We use cookies to make your experience of our website better. You can delete or block cookies, but some parts of this website won't work without them. By using this website you accept our use of cookies.

Brookr

10/05/2016 16:19:35

 Good afternoon.

This is a fairly general question around the use of Theta (Advanced Paramaters).

I am currently working on a third party model that appears to unstable towards the downstream. I have tried various things to rectify the problem but to no avail. I have found that if I increase theta the poor convergance is removed.

It is set to 0.7 as default, however if we increase this value 0.85 or greater that seems to work.

Within the help file it states that a higher value will make the model run more stably, if this is the case why not have it set to 1 all the time? It obviously has some draw back/ limitation.

I was hoping you might be able to explain these limitations.

Regards Richard Brook

Konrad

Flood Modeller Support

10/05/2016 17:52:13

Hi Richard,

There are deifnitely a couple of trade-offs; one is time - it may take more iterations to converge when theta is set to 1. The other, probably more pertinent, is of numerical diffusion, which can be introduced for higher values of theta. 0.7 is generally a good balance between stability, accuracy & time.

There's some discussion on it here from the EA 1D model benchmarking study (Section 5.1):

https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/290893/scho0305bixt-e-e.pdf

Konrad

Sign in to create or reply to forum posts.

Privacy & Terms

Copyright