Interval & Time range
From S/V Morgan's Cloud request,
Start 6 hrs after the model run time
then every 3 hrs to 18 hrs
then every 6 hrs to for 3 days
then every 12 hrs for 7 days.
How to do this?


|WIND
Data Types Request, as in the Dialogue

|7.0,58,2014012718
Movement Vector
Knots 7.0 (nm)
Direction 58 (True)
Date & Time to start grib: 201401718 (Jan 7, 2014, 18 hours in advance?)
What happens when we change 18 to 12 or 24? What changes?

Maybe have an "Efficient Grib:
-Checkbox that sets the above settings and allows change.

[rgleason]

I have added the above to the tracker FS#1087
FS#1087 : Retrieval: Support Airmail's Stepped "Moving" Grib - Add details to GRIB Request Box
and changed it from "Bug" to "Feature Request"

PLEASE VOTE FOR THIS FEATURE " MOVING GRIB" FEATURE on TRACKER!!!
There are only two votes so far!

[HUR450502]

.... of course, I would vote for this feature but do not have any means to do so, I am afraid.

However, coming back to the above post(s). Why not asking the user on which diameter or radius in nautical miles he/she wants the grib to be created and then "moved" by (user entered) estimates of ship's speed and course. OpenCPN would know the proper logitude and latitude (the longitudes would differ the nearer you get to the poles) by simple calculation and could build the "square" by the uprounded proper longitude and latitude?!

What do you think?

Regards, Uwe