wxtide32 in OpenCPN

[reflectionsv37]

I was just comparing a few different locations in Terry's file vs. WXTides32. Bangkok and Phuket are just a couple of minutes different between the two. But, other locations along the west coast of Thailand, i.e. Pulau Lelas are off by about 1 hour.

It appears to having something to do the UTC time being assigned to the stations that is different. Bangkok shows UTC +7:00 in both files and times are a few minutes apart. Pulau Lelas shows UTC +7:00 for both files but times are different by about 1 hour. Bangkok is a tidal reference, Pulau Lelas is a tidal substation based on Belawan, Sumatra Indonesia. For Phuket Harbor, also a tidal substation, both show +8:00 and the times are only a few minutes apart. What does this mean? I have no idea!!

Just looking at the times between Phuket and Pulau Lelas, I suspect Terry's file has the correct time but I'm too far away from Pulau Lelas to look outside to see for myself.

[rgleason]

p_dub has brought up changes in daylight savings quite frequently, as an issue that must be resolved.



I believe this really has nothing to do with Opencpn which is just using the harmonics as presented. Changes are needed to the files I think.

[P_Dub]

Originally Posted by rgleason

p_dub has brought up changes in daylight savings quite frequently, as an issue that must be resolved.
I believe this really has nothing to do with Opencpn which is just using the harmonics as presented. Changes are needed to the files I think.

The files works OK in WTides but not in OpenCpn. So this has something to do with the daylight time not use anymore in many Countries. One Should take a look at this please.
Best Regards.

[reflectionsv37]

I don't think the discrepancy between the 2 files has anything to do with the OpenCPN. I was simply comparing Terry's file to WXTides32. Apples to oranges comparison. I actually think there is an issue with the WXTides32 file for that particular station.

I know the WXTides32 file is correct for Phuket and have had issues with tides being incorrect when using the Pulau Lelas station. The inconsistency seems to be within the WXTides32 file.

Sorry if I implied there was an issue with OpenCPN. And FYI, no daylight savings time in Thailand, Malaysia or Singapore. But, Thailand is an hour different than Malaysia. Hard to believe, but Malaysia and the Philippines use the same time zone.

[rgleason]

Yes, I believe the solution(s) are small changes needed too the files. OpenCPN provides the framework to run the files and is not the culprit. Someone very interested in solving the problem needs to learn the format and make the right changes. I have no interest in doing this right now, nor do the developers.


There should be Tides thread to share knowledge about how to make the changes.
Eventually that could be added to Supplemental wiki documentation.


If a file is improved there should be some mechanism to make it publicly available that describes exactly what was changed and improved, otherwise that work is lost.

[P_Dub]

rgleason would you please check your private message. I have send you my email if you are interested in the subject.

As far as those stations mentioned above, the main differences are cause by using a different recorded set of data for the analysis and also using or not daylight saving time. Further more, the proram use for the analysis (either Task2 of BA or xtide or others and the number of constituents use in the analysis. Of course, one would think that it is all the same but unfortunatly not.

Best Regards

[SerialHobbyist]

Originally Posted by stormalong

That is the solution. Thank you very much. Could we get that into the documentation?

I use the .tcd file from the xtides website (https://flaterco.com/xtide/). Isn't wxtide32 based off that? Anyway, xtides gets a new version every year.



I am also in the Pacific Northwest, and I manually edited the HARMONIC.IDX file to remove many stations that were duplicate from xtides because they appeared to have lower accuracy.

[rgleason]

Pdub will check pm, but am out sailing w limited inet.

[skipperearly]

I know about the copyright issues on the tidal data in UK Irland Atlantic France Portugal and Gibraltar but want to enquire if someone has made harmonix file enhancements for this geography - obviously such downloads would be only used for testing

[DavidBurch]

We have been checking world tides in the stock harmonics provided by OpenCPN by comparing multiple stations from all nations represented with official data from those nations. I will write this up shortly to show the data. These international tides are a notable asset of the program.

All checked out well within the accuracy expected, especially US data, which are very close… and then we checked Sydney AU, and it was way off. After some effort, I think we know why and how to fix it.

There is this note in the change log, top of the Harmonics file

# 2000-01-26 Replaced Sydney constants that were fudged using Harmgen
# 2.0 with marginally better constants done right using Harmgen 2.1.

This follows an earlier correction:

# predictions starting in 1999. Updated Sydney, Australia with
# constants derived by Harmgen 2.0 from raw data provided by
# David Gorham (DGorham@mhl.nsw.gov.au) covering 1996-06-30 to 1997-07-01.

These corrections have left two complete sets of Sydney constants in the file, which can be found searching on -33.8500.

If you delete the first set, the Sydney tides are then much better, as shown in the attachment. We then changed these latter constants to the latest official values and we get even better data except for the one low tide. I recall seeing a note in the forum or manual that a low tide discrepancy might be a computational issue rather than constants… but I cannot find that reference now.

I am not sure how the pros do this, but we just edited the harmonics file with Xcode in a Mac and did not change the idx file. Not sure how this should be done. I post this to learn more about it. It seems that even though there are many stations based on Sydney, the constants for them seem all independent and more or less ok. Also attached is a spread sheet with the data.

Note that OpenCPN uses +11h (ie daylight time) whereas it is winter there and the correct value is +10h, so we need to make an hour correction in the comparison with official data, available for 2019 at this location: Tide Predictions for Australia, South Pacific and Antarctica. It would seem we would want to remove that first set of Sydney data at some point in the future.



https://www.starpath.com/temp/sydney..._compared.xlsx

[P_Dub]

Has you can see the cat is always coming out of the bag. I was the devil when I stated that and of course I was wrong, according to a suppose to be specialist that had ckeck it.
So now what is next!
Best regards.

[HappySeagull]

To edit the timezone, goto the HARMONIC.IDX and change it,
merely as demo that you can do on a copy:

example:
TNAP:Can:BC: -126.9333 50.5833 -8:0 Alert Bay, British Columbia


change to
TNAP:Can:BC: -126.9333 50.5833 -10:0 Alert Bay, British Columbia

you need edit both HARMONIC and HARMONIC.dx or, even while the tide may be "corrected", the timezone indicated in the graph will not be.

Don't change it per "daylight savings".

don't use windows notepad that add in their peculiars to a "text file".(I don't know about Macs Linux works fine.)
use a Hex editor, or Notepad++ in Windows...



(old versions of OpenCPN: Delete MRU_DAT in opencpn before restarting with your fixes!(it'll restore itself..)


As said, use copies of the files-put the originals in a safe place. OR if you must keep them where they are, rename them.