Regarding preparing input files for Hydrological and Sediment transport modelling in SHETRAN

[RESMI S R]

Dear all,

I am currently reserching on the contrasting features of 2 sub-catchments (Areas: 10,612 sq.km and 19,000 sq.km respectively) of the same river basin, considering their hydrology and sediment transport capability. Since, I wanted to use a distributed physical based model, my Ph.D. Supervisor advised me to use SHETRAN. For the first sub-catchment, we fixed the grid size as 550 m*550 m. Since my sub-catchments are large, using SHETRAN Windows version is not possible. I have to run both the hrdrological and sediment tranport module for my present study. My queries are as follows:

1) How do I prepare the input files relevant to my modelling, using the SHETRAN Standard Version. Since I am a beginner and the tutorial videos do not show this procedure, I am not getting an idea what to begin with.

2) While going through the user guide, I came across that SHETRAN Standard version has full capability, but without WINDOWS environment. I am currently using WINDOWS Operating System. Is it possible to set up the catchment and Input data in WINDOW OS. Otherwise, what is the alternative? 

Kindly help me with this issue.

Thanks and regards,

Resmi

[Steve Birkinshaw]

Dear Resmi,

1) You need to use the Shetran easy setup for sediment. https://research.ncl.ac.uk/shetran/Shetran%20Easy%20Setup-%20sediment%20and%20snow.pdf. This takes you through an example for the "Aire at Kildwick Bridge".  The example shows what data is needed. This is a two stage process it first creates the Shetran input files and then runs Shetran taking the Shetran input files and producing Shetran output files. To get further with the sediment simulation you will need to manually modify the input_catchmentname_syd.txt file and use the standard version of Shetran. To run the sediment component of Shetran it is usually best to calibrate the water flow component first.

If your question is how do you actually prepare the files (such as the DEM file or land cover file in ASCII format). This is really a GIS problem. The input files using the Easy Setup are all in standard ASCII format apart from the xml file which can be manually edited.

2) The Shetran standard version runs on a Windows operating system it just does not have a Graphical User Interface (GUI)

Due to memory issues there is a limit of 200x200 grid cells in Shetran. If you catchments are 10,612 sq.km and 19,000 sq.km you might exceed these limits (and the runtimes will be very long). It will probably be better to use 1000m vy 100m grid sizes.

Regards,

Steve

[RESMI S R]

Dear Steve, 

As per your suggestion, I prepared all my input files and started running the model with grid size 1km*1km. Inorder to increase the drainage density of my river network, I reduced the number of 'RiverGridSquaresAccumulated' from 2000 to 50 (Initially, I had given the value 2000, but only a single river was formed, still the simulation was running fine). After 86% the simulation encountered 3 warning errors (ERROR NO: 1030) and a fatal error (ERROR NO: 1060). What could be the possible causes and remedies for these? Kindly help me.

Thanks and regards,

Resmi

[Steve Birkinshaw]

Dear Resmi,

It sounds like you are making good progress. As you say the error 1030 is just a warning error. The important error is 1060 which causes the simulation to stop. There are a variety of reasons for this error but generally it is caused at the junction between rivers and the simulation becomes unstable. It often happens during a recession in the discharge. In Shetran the basic timestep is reduced during heavy rainfall and then gradually increase back to the standard 1 hour timestep. If it increases back to 1 hour too soon you get this error. If you are using the easy setup first thing to try is to change this value in the xml file

<IncreasingTimestep>0.05</IncreasingTimestep> speed of increase in timestep after rainfall back to the standard timestep. The standard value is 0.05. If if there are numerical problems with error 1060 it can be reduced to 0.01 but the simulation will take longer.

from 0.05 to 0.01

or reducing the standard timestep

<RegularTimestep>1.0</RegularTimestep> This is the standard Shetran timestep it is autmatically reduced in rain. The standard value is 1 hour. The maximum allowed value is 2 hours

from 1.0 to 0.5

If neither of these work then let me know.

Regards,

Steve

[RESMI S R]

Dear Steve,

Before mailing you, I had gone through the queries in SHETRAN Google group. I had come across a query highlighting similar issue and your reply to it. As per that, I modified both my 'Increasing Timestep' and 'Standard Timestep' to 0.01 and 0.5 respectively. Still, the problem persis. Is there anything else that can be done to rectify this issue? If so, please guide me.

Thanks and Regards, 

Resmi

[Steve Birkinshaw]

 Dear Resmi,

I think the easiest thing is for you to send me all the input files you are using and I will have a look. Can you zip them all up and send them to my email (s.j.bir...@ncl.ac.uk).

Thanks,

Steve

[Felix Op de Hipt]

Dear Resmi,

I remember that I had similiar problems when I started and if I remember well Steve advised me (or I played around, dont remember exactly:) to change the "RiverGridSquaresAccumulated" to a rather large number (in my case 200). If I remember well that changed a bit the details of the river system generation. I am absolutely not sure if it was the same error type but may be you can just try it.

Best regards

Felix