Jump to content
GMS, SMS, and WMS User Forum
Sign in to follow this  
LudekStr

GSSHA - G&A multilayer issues

Recommended Posts

Recently I've been experiencing issues trying to implement more soil layers into my model. The model which worked fine with classic G&A redistribution infiltration routine suddenly produces weird results with G&A multilayer routine.

The only change I made was switching to multilayer routine and assigning the depth of all layers. No matter what the depth was, disregarding any less permeable second layer I tried to introduce - the model always produces the same hydrograph with the peak as much as 20x higher then before and much faster response to all precipitation pulses. Moreover, when I tried to assign less permeable parameters to all layers, the peaks reached astronomic values, the model responded with NEGATIVE infiltration and the discharge volume reached about 200% of precipitation total.

Does anyone has similar experience and/or suggestions, what went wrong? I suspect there's a numeric calculation bug inside :-)

Thx

Ludek S.

CTU Prague

Share this post


Link to post
Share on other sites

Ludek,

I agree, this sounds like a bug in GSSHA. I would suggest contacting the model developers to report the bug. You can find their contact info on the GSSHA wiki page: www.gsshawiki.com

Thanks,

Clark

Share this post


Link to post
Share on other sites

Ludek,

I'd make sure you're using the most recent version of WMS 9.1. There was a bug we fixed in WMS last May that could have caused problems running the Multi Layer Green Ampt (MLGA) method in GSSHA. You can view more information about this fix here:

http://xmswiki.com/xms/WMS:Bugfixes_WMS#Intermediate_Release_9.1.6_.E2.80.93_July_18.2C_2013

I'd recommend looking at your mapping tables to make sure you've defined the same parameters in the same places in the MLGA method vs. the Green Ampt method. These parameters must be defined for all the soil layers (there are 3 layers). In all of our tests, we're getting basically the same hydrograph when we run the MLGA method and compare it to the regular GA method results.

One other thing I would recommend: make sure your Residual Saturation < Wilting Point < Field Capacity < Porosity for all your layers. If this is not the case, you'll have errors or your model won't run. Make sure your time step is low enough, maybe on the order of 5-10 seconds. Initial moisture has a big impact on your model. This value needs to be higher than the residual saturation but lower than the field capacity. I’d recommend different values for each soil type, maybe values near the wilting point if you have a pretty dry area.

Share this post


Link to post
Share on other sites

Thanks for your reply!

I installed a brand new version ow WMS only last week, when I found out that there's been a fix of multi-layer routine in GSSHA.

What I did is that I took the old project using G&A with redistribution and tested it on the new WMS release. It worked perfectly fine as the old one did. Then I switched in the Job Control to G&A multilayer routine. The mapping table dissapeared, so I refilled it with the same parameters (there were some changes in order of parameters, but I reflected it). I assigned the same parameters for every one of all three layers, including the same depth of the layers 100 cm.

?? It just occured to me - is the parameter "depth" supposed to mean "thickness" of the layer or is it the depth of the interface where the layer starts/ends?

As you suggested I checked the parameters whether they fulfill resid < wilting Pt < FC < porosity and I had it right. Only the initial moisture was higher then the FC (still lower than porosity), as I need to reflect the antecendent rainfall.

For this setup the model responded with very sharp and high peaks, they don't seem to correspond to peaks in the hyetograph and the model doesn't have any troubles to produce even higher volume of discgarge than was the precipitation total either. The output is very sensitive to the timestep, I tried several options between 2 min and 15 seconds, but the character of output is always the same - up to 100x higher and sharp peak discharges comparing to G&A redistrib. routine.

Ludek S.

*************************************************************************************************************************************

In case you'd like to see the mapping table:

ID DESCRIPTION1 DESCRIPTION2 HYDR_COND CAPIL_HEAD POROSITY PORE_INDEX RESID_SAT FIELD_CAPACITY WILTING_PT
1 8, sand 7, arrable land 21.282600 4.950000 0.437000 0.694000 0.020000 0.090000 0.033000
2 8, sand 4, gardens 21.282600 4.950000 0.437000 0.694000 0.020000 0.090000 0.033000
3 6, sandy loam 8, impervious 0.000010 11.010000 0.453000 0.378000 0.041000 0.207000 0.095000
4 6, sandy loam 4, gardens 3.359100 11.010000 0.453000 0.378000 0.041000 0.207000 0.095000
5 8, sand 3, meadow 21.282600 4.950000 0.437000 0.694000 0.020000 0.090000 0.033000
6 6, sandy loam 7, arrable land 3.359100 11.010000 0.453000 0.378000 0.041000 0.207000 0.095000
7 8, sand 8, impervious 0.000010 4.950000 0.437000 0.694000 0.020000 0.090000 0.033000
8 6, sandy loam 1, forests 3.359100 11.010000 0.453000 0.378000 0.041000 0.207000 0.095000
9 5, loam 4, gardens 1.120000 8.890000 0.463000 0.252000 0.027000 0.270000 0.117000
10 5, loam 8, impervious 0.000010 8.890000 0.463000 0.252000 0.027000 0.270000 0.117000
11 6, sandy loam 3, meadow 3.359100 11.010000 0.453000 0.378000 0.041000 0.207000 0.095000
12 7, loamy sand 7, arrable land 6.280000 6.130000 0.437000 0.553000 0.035000 0.125000 0.055000
13 5, loam 7, arrable land 1.120000 8.890000 0.463000 0.252000 0.027000 0.270000 0.117000
14 5, loam 2, bushes 1.120000 8.890000 0.463000 0.252000 0.027000 0.270000 0.117000
15 5, loam 3, meadow 1.120000 8.890000 0.463000 0.252000 0.027000 0.270000 0.117000
16 6, sandy loam 2, bushes 3.359100 11.010000 0.453000 0.378000 0.041000 0.207000 0.095000
17 7, loamy sand 3, meadow 6.280000 6.130000 0.437000 0.553000 0.035000 0.125000 0.055000
18 1, clay 3, meadow 0.060000 31.630000 0.475000 0.165000 0.090000 0.396000 0.272000
19 1, clay 4, gardens 0.060000 31.630000 0.475000 0.165000 0.090000 0.396000 0.272000
20 8, sand 1, forests 21.282600 4.950000 0.437000 0.694000 0.020000 0.090000 0.033000
21 1, clay 7, arrable land 0.060000 31.630000 0.475000 0.165000 0.090000 0.396000 0.272000
22 5, loam 1, forests 1.120000 8.890000 0.463000 0.252000 0.027000 0.270000 0.117000
23 7, loamy sand 1, forests 6.280000 6.130000 0.437000 0.553000 0.035000 0.125000 0.055000
24 3, silty clay loam 7, arrable land 0.200000 27.300000 0.471000 0.177000 0.040000 0.366000 0.208000
25 1, clay 1, forests 0.060000 31.630000 0.475000 0.165000 0.090000 0.396000 0.272000
26 1, clay 6, water bodies 0.000010 31.630000 0.475000 0.165000 0.090000 0.396000 0.272000
27 6, sandy loam 6, water bodies 0.000010 11.010000 0.453000 0.378000 0.041000 0.207000 0.095000
28 3, silty clay loam 1, forests 0.200000 27.300000 0.471000 0.177000 0.040000 0.366000 0.208000
29 3, silty clay loam 3, meadow 0.200000 27.300000 0.471000 0.177000 0.040000 0.366000 0.208000
30 7, loamy sand 8, impervious 0.000010 6.130000 0.437000 0.553000 0.035000 0.125000 0.055000
31 7, loamy sand 4, gardens 6.280000 6.130000 0.437000 0.553000 0.035000 0.125000 0.055000
32 3, silty clay loam 2, bushes 0.200000 27.300000 0.471000 0.177000 0.040000 0.366000 0.208000
33 1, clay 8, impervious 0.000010 31.630000 0.475000 0.165000 0.090000 0.396000 0.272000
34 3, silty clay loam 4, gardens 0.200000 27.300000 0.471000 0.177000 0.040000 0.366000 0.208000
35 4, sandy clay loam 1, forests 0.300000 21.850000 0.398000 0.319000 0.068000 0.255000 0.148000
36 4, sandy clay loam 7, arrable land 0.300000 21.850000 0.398000 0.319000 0.068000 0.255000 0.148000
37 4, sandy clay loam 3, meadow 0.300000 21.850000 0.398000 0.319000 0.068000 0.255000 0.148000
38 3, silty clay loam 8, impervious 0.000010 27.300000 0.471000 0.177000 0.040000 0.366000 0.208000
39 8, sand 2, bushes 21.282600 4.950000 0.437000 0.694000 0.020000 0.090000 0.033000
40 1, clay 5, marches 0.000010 31.630000 0.475000 0.165000 0.090000 0.396000 0.272000
41 1, clay 2, bushes 0.060000 31.630000 0.475000 0.165000 0.090000 0.396000 0.272000
42 4, sandy clay loam 2, bushes 0.300000 21.850000 0.398000 0.319000 0.068000 0.255000 0.148000
43 4, sandy clay loam 4, gardens 0.300000 21.850000 0.398000 0.319000 0.068000 0.255000 0.148000
44 4, sandy clay loam 8, impervious 0.000010 21.850000 0.398000 0.319000 0.068000 0.255000 0.148000
45 7, loamy sand 2, bushes 6.280000 6.130000 0.437000 0.553000 0.035000 0.125000 0.055000

*****************************************************************************************************************************************

Ludek,

I'd make sure you're using the most recent version of WMS 9.1. There was a bug we fixed in WMS last May that could have caused problems running the Multi Layer Green Ampt (MLGA) method in GSSHA. You can view more information about this fix here:

http://xmswiki.com/xms/WMS:Bugfixes_WMS#Intermediate_Release_9.1.6_.E2.80.93_July_18.2C_2013

I'd recommend looking at your mapping tables to make sure you've defined the same parameters in the same places in the MLGA method vs. the Green Ampt method. These parameters must be defined for all the soil layers (there are 3 layers). In all of our tests, we're getting basically the same hydrograph when we run the MLGA method and compare it to the regular GA method results.

One other thing I would recommend: make sure your Residual Saturation < Wilting Point < Field Capacity < Porosity for all your layers. If this is not the case, you'll have errors or your model won't run. Make sure your time step is low enough, maybe on the order of 5-10 seconds. Initial moisture has a big impact on your model. This value needs to be higher than the residual saturation but lower than the field capacity. I’d recommend different values for each soil type, maybe values near the wilting point if you have a pretty dry area.

Share this post


Link to post
Share on other sites

Ludek,

Regarding your question above, the depth is the thickness of the layer. I'd just make sure you have defined the correct parameters for all 3 layers in your mapping table. Check your .cmt file to make sure non-zero values are defined for all your soil ID combinations. Your mapping table values look fine at first glance. You could try sending your files to the GSSHA developers (see gsshawiki.com for contact information) and see what they say. Aquaveo also has various levels of support (for simple model issues) available if you need it. You could send your files to Aquaveo tech support and see what they say. I'd strongly recommend going through your mapping table parameters first and make sure everything is set up right.

Let me know if I can help with anything else,

Chris

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
Sign in to follow this  

×