LinTerp: Warning Message

[rhainb]

Hi Harc,

I think I have found another bug. When performing a Monte Carlo Simulation I received this warning message; “LinTerp: X input array not sorted in increasing order. This message will not be shown again for this duration.”.

I am not concerned about the message itself at the moment, I will look into that myself. But, what I believe to be a bug is that despite the warning saying that “This message will not be shown again for this duration.” the same message pops up again when I close the message, I believe that it is doing this every time a different run takes place (50-100 times each duration). Eventually with enough clicking I can get through the Monte Carlo simulation.

• This topic was modified 6 years, 9 months ago by rhainb.

[Matt Scorah]

Thanks for raising this. In order to replicate this could I please check the following:
What version of RORB are you using?
Do you have a print node at a location which is receiving zero flow (e.g. if you turn off some print nodes does the error go away?)

• This reply was modified 6 years, 9 months ago by Matt Scorah.

[rhainb]

Hi Matt,

I am using RORB 6.32.

The issue I had was that I didn’t have any print nodes in my catchment file. Putting in a print node has stopped the warning from appearing.

[reevesm]

I am receiving the same LinTerp warning message using version 6.42 with ARR2016 ensemble events with initial losses varying by ARI. The messages seem to appear when the simulation reaches an ARI where the initial loss factor is greater than 1. The batch.out indicates “Error interpolating initial loss factor, ARI factor set to 1.0”. The storm files confirm that the initial loss factor was set to 1 instead of the value specified. This has occurred in two separate models that have been updated to ARR2016 hydrology. Workaround at the moment is to run in separate batches where initial losses remain constant across ARI’s.

[Benson Liu]

Hi Reevesm,

I can’t seem to be able to replicate the bug you are experiencing with my own catchments, would you be able to send me an email with your .catg, .par and other input files as well as a brief description of the other modes/options you may be using in RORB when this bug occurs? My email is benson.liu@dev7.preflight.com.au.

Kind regards,

Benson

[mnogueira]

Hi Benson,

Did you guys figure out what was this problem? I’m running a very simple model with ARR ensembles and I get that warning everytime the simulation reaches the 2h duration.

Cheers,
Mateus

[Benson Liu]

Hi Mateus,

That particular error that reevesm encountered is fixed in the latest release of RORB. Are you getting the same error output in your batch.out file or are you only getting the LinTerp warning message?

Kind regards,

Benson

[Anonymous]

Hi HARC,

I am running RORB v6.45, doing a Monte Carlo Simulation for a reasonably large model. I am receiving the Linterp X warning for every single duration and it does pop up about 50 – 100 times as has been mentioned previously on this thread, despite saying that it won’t pop up again for this duration.

Wondering if there is a workaround for this error.

Thanks so much,
Kylee

[mnogueira]

Hi Benson,

I was only getting the LinInterp warning message. The model was very simple, I re-created it from scratch and couldn’t replicate the error – Working fine now.

Cheers
Mateus

[Benson Liu]

Hi Mateus and Kylee,

Mateus, that is a little odd but if it is working then that’s good.

Kylee, it is a little difficult to diagnose without looking at the files but there are 2 common causes for the Linterp warning.
1) There is something odd with the design rainfalls. The Linterp warning can pop up if the AEPs are mislabelled or rainfall in
the same duration are for some reason extremely similar or decrease with rarer AEP.
2) As Matt said earlier in the thread the warning can occur when a print node has 0 or the same flow for multiple AEPs.

Would you be able to check the design rainfalls and the print nodes?

Kind regards,

Benson

[daveM85]

Hi HARC,

I’m having the same issue as per above. I have checked the IFD csv and other output files but I don’t seem to find the issue impacting my model.
Can you please review the files attached and let me know?

Thanks.
Regards,
David

Attachments:

You must be logged in to view attached files.

[Matt Scorah]

Hi David,

I couldn’t replicate your issue with the files provided. Feel free to email enquiries@dev7.preflight.com.au with more info (e.g. RORB version number, type of run error was encountered, screenshot of error etc.)

Cheers,
Matt

[daveM85]

Thanks Matt.
I think I have found the issue. I have sent an email to enquiries@dev7.preflight.com.au in any case but it looks like that the Montecarlo simulation doesn’t like long path for the input and output directories, whereas I can run the other simulations with no issues.

Considering that linterp is a C++ library with N dimension, would there be a chance that the dimension assigned within the coding has somehow an impact on the length of the path for the input and output directories?

David

[Author]

Posts