Lake Augmentation Problem

SUMMARY: Home owners from a beautiful lake community faced a water problem: their lake was ‘sick’ – it was losing water and lake levels were getting lower. An expert "fixed" the problem by pumping deep groundwater directly into the lake. But years layer, the same problems returned. A second expert says the plan never really worked in the first place. Determine if the lake augmentation project was doomed from the beginning. 

The lake, augmentation well, and other salient features of the site.

Background & Problems

In 2005, home owners from a beautiful lake community faced a water problem. Their lake was ‘sick’ – it was losing water; lake levels were getting lower. This was impacting the value of their homes, especially for those on the lake front. One lake front owner whose home value was particularly impacted had an idea: maybe we can pump water from underground and add it to the lake.

The community felt it was a good idea, so they hired a consultant to explore its feasibility. The consultant told the home owners association that – yes – it is possible, but it will require that the water is pumped from a source that is ‘external’ to the lake. In other words, he said, if you pump groundwater, you want to make sure you go deep enough, and hopefully you are getting water from a confined aquifer that is not hydraulically connected to lake.

The consultant suggested drilling a deep test well and pumping really hard to see what happens to the water levels in the shallow aquifer surrounding the lake. He claimed, if the test results showed that the shallow wells around the lake do not respond to the pumping from this deep test well, it means the deep aquifer is disconnected or separated from the shallow aquifer and the lake . The home owners thought this made perfect sense and agreed to give it a try. The consultant installed a test well, penetrating 450 ft  deep into the ground and equipped with a high-capacity pump. He completed the test in collaboration with the home owners with water wells.

Here is the gist of the results: despite pumping 1000 GPM continuously for two weeks, there was no measurable responses observed in any of the shallow wells. In fact, the responses were very small and quickly stabilized even in the existing wells in the deep aquifer. So, the consultant concluded that the results were better than expected! He thought he had found a very good external source of water that was separate from all the various existing wells and lakes, and that the deep aquifer is not connected to the shallow aquifer. He recommended moving forward with the lake augmentation project.

The community was very happy to hear this! They decided to immediately move the proves forward and quickly build the necessary infrastructure. After a little time had passed, the community was pleased with the results: the lake responded to continuous augmentation pumping, although sometimes the lake behaved in a way they did not expect (occasional ups and downs). So the community declared the project a success.

But the success did not last. A few years later, the lake got sick again, especially during the time following the drought of 2012. The community waited another two years, and the lake seemed to remain sick. It even seemed to take longer to respond to the augmentation pumping.  So they hired a different expert. He gave the community a much different response after reviewing all of the information.

The second expert said that, "Although many of the borehole records of deep water wells surrounding the lake showed a confining layer of clay separating the shallow sandy aquifer and the deep aquifer, a few of the deep wells around the northwest portion of the Lake did not show a presence of any clay at depth. There is, therefore, a strong possibility a small 'break' or 'gap' in the clay layer, meaning the deep aquifer may actually be somewhat hydraulically connected to the shallow aquifer and lake. If that is the case, the project might not really have worked in the first place; it appeared to have worked because the years immediately following the installation and implementation coincided with a few wet years. It was all controlled by mother nature."

Objectives & Deliverable

You have been asked to evaluate the arguments made by each expert and determine what is really going on. You will develop a 3D groundwater model to assess the sources of water to the deep augmentation well that pumps water into the lake, and discuss the implications on lake level control.

In your analysis, you should specifically address / evaluate:

• Do the aquifer test data support a hydraulically disconnected deep aquifer, as suggested by the first expert? Why?
• What are the implications of a possible break in the clay layer? Would you expect the aquifer level data to look different?
• What are the groundwater flow patterns during natural and "managed" conditions (no pumping and with pumping, respectively)?
• What is the source of water to the deep augmentation well? Is the lake providing any water? If yes, how much (as a percentage of the pumped water)?

Prepare a 1-2 page formal report that summarizes your approach and findings. Include any detailed model results / graphics in support of your conclusions in an appendix. 

Given Information

Site Geology

Available borehole lithologic records from nearby and along the lake (see above) suggest that a thick layer of confining materials (clay, silts, etc.) separate sandy surficial deposits from deeper, highly transmissive aquifer materials. This clay layer is expected to extend toward the deeply-incised (i.e., well-connected) River and Stream, where this is an abrupt break below the surface water. However, it is not clear from the available data if a similar break exists below / near the Lake (see below).

Conceptual cross-section of the groundwater-surface water system.

Field Data

The following information/data are available from a preliminary study:

• Average land surface elevation:                                               …208m
• Average long-term recharge ε:                                               …15 in./yr.

• Present-day water level in the Lake:                                       …206m 
• Average depth of the Lake  …6m
• Lake leakance … … 5d^-1

• Present-day water level in the River:                                      …205m
• Average depth of the River: …2m
• River leakance: …100 m/d

• Present-day water level in the Stream:                                   …205m
• Average depth of the Stream: …1m
• Stream Leakance: …50 m/d

• Average thickness of sandy surficial deposits                        …b_sand = 90m
• Hydraulic conductivity of sandy surficial deposits                …K_sand =125 m/d
• Vertical anisotropy of sandy surificial deposits ...(K_x/K_z)_sand=2

• Hydraulic conductivity of the River channel deposits           …K_channel = 150 m/d
• Vertical anisotropy of sandy surificial deposits ...(K_x/K_z)_channel=2

• Average hydraulic conductivity of clay layer                        …K_clay = 1 m/d
• Average thickness of clay layer                                              …b_clay = 15m
• Vertical anisotropy of clay layer ...(K_x/K_z)_clay=20
• Average thickness of clay layer                                              …b_clay = 15m

• Hydraulic conductivity of the break in the clay                     …K_break = 125 m/d
•  Vertical anisotropy of sandy surificial deposits ...(K_x/K_z)_break=2

• Average hydraulic conductivity of the deep aquifer              …K_deep =200 m/d
• Average thickness of the deep aquifer              …b_deep =40m
• Vertical anisotropy of deep aquifer ...(K_x/K_z)_deep=2

• Proposed well withdrawal rate                                               …Q = 1000 GPM
  

MAGNET/Modeling Hints:

• Use ‘Synthetic mode’ in MAGNET to create a model domain with the same dimensions as shown in the map
• Overlay the provided SiteMap image file included in the problem description. Choose ‘Use Domain Extent’ to fit the image to the established domain size. 
• You may represent the River, Stream, and Lake as two-way head-dependent boundary conditions.
• You can assume the shape / extent of the River channel deposits is the same in the first layer (sandy shallow aquifer) and second layer (clay layer). 
• Note that the River channel deposits extend down through the clay layer, so the zone feature should exist in both the first and 2^nd layer
• Use zones features to assign hydraulic conductivities to the River channel deposits and the break in the clay layer
• Use a relatively fine grid (e.g., NX=75) for your simulations
  
  
• Add "computational layers" (or sublayers) within each GeoLayer to resolve 3D head variability (impacts of anisotropy ratio)
  • You can add layers by going to: 'Conceptual Model Tools' > 'DomainAttr' > 'Simulation Settings' tab > 'Number of Sublayers' under Grid & Layer Settings. 
  • First, simulate the model without using computational in any of your 3 Geolayers. This creates the initial water table.
  • Then, check the boxes next to 'Number of SubLayers' and 'Water Table as Top' and re-simulate. The shallow aquifer layer and deep confined aquifer layer will use three (3) sublayers each. The clay layer can be assigned two (2) sublayers.
    
     
• Perform a water balance analysis before the pumping well is turned on to determine the baseflow to the stream under "natural conditions". Then turn the well on and perform a water balance under managed conditions.
  • To simplify your analysis, assume that the Lake water level does not change quickly during management, i.e., evaluate the water balance right when pumping begins, but the lake level has not changed yet.