Tidal Gribs can't be read with 5.8.4?

[AedanC]

Hello,


I have discovered that OpenCpn version 5.8.4 with Grib reader 4.1 cannot read tidal grib files that could be read by 5.8.2. It indicates "no valid GRIB file".


5.8.4 running on Windows 11 Pro, 5.8.2 running on Windows 10 Home.


An example failing file can be found at https://drive.google.com/file/d/13xv...usp=drive_link


If any more information is required please let me know.


Aedan

[Maelruan]

Aiden C,
I have same problem with IMI Irish Sea gribs for past few weeks using OPenCPN 5.6.2 on Win 10. The file download looks 1 bit short, i.e. 21,174 instead of 21,175.
Email sent just now to IMI.
Maelruan

[AedanC]

Originally Posted by Maelruan

Aiden C,
I have same problem with IMI Irish Sea gribs for past few weeks using OPenCPN 5.6.2 on Win 10. The file download looks 1 bit short, i.e. 21,174 instead of 21,175.
Email sent just now to IMI.
Maelruan

I downloaded the IMI grib for the West Coast yesterday (the one linked above), it opened fine with 5.8.2 but not 5.8.4. Please let us know what the IMI say.


I'll be out on the boat today and I'll try today's version.


Aedan

[AedanC]

On further investigation it turns out that neither 5.8.2 or 5.8.4 will read the recent tidal gribs from the Marine Institute, but both will read older ones, so it looks like their format may have changed, I’ll take it up with them tomorrow.

[did-g]

Hi,


It looks like they are using a more recent table (updated 22 years ago) , sort of:
https://codes.ecmwf.int/grib/format/grib1/ctable/4/
https://www.nco.ncep.noaa.gov/pmb/do...88/table4.html

As far I can tell OpenCPN is using Zygrib one which doesn't decode 13 (15 mn) an 14 (30 mn)


Sort of because P1 is always 0, at least in your sample, so they define it but don't use it.

diff --git a/plugins/grib_pi/src/GribV1Record.cpp b/plugins/grib_pi/src/GribV1Record.cpp
index 9c2d92a8d..ffa97fd62 100644
--- a/plugins/grib_pi/src/GribV1Record.cpp
+++ b/plugins/grib_pi/src/GribV1Record.cpp
@@ -928,6 +928,7 @@ zuint GribV1Record::makeInt2(zuchar b, zuchar c) {
     return ((zuint)b<<8)+(zuint)c;
 }
 //----------------------------------------------
+// https://codes.ecmwf.int/grib/format/grib1/ctable/4/
 zuint GribV1Record::periodSeconds(zuchar unit,zuchar P1,zuchar P2,zuchar range) {
     zuint res, dur;
     switch (unit) {
@@ -943,6 +944,10 @@ zuint GribV1Record::periodSeconds(zuchar unit,zuchar P1,zuchar P2,zuchar range)
             res = 21600; break;
         case 12: //    12 hours
             res = 43200; break;
+        case 13: // 15 minutes
+            res = 900; break;
+        case 14: // 30 minutes
+            res = 1800; break;
         case 254: // Second
             res = 1; break;
         case 3: //     Month

[Maelruan]

AedanC,
IMI say they will try to fix the problem. They have had some updates to their system & their GRIBS are readable on other packages, but now find OpenCPN will not read them. I've asked for details of those packages.
They know that older versions of OpenCPN could not read netcdf format GRIBS, but ask if latest versions could. I found a reference to that format in the manual for plugins, but quite an amount of programming seemed to be necessary to allow it to work.
Hopefully the Institute can sort it out soon - their GRIBS are the best for these waters!
Maelruan

[Maelruan]

Further to my last post I've just checked QtVlm, and it is reading IMI GRIBS.

[AedanC]

I had emailed my contact but she is away until tomorrow, I’ll cancel my mail to her as you seem to have it under control. Thanks.

[Maelruan]

AedanC
We've got our GRIBS back!
Many thanks to the Marine Institute for their speedy response!
Maelruan

[AedanC]

Maelruan,

That was a mighty fast response all right. I’ll give them a try later on today, I’m on the road at the moment.

Cheers,

Aedan