Lakes

[Russell Adams]

Has anyone had a good experience of trying to model shallow lakes with SHETRAN? I have two in my catchment that are 1.5m and 5m deep (approximate mean). They both have influent and effluent rivers. When I set the model up I made the channel elements where the lakes are situated wider than usual and set the elevation of the elements to be just below the lake bed. The ground surface elevation is the lake bed elevation. At the end of the lakes I set up intermal weirs with the crest height set to the desired lake water level. (I have this from the DEM). I set the initial water level in the channels to be bank-full, i.e so that the water surface is at the lake bed.

The model runs for about 3 months on a daily timestep but only when I start with quite a low phreatic surface, when the water level in the channels reaches the crest it appears to start instabilities. The error message 1030 then fatal error 1060 is generated from the OC component.  I tried setting the surface water depth on the grid elements to fill the "lake" grid but this didn't work very well either. Would relaxing the OC convergence criteria help? Or even using banks? Any help is appreciated as I cannot get past this point.

Russell

[Russell Adams]

Thanks Geoff, I'll try that. I only have daily CEH-GEAR rainfall hence the daily timestep.

I'm actually thinking (off topic) of just running a groundwater model of this catchment using the VSS, I have a recharge input layer that I can use for a steady-state simulation instead of the usual  soils/veg/channels set up. I could then model the lakes as fixed head boundary conditions instead.

I'll keep you posted on that project, there's even the possibility of setting up some conduits as there's karst (similar to Pennine karst) and GSI/BGS(NI) have been doing some spring monitoring and tracer tests...

Russell

Virus-free. www.avast.com

On Tue, 21 Apr 2020 at 09:51, Geoff Parkin <geoff....@newcastle.ac.uk> wrote:

Hi Russell

 

We have been exploring similar issues for catchments in the Lake District, and have just developed a new modified version of Shetran to include lake management, and a paper is actively under preparation now. This has been led by Daryl Hughes, PhD student at Newcastle, who has explored various ways of setting up lakes.

 

In general, I would expect daily timesteps would cause problems here, so I would suggest trying the more convential (for Shetran) hourly timesteps.

 

Best wishes

Geoff

 

From: she...@googlegroups.com [mailto:she...@googlegroups.com] On Behalf Of Russell Adams
Sent: 21 April 2020 09:26
To: Shetran <she...@googlegroups.com>
Subject: Lakes

 

⚠ External sender. Take care when opening links or attachments. Do not provide your login details.

Has anyone had a good experience of trying to model shallow lakes with SHETRAN? I have two in my catchment that are 1.5m and 5m deep (approximate mean). They both have influent and effluent rivers. When I set the model up I made the channel elements where the lakes are situated wider than usual and set the elevation of the elements to be just below the lake bed. The ground surface elevation is the lake bed elevation. At the end of the lakes I set up intermal weirs with the crest height set to the desired lake water level. (I have this from the DEM). I set the initial water level in the channels to be bank-full, i.e so that the water surface is at the lake bed.

 

The model runs for about 3 months on a daily timestep but only when I start with quite a low phreatic surface, when the water level in the channels reaches the crest it appears to start instabilities. The error message 1030 then fatal error 1060 is generated from the OC component.  I tried setting the surface water depth on the grid elements to fill the "lake" grid but this didn't work very well either. Would relaxing the OC convergence criteria help? Or even using banks? Any help is appreciated as I cannot get past this point.

 

Russell

--
You received this message because you are subscribed to the Google Groups "Shetran" group.
To unsubscribe from this group and stop receiving emails from it, send an email to shetran+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/shetran/ad908be3-e9a0-4870-ac34-7f3b01a82c5d%40googlegroups.com.

Virus-free. www.avast.com

[Steve Birkinshaw]

Hi Russell,

It sounds like you are trying the right things. The error messages you get are fairly normal when you try to run with lakes. There are three possibilities I have done

1) In the Easy Setup you can have a lake mask. What this does is set the Strickler Coefficient for river channels near the lake to a value of 3 instead of the default value of 20 for rivers. This seems to work well in attenuating the flow in a similar manner to a lake. You can obviously manually set the Strickler values in the ocd file.

2) Turn the 1060 error to a warning (I can send you an executable). This obviously needs to be done with care but if it is just to get the simulation past a particular difficult timestep then it can work OK.

3) reduce the timesteps (it will obviously take longer to run). Either reduce the basic timestep or the amount of rainfall allowed in timestep and the recovery back to the standard timestep. In the frd file in the standard version the 1,2 and 4th values have been set to reduce the 1060 problem

:FR20 - BASIC TIMESTEP DATA                                                    
 0.2000 0.0100 9999.0 0.5000

If none of this works I am happy to look at the dataset and see if I can sort it out.

Regards,

Steve    T