Raven

Hello,

uhm...this might be a stupid question. But if I want to run Raven in an hourly time step. What do I tell the :TimeStep command? Does Raven understand 1/24 or 0.04?

Thx!

01:00:00 (read the manual)

Hi Symona

Either format actually works, either 01:00:00 or as a double (a fraction of 1 day, i.e. 0.0416667). For a 1 hour timestep, 01:00:00 is probably better since it avoids having to put in many repeating decimal places :)

Cheers,
Rob

Would you please give an example for grid set up

Hi Elizabeth

I'm afraid I don't have an example on hand. However, it will likely involve some GIS processing with something like a fishnet function (ArcMap setup or QGIS setup) to get coordinates of grid cells within a given watershed. Once you have the coordinates of centroids of each grid (an output of the fishnet function), you can basically setup the rvh file for each centroid an tie an HRU to each subbasin. The rest is normal Raven setup.

Hope that helps!
Rob

ArcMap Fishnet function:
http://desktop.arcgis.com/en/arcmap/10.3/tools/data-management-toolbox/create-fishnet.htm

QGIS equivalent functions:
https://gis.stackexchange.com/questions/123444/creating-fishnet-grid-shapefile-in-qgis

Hello,
a question regarding the VIC Soil Evaporation Algorithm:

When it is used without using other VIC Algorithms (e.g. the VIC Infiltration) it is always asking for the VIC_ALPHA, VIC_ZMAX and VIC_ZMIN soil properties. But are they really necessary when the VIC Infiltration is not used?

Thx!

Hello Symona,
Thanks for the comment.
Looking at the source code, it does state that the three terms are not being used in the equation (although it requires it). There is a flag in the code telling the devs to look at it.
I think you can put whatever number and it would not have an effect the process (given you aren't using other VIC processes). Let me know if having random values of the three terms messes things up.

Regards,
Konhee

rchlumsk wrote:Hi Symona

Either format actually works, either 01:00:00 or as a double (a fraction of 1 day, i.e. 0.0416667). For a 1 hour timestep, 01:00:00 is probably better since it avoids having to put in many repeating decimal places :)

Cheers,
Rob

Follow-up question: How about the Time Series Input Files? The manual states the time interval has to be specified as a double. So for hourly data I'd have to put 0.0416666666666667 and the values itself still need to be in mm/d. Is that correct? Because right now Raven is giving me the following error with this setup:
======================================================
Initializing Model...
HWB 7358 -1.83354e-10
===============Exiting Gracefully==============
Exiting Gracefully: CGauge::Initialize: negative or excessively large (>10000mm/d) precipitation intensity reported at gauge
Error Type: Bad input data
===============================================

However, when looking at my time series I can't find any negative (or excessively large) values.

When I put 0.04 as the time interval the simulations run fine. But my day would have 25 hours now. So are their any experiences or suggestions how to best run Raven in a daily mode?

Thx!

Hey Symona,

My guess might be is because 0.0416666666666667 don't perfectly go into 24, there is some sort of rounding error.
Like you mentioned, using a fraction like 1/24 and it will give you an exact double.
But like Rob said, the standard is to just use 01:00:00 for hourly timesteps.

Regards,
Konhee

Hey Konhee,

I was guessing it must be a rounding issue but I didn't realize a fraction would actually work. And the manual states that the format 01:00:00 should not be used in the RVT files. When trying it anyways it also gave me the above mentioned error. Thanks for the clerification. It works fine now! :)