null

Message Boards

"Wash-out" of initial conditions for certain constituents

Martin Schueder, modified 4 Years ago.

"Wash-out" of initial conditions for certain constituents

Padawan Posts: 52 Join Date: 10/8/13 Recent Posts
Hello!

In my simulation I have created a .q3d initial conditions map for TIC, ALKA, DO, ULVA, and Temp. My simulation shows the following after 15 minutes (a single output file interval):

- DO and TIC have dispersed throughout the grid (5.34 hectares in size) and become relatively uniform
- ALKA has retained most of its spatial heterogeneity as expected

For example: initial conditions for DO range between 7 and 15 mg/L over my grid, but after the first 15 minutes the range reduces to between 12 and 14 and quickly diminishes to a range of 10.219-10.211 after 24 hours. The same can be said for TIC. Alkalinity on the other hand retains a range of 145-185 mg/L after 24 hours which is similar to its initial condition

The only explanation I could think of is that ALKA is dependent on non-volatile components where TIC and DO are dependent on soluble oxygen and carbon dioxide concentrations. DO and TIC might reach uniform values after the first 15 minutes because of a quick equilibrium exchange with the air.

To try and address this issue I made the re-aeration constants of O2 and CO2 extremely low to see if this would fix the problem. The initial condition "wash-out" still occurred.

Any thoughts as to why not all constituents disperse in the same way?

Your comments are most appreciated,

Martin
Christophe Thiange, modified 4 Years ago.

RE: "Wash-out" of initial conditions for certain constituents

Jedi Knight Posts: 125 Join Date: 11/15/12 Recent Posts
Hi Martin,

Does your ALKA pattern match your fixed ULVA pattern? If so that might be part of the answer.

You could print out the balance for a segment where ALKA remains relatively constant and DO/TIC change significantly. This will tell you if the changes are process or transport driven.
Rudy Schueder, modified 4 Years ago.

RE: "Wash-out" of initial conditions for certain constituents

Padawan Posts: 52 Join Date: 10/8/13 Recent Posts
Hi Christophe,

Sorry it took so long to figure this out on my end. Firstly, my ULVA currently do not match my ALKA pattern.

From my print out it appears that changes in TIC, DO, and ALKA are transport driven. Here is an example at one of my observation stations:

Observation Point S7 Layer 3/5 (bottom left end of model area)
taken after 5 mins (1 .hda file write)
No loads at layer 3 or at any layers near S7

DO
Processes 0.017568 g/m2
Transport in 54.06 g/m2
Transport out 55.62 g/m2

TIC
Processes 0.00804 g/m2
Transport in 34.09 g/m2
Transport out 33.07 g/m2

ALKA
Processes 0.000247 g/m2
Transport in 38.56 g/m2
Transport out 38.9 g/m2

Nothing seems to be extremely different between the three parameters (transport dominates all three) so I am not quite sure where to go from here. I have attached an image to better display the problem. Please note that there is a man-made baffle separating the two halves of the pond. Any suggestions on something to check next?

Thank you again,

Rudy
Christophe Thiange, modified 4 Years ago.

RE: "Wash-out" of initial conditions for certain constituents

Jedi Knight Posts: 125 Join Date: 11/15/12 Recent Posts
Hi Rudy,

Thanks for the pictures, it helps a lot.

From your results at station S7 we can indeed see that, for that location, changes are mainly transport driven.

Since we want to understand why, in each of the basins, TIC and DO seem to mix faster than ALKA, it would be better to generate a balance for larger regions, covering the entire domain together. I would define two monitoring areas per basin. One in the SW half and one in the NE half (of each basin). All areas including all layers if your model is 3D. So basically we will end up with four "monitoring blocks".
Generating a balance output without lumped transport will then show the exchanges between each of the four blocks and what was produced/consumed within those blocks.

You can use Dido to make an *.obs defining the four observation areas. Make sure they do not overlap but that they cover the entire domain together. The *.obs file will only contain segment numbers of the top layer. Once imported in the GUI, you can click the observation areas and change their Layer setting to "Average over depth", the GUI will then generate the additional numbers for segments in lower layers in the *.inp file.

Let us know if something is not clear.
Rudy Schueder, modified 4 Years ago.

RE: "Wash-out" of initial conditions for certain constituents

Padawan Posts: 52 Join Date: 10/8/13 Recent Posts
Hi Christophe,

Thanks for the reply. It will take a little time for me to try this out but I will get right to it and report back.

Thank you for the insight,

Rudy
Rudy Schueder, modified 4 Years ago.

RE: "Wash-out" of initial conditions for certain constituents

Padawan Posts: 52 Join Date: 10/8/13 Recent Posts
Hi Christophe,

I have done as you recommended, and as per the attachment showing where the quadrants are in my pond, here is what my results show at 09:05:00 (5 mins after initialization):


OXY TIC ALKA
Proc In Out Proc In Out Proc In Out
SW -306 19686 26051 33.7 6223.9 6177 -0.297 21522 21617
NW -211 43278 36587 1.93 5516.5 5868 -0.296 24221 24268
SE -401 19003 29795 43.52 6186.8 6273 -0.358 21718 21874
NE -257 43613 39436 11.8 6244.5 5094 -0.386 24760 24736

SW and NW quadrants will have large exchanges with in another and SE and NE quadrants will have large exchanges with one another. SW and SE will exchange a small amount of material through a 2 m opening in the pond baffle. NW and NE quadrants will not exchange material.

What sorts of clues should I be looking for to identify the source of this problem? It looks like there is a large mass transfer of OXY from SW to NW and SE to NE, which makes sense because concentrations are initially higher in these regions. It also makes sense that there is a transfer of TIC from NW to SW as SW and SE have lower TIC than NW. The net transport (In-Out) represents 2.4%, 1.6%, and 0.3% of the total transport in the system for OXY, TIC, and ALKA. It perhaps can be said that based on these numbers there is a larger net transport of OXY and TIC between quadrants than ALKA.

Any other ideas about how to further deduce the problem?

Thank you,

Rudy
Christophe Thiange, modified 4 Years ago.

RE: "Wash-out" of initial conditions for certain constituents

Jedi Knight Posts: 125 Join Date: 11/15/12 Recent Posts
Hi Rudy,

That is exactly what I had in mind.

Advection and dispersion are the same for all active substances (unless you make use of additional dispersion arrays, but that isn't your case), so what I am trying to prove out of the balance numbers is that the pattern differences must be process and/or load related.

Could you attach the *.prn file you got from the mass balance?

Thanks,
Christophe
Rudy Schueder, modified 4 Years ago.

RE: "Wash-out" of initial conditions for certain constituents

Padawan Posts: 52 Join Date: 10/8/13 Recent Posts
Hi Christophe,

I believe I have attached the appropriate file with the model at 1% completion (~7 hrs in). Let me know if I can provide you with any further information.

Thank you,

Rudy
Christophe Thiange, modified 4 Years ago.

RE: "Wash-out" of initial conditions for certain constituents (Answer)

Jedi Knight Posts: 125 Join Date: 11/15/12 Recent Posts
The differences were due to an error in the process library definition causing HDisperVel not to affect the dispersion of Alka. This has been fixed in revision 4910.

Thanks Rudy for pointing that out.