Barron Lake — A Fully-Coupled Lake-Aquifer Model

The site and the monitoring story

Barron Lake sits in a glacial terrain near Niles, in southwestern Michigan's Lower Peninsula. The lake has been monitored since 2006, with stage records through 2013. Since 2013, groundwater levels have been recorded at 11 private wells surrounding the lake — enough spatial coverage to infer boundary-condition dynamics but not so many that local flow gradients can be interpolated with high confidence.

The system has three distinctive features that drive the modeling approach:

• A pumping well near the lake moves water from the deep aquifer into the lake as surface water, augmenting lake levels during drier periods.
• The surrounding geology is characteristically heterogeneous — typical glacial stratigraphy with 4 distinct material types identified from Wellogic borehole data via Transition Probability (TP) geostatistics.
• The private-well monitoring records function as transient boundary conditions — what the model should match at the lake's perimeter, imported via ESRI shapefiles and per-well CSV stage time series.

1. Load domain and set baseline aquifer attributes

The model starts by loading a pre-defined domain polygon via LoadModel using the provided Domain.txt. Domain-wide aquifer attributes are then assigned via DomainAttr:

The aquifer parameters to assign at the domain level:

2. Import a T-PROGS 3D geology model as the conductivity field

The aquifer conductivity field comes from a pre-computed Transition Probability (TP) geostatistical model built from Michigan Wellogic borehole data. To apply it to the aquifer:

Inside the DM zone's Flow Properties, check Conductivity and select Borehole Simulation as the sub-option. Click the … options button to open Borehole Simulation Options. Check Import and browse to BarronLakeTP.tp (the T-PROGS result file). Assign typical x-direction conductivity values for the 4 material types identified by the geology model:

3. Configure overland / watershed modeling

Still inside the DM zone's Flow Properties, click the Overland Flow… button towards the bottom-right. This opens the Surface Water Overland Flow Options menu. Check Overland Flow to direct IGW-NET to simulate land-surface hydrologic processes — runoff, recharge, evapotranspiration, snowpack dynamics.

4. Load the lake polygon

Load BarronLake.txt as a zone via the text-file import. Two prompts require attention: tolerance (for polygon smoothing) should be 0 (no smoothing applied); "delete all existing attribute zones" should be Cancel (keep the DM zone intact).

5. Parameterize the lake — two-way coupling with bathymetry and stage

In the lake zone's Zone Attributes, navigate to the Source and Sinks Head Dependent tab and check Two-way Head Dependent. Configure:

Click the From coupled… button to open the Surface Water Source and Sink interface. Configure transient inputs:

• Under SW Prescribed Source/Sink, select Transient and click …. Load Qsw.csv — the augmentation pumping into the lake.
• Check SW Stage Measurement and click …. Load LakeLevel2013.csv — observed lake levels for validation.

6. Activate the Watershed Solver

Open DomainAttr and navigate to Simulation Settings. Click the Watershed Solver button.

7. Add the lake augmentation well

The augmentation well extracts water from the deep aquifer and delivers it to the lake — a common management strategy for small recreational lakes with marginal water balance. Draw a well anywhere using DrawWell, then correct coordinates in the Well Input Options menu:

8. Transient boundary conditions from a shapefile

Time-varying boundary conditions around the lake perimeter are based on regional flow patterns and water-level records at 11 private wells. They're modeled as transient "rivers" with very high leakance (mimicking prescribed-head BCs) but conceptualized as two-way head-dependent source/sink zones.

Import the shapefile

Conceptual Model Tools → Zones → Zone from a shapefile opens the Shapefile Import Options interface. Upload StagePolygons.shp, .dbf, .prj (all three required). Filter by MUNICIPLTY and select all entries. Click Next.

Under Select Attributes Options:

• Choose Used as Zone (more sub-options appear)
• Choose Shape-specified and select MUNICIPLTY from the dropdown — each zone is named by its MUNICIPLTY entry
• Check Two-way flux; group name River (for mass balance differentiation)
• Constant water depth: 1 m
• Constant Leakance: 5000 1/day (very high → mimics prescribed head)
• Click Add to MAGNET4WATER Model

Add time-varying stage to each zone

For each imported zone:

• Click Other Tools → Utilities → Geometry unlock, then click inside the zone. A prompt confirms which zone was selected (by MUNICIPLTY name).
• In Source and Sinks Head Dependent, check Two-way Head Dependent
• Choose Const, check Transient, click … to open Transient Data Input
• Click Load File and browse to the appropriate stage CSV — StageMyer.csv, StageKelly.csv, etc.
• Copy the initial water level (first row), paste into the Const field — this sets the initial value for the transient simulation
• Click Save

9. Grid, layers, and transient flow settings

10. Run and analyze the coupled simulation

Simulation Tools → SIMULATE. Accept the suggested projection. The status bar reports progress; after the first time step (1 day) is solved, head contours and velocity vectors appear over the domain.

Analysis Tools → Analysis → Display Charts opens the full chart suite. Configurations of note:

• In the Cross Section Diagram, enable Conductivity (Kzz) and Big Plot. Colors show vertical conductivity (red = high; blue = low), revealing the T-PROGS heterogeneity structure.
• In the SW Lake Chart, enable Show Obs Data and Lake Budget, then Redraw. This overlays the observed lake levels (from LakeLevel2013.csv) on the simulated stage, and opens the Lake Budget Chart showing inflows and outflows.

Technical notes and advanced features exercised

This model exercises several advanced platform features that are each touched by individual quick tutorials but rarely combined. The coupling patterns are worth noting:

• T-PROGS ↔ groundwater flow: The 3D geologic model drives conductivity assignment. Material-specific K values are the user-supplied interpretation of the TP simulation.
• Watershed solver ↔ groundwater recharge: The SW Infiltration as GW Recharge flag converts watershed-scale infiltration into GW recharge directly, without user-supplied recharge rates.
• Lake ↔ watershed ↔ groundwater: The lake receives runoff from the watershed solver, exchanges flow with groundwater through bathymetry-based leakance, and receives augmentation pumping as a direct SW input. All three are active simultaneously.
• Shapefile BCs ↔ private-well observations: The transient stage data at BC zones is inferred from private monitoring wells — the model's boundary conditions are data, not idealization.