Diagnostics and Calibration

[rchlumsk]

Hi Elizabeth

The weighting file can handle any positive value there, the weighting of each point will be adjusted accordingly. Using values between 0 and 1 makes the most sense to me. I can't comment on whether 0.95 is the correct value to use for ice cover, although something less than 1 (if 1 is used for all other non-ice cover points) makes sense.

Cheers, Rob

[ephilip]

Hi Rob,
I was trying to run DDSAU with just NS with threshold of -0.7. The model is running and giving me only OstModel0.txt and OSTOutput0.txt. No OstModel0_DDS0.txt (set), OStDDSPn.txt files getting generated. Would you please help.

Thank you
Elizabeth

[rchlumsk]

Hi Elizabeth

I think I need a bit more info to help you on this one. Please check the OstErrors.txt file and let me know what the errors are, and if the issue is causing an error or if it is something else in the run.
Rob

[ephilip]

Good Morning Rob,
The out put from the OStErrors.txt is below:

Ostrich Error Report for Processor 0
A total of 0 errors and/or warnings were reported
NO ERROR : No messages

Info from the OstIn.txt is

ProgramType DDSAU
ObjectiveFunction GCOP
ModelExecutable Ost-RAVEN.bat
...
BeginResponseVars
#name filename keyword line col token
NS_Q .\model\run1_Diagnostics.csv; OST_NULL 1 3 ','
EndResponseVars
...

BeginDDSAU_Alg
PerturbationValue 0.2
NumSearches 3
MinItersPerSearch 40
MaxItersPerSearch 40
ParallelSearches no
Threshold -0.7
Randomize no
ReviseAU no
EndDDSAU_Alg

Thank you
Elizabeth

[ephilip]

Hi Rob,

Sorry for disturbing you again. Previous problem got solved. Now I am getting error

Ostrich Error Report for Processor 0
A total of 1 errors and/or warnings were reported
FILE I/O ERROR : Using default DDS algorithm setup.

Is it because I am only using NS_Q? Please help.

Thank you
Elizabeth

[rchlumsk]

Hi Elizabeth

Glad that you got the first one resolved. The issue should not be the single objective since DDSAU only uses 1 objective (as per Table 1 in Ostrich documentation). Only difference I can see in the syntax is a missing underscore in the "BeginDDSAU_Alg" and "EndDDSAU_Alg", where the documentation has "Begin_DDSAU_Alg" and "End_DDSAU_Alg". Try that and double check the other syntax and let me know if that works.

Rob

[ephilip]

Hi Rob,
Thank you for the reply.

The version I ran after correcting the first issue had the correct syntax. It is giving the same error as before

Ostrich Error Report for Processor 0
A total of 1 errors and/or warnings were reported
FILE I/O ERROR : Using default DDS algorithm setup.

ProgramType DDSAU
ObjectiveFunction GCOP
ModelExecutable Ost-RAVEN.bat
...
Begin_DDSAU_Alg
PerturbationValue 0.2
NumSearches 3
MinItersPerSearch 40
MaxItersPerSearch 40
ParallelSearches no
Threshold -0.7
Randomize no
ReviseAU no
End_DDSAU_Alg

Would you please help.

Thank you
Elizabeth

[rchlumsk]

Hi Elizabeth

I am afraid that I am not sure what the solution is for this error. I recommend that you check through the Ostrich manual in case there are some clues to solving this problem. You can also try to contact Dr. Mattot or Dr. Tolson if they have any suggestions, and let me know if you figure out the issue in the next few days. Otherwise you can try to send me your files and I can try to take a peek at your setup.

Cheers,
Rob

[Konhee Lee]

Hello Elizabeth,

There has been an issue with Ostrich and DDSAU.
https://app.box.com/s/da98sdck9ygkuf2pm4oajwofsb5z3lat
There is a beta version of Ostrich.exe uploaded in the link above (download the .piz file and change the extension to zip).
Compare your OstIn file to the one uploaded in the dropbox as well.
Hope this could solve your problem.

Regards,
Konhee

[ephilip]

Hi Konhee,

Thank you for the link. Even with the beta version, it is giving the same I/O Error. But the OstOutputDDSAU.txt with the behavior solutions generated. Any help is highly appreciated.

Thank you
Elizabeth