General Questions in Raven

[ephilip]

Hi Elizabeth

With the flowdurcurve.plot function it uses the hydrograph object from hyd.extract function, so it looks for the $sim and $obs objects within the hydrograph argument. If you use the hyd.extract function and pass that on to the flowdurcurve.plot function, it should work.

For the other errors, I am unable to reproduce them. Could you send me an example of the code breaking?

Thanks,
Rob

Hi Rob,
Let me clarifying a bit more.
Yes hyd.extract function was called first for Flow Duration Curve. Yet getting error.

For other errors, in the previous version, if the observed flow data is not complete, it was giving warning. Now with the newer version, it is giving the error (if not the observed data is complete). This is a typical situation when calibration and validation data are complete and for the application period the data is incomplete.

Please help.

Thank you
Elizabeth

Good Morning,

Just a friendly reminder. Did you get a chance to read the question.

Thank you
Elizabeth

[pcampbell]

Hi,

I'm trying to build a model using the UBCWM emulation and have a question about the Hydrologic Processes section of the example input file (Appendix D of the user manual). The example rvi file has a section commented as "# from infiltration to routing", which appears to move water from runoff back into the soil profile, specifically:

# from infiltration to routing
:Flush RAVEN_DEFAULT SURFACE_WATER INT_SOIL2

Is the UBC emulation using the INT_SOIL2 soil layer (alias for SOIL[4]) as a substitute for in-channel routing? I'm not quite sure what this command does and whether it needs to be included. Any explanation would be appreciated.

Thanks
Pete

[jrcraig]

Hi Pete,

In the UBC watershed model, there are intermediate routing stores used for controlling the timing of outflow from the system. These are denoted INT_SOIL2 (fast runoff) and INT_SOIL3 (slow runoff). While they technically are 'soil' stores in the Raven emulation - Raven uses soil storage as a catch all for any bucket-type storage unit in a conceptual model - these should be considered routing stores, where water is delayed before reaching the subbasin outlet. The "Baseflow" and "Percolation" processes controlling flows from these stores are really just linear storage relations where outflow from the routing stores is proportional to storage of water within the stores.

You will want to keep these in the configuration for a strict representation of UBCWM. Alternately, you could represent this time delay using the in-catchment routing options of raven, but this does not enable the independent control of snowmelt vs. rainfall runoff timing that is implicitly supported in the UBCWM model.

Regards,
James

[anya]

Hi,

I am trying to run Raven with UBCWM set up and .tb0 file as an input. The model runs but generates empty outputs (sometimes only headers are present) and no errors. Since the errors.txt file is empty, I was wondering if you maybe know what could be the problem and what can I look into to solve it? I am using 2.9.9 version.

Thank you!

[rchlumsk]

Hi Anya

Can you please try using the version 3.0 (latest version from the website) and then post back with both the Raven_errors.txt and the screenshot from the command prompt/ terminal where Raven was run from? There may be some issues even before the inputs are read in if no errors file is generated.

Thanks,
Rob

[anya]

Hi Rob,

Thank you for your reply! I am actually using the latest version of Raven (version 3.0 on the website), but in command prompt it is reported as 2.9.9. Here is the copy of command prompt window:

. . . . . . . . . . . . . . .

C:\Users\Anna>cd C:\Users\Anna\Documents\Raven_files\BC_Hydro\La_Joie_ws

C:\Users\Anna\Documents\Raven_files\BC_Hydro\La_Joie_ws>Raven.exe .\La_Joie_ws -o .\output\
============================================================
RAVEN
a robust semi-distributed hydrological modelling framework
Copyright 2008-2020, the Raven Development Team
Version 2.9.9 w/ netCDF
BuildDate May 20 2020
============================================================

C:\Users\Anna\Documents\Raven_files\BC_Hydro\La_Joie_ws>
. . . . . . . . . . . . . . .

And Raven_Errors_txt is empty (0 KB) (apologies, I do not know how to attach files here). The model generates the output folder and all user-specified outputs, but they are empty/only have headers.

[jrcraig]

Hi Anna,

I think you are missing the filename here. IF the .rvi file is in the La_Joie_ws directory, and named La_Joie_ws.rvi, then you should be calling

C:\Users\Anna\Documents\Raven_files\BC_Hydro\La_Joie_ws>Raven.exe La_Joie_ws -o .\output\

As is, this is looking for an empty .rvi file in the C:\Users\Anna\Documents\Raven_files\BC_Hydro\La_Joie_ws\La_Joie_ws directory. The empty filename might be the cause for the error.

[anya]

Thank you, James!

Actually, it turned out that for some reason it does not run (i.e., no error message, empty output files) when I use the most current version of Raven.exe, but it worked (no errors, complete outputs) when I switched for older version (the one which is available together with tutorials, Version 2.9.2, BuildDate May 30 2019.

[anya]

Hi!
I was wondering about the available diagnostic algorithms: is it possible to make Raven output Monthly Mean Squared Error (TMVOL) for each month (i.e., 12 values instead of 1)?
Thank you!

[rchlumsk]

Hi Anya

I don't believe this is currently an option in Raven. The TMVOL metric is based on each month in the simulation, this this would produce more than 12 values if the simulation is longer than a year.

If you are looking for an average monthly error for each month in the year, over the simulation period, this would be the type of metric that would require post-processing of Raven output and generating your own diagnostics. There are some utilities in the RavenR package (such as rvn_monthly_vbias) that may be useful in this regard.

Hope that helps,
Rob