I am modelling groundwater sustainability in a 1885.14 km2 watershed. The annual recharge in the area was estimated to be 55 mm, while the current extraction rate is 484,000m3/day. I tried to estimate the safe yield of the aquifer system. I used MODFLOW to accomplish my work. I run it for 30 years simulation period. And, my model result shows that the aquifer will be safe with a maximum pumping rate of up to 530,000 m3/day.
Now, this confused me. Because, the only recharge amount I used in my modelling work was 55 mm, which is 103,682,700 m3/year in whole area, while the annual discharge is 176,660,000 m3/year.
So, my question here is:
How could it be possible for the aquifer to be safe (almost no draw down) after pumping it with up to 530,000 m3/day, which is 193,450,000 m3/year? Recharge is greater than discharge in this case. So, how could it be possible? Any suggestion please!
Thanks in advance for your suggestions and comments!
Cheers,
H
This may be a stupid question, but are there any other BC's that may be injecting water into your system? You're taking into account surface recharge as your only source of water, but if you have any Dirichlet or Cauchy BC's, those could make the difference, by either adding or substracting water from your aquifer.
Stupid question? Really?
I don't have any other boundary condition my friend (no flow boundary in all sides). It's a blocked system. I don't have any specified head boundaries as well. Go back and read my question again. I think you didn't get what it means!
I meant my question might be stupid (as it was), as you didn't mention the side BC's. I didn't mean your question was stupid in any way (it isn't!).
Seeing as the only source of water on your model is recharge, the only other possible "source" MODFLOW could be considering is the Specific Storage or Specific Yield you have to set in a transient model (depending on which type of aquifer). If those are really high the model might be missrepresenting drawdowns.
I can't think of any other way you could get low drawdowns over such a large amount of time with the pumping being greater than the recharge, other than convergence problems causing the problem to be non-conservative (which I assume there are none).
I know what I'm saying is basic stuff, but it's really weird that you can pump out that much more water than the recharge without big drawdowns. Hope it helps!
Well,
What the water balance, means What about evapotranspiration, If this were the case, the groundwater would flow up in springs and form water bodies. Is this what happens in the drainage basin?
Sincerely
I think I have to give some details about the system. The aquifer is a heterogeneous anisotropic unconfined unit. It’s in a valley surrounded by a basement outcrop (basalt) in all sides. So, I used a no flow boundary in all sides. There is no specific head boundary in the system. No evaporation from water table, for the water table is deep (averagely 26 m from the groun surface. The average thickness of the aquifer is 170 m.
I agree with you that the model output is not reasonable. now comes the fun part of figuring out why. One way modelers go about this when they think a certain parameter is responsible is to change that parameter until they get a reasonable result. In your case if you think the recharge is the problem, it would be useful to understand the potential error in the estimation. change it and see the effect on the model output. you could reduce it until you get a reasonable result. Is the 'new' recharge value reasonable? is it within the error of the recharge estimation? what happens when recharge = 0? since the water table is so deep is it possible that there are perched water zones where most of the recharge is held up so that it never reaches the aquifer?
It is always good to use field data to validate your model in the form of water levels in the pumping well and other wells in the area. support from geological data is essential for setting up the model. I know you said the aquifer is heterogeneous but how is the aquifer portrayed in the model? are there geologic layers in the aquifer with different K? are there shallow low K zones that could form a perched water table? Is the basement rock really a no flow boundary? a pure closed reservoir is unusual- typically it leaks somewhat. do you have pumping well data? what does the pumping well derivative plot look like? what other data support a no flow boundary? is there any information concerning the degree of weathering of the basalt? fracturing? Troubleshooting a model is tough, but data is a very useful guide into understanding the flow system.
I am talking strictly about the numerical model. The reason is storage!! Remember the first lesson of the Theis model. Whenever extraction takes place the water comes first from the storage and when the storage water is depleted then the cone of depression starts increasing till the time the rate of extraction equals the water coming inside the cone of depression. Run your model as steady state and I am damn sure you will get the right result.