1Spatial Analysis of Groundwater Levels
Step 1 — Load and Simulate the Regional Model
Click 
to load the regional model from Tutorial 1. Submit the model for simulation to generate the simulated head distribution.
Step 2 — Launch Calibration Tool
Click 
the 'Calibration' button. When the prompt appears, choose 'Cancel' — this tells IGW-NET to perform a model-wide spatial comparison rather than a single-point calibration. The tool will compare simulated heads against observed water levels across the entire domain.
Step 3 — Select Observed Data Source
In the window that appears, select a data source for water level measurements from the drop-down menu. This example extracts data from the IGW Server — live-linked observed water levels from the MAGNET4WATER Data Center, sourced from USGS and state monitoring networks. No manual data entry needed.
Step 4 — Apply Data Filters
Apply filters to select the appropriate observation data — by date range, well depth, data quality, or spatial extent. Select 'OK' to extract the filtered data from the server. Once extraction completes, select 'OK' again to proceed.
Step 5 — View the Calibration Chart
The calibration chart shows simulated vs. observed head for every observation point. Check 'Show Std' to add confidence intervals and 'Add Band-mean' to add a moving window average. Points near the 1:1 diagonal line indicate good calibration. Systematic deviations indicate model bias that needs correction.
Step 6 — Export Data for Further Analysis
Copy the downloaded data displayed below the chart for further offline analysis — statistical evaluation, custom plotting, or report generation.
2Time-Series Analysis
Step 1 — Create a Submodel and Save Parent
Click 
to save the 'Latest Model Zipped File' of the regional model. Add a submodel and apply 'Boundary Conditions from Parent Model' in the Default Attributes menu. Submit the submodel for simulation.
Step 2 — Add Pumping and Monitoring Wells
Click 
to add a pumping well (1250 GPM) in the northeast portion of the model domain. Then add a monitoring well just north of the pumping well. The monitoring well will record simulated heads over time for comparison with observed data.
Step 3 — Enter Observed Water Levels
Click 
the 'Observation H' button to enter observed water level measurements at the monitoring well location. Enter date-value pairs representing actual field measurements. These become the "truth" against which the model is evaluated.
Step 4 — Enable Transient Mode
Click 
to open the Default Attributes menu. Check 'Transient'. Set the starting date to match the first observation date from Step 3. Use a time step of 100 days. This ensures the simulation timeline aligns with the observation period.
Step 5 — Set Initial Conditions from Parent
Click — still within the Default Attributes menu, select 'Parent' under 'Initial & Boundary Condition for Head'. Then apply 'Boundary Conditions from Parent Model', check 'Import', and upload the zipped file from Step 1. This ensures the transient simulation starts from the calibrated steady-state condition.
Step 6 — Submit for Transient Simulation
Click 
to submit the model. The solver computes the transient response to the 1250 GPM pumping well — water levels draw down over time, and the monitoring well records the evolving head.
Step 7 — View Time-Series Calibration
Click 
the 'Analysis' button, then select 
'Display Charts'. The Monitoring Well chart displays simulated and observed water levels on the same plot — showing whether the model correctly captures the timing, magnitude, and shape of the drawdown response.
Step 8 — Save or Publish
Click 
to save or publish the calibrated model.
Interpreting Calibration Results
Spatial calibration (Part A) tells you: Is the overall flow pattern correct? Are heads too high everywhere (recharge too high)? Too low (conductivity too high)? Scattered (missing heterogeneity)? Biased in one region (wrong boundary condition)?
Time-series calibration (Part B) tells you: Does the model respond correctly to stress changes? If drawdown is too fast, specific yield may be too low. If drawdown is too slow, conductivity may be too high. If the shape is wrong, the boundary conditions or aquifer geometry may need revision.
Calibration is iterative: Compare → identify misfit → adjust parameters → re-simulate → compare again. Each cycle improves the model. The goal is not perfection — it's confidence that the model captures the essential behavior of the system for the decisions you need to make.
Data from the Server — No Manual Collection Needed
A powerful feature of IGW-NET calibration is the live connection to the MAGNET4WATER Data Center. Observed water levels from USGS, state monitoring networks, and other sources are extracted directly — no manual file preparation, no reformatting, no data wrangling. The platform connects model to reality with a few clicks. This is the "algorithms come to data" philosophy applied to calibration.
3What's Next
With calibration mastered, continue the learning path:
Tutorial 9: Synthetic Model — generate heterogeneous aquifers and explore how calibration changes with realistic heterogeneity
Tutorial 10: Aquifer Layers — calibrate multi-layer models with observations at different depths
Tutorial 19: Automatic Parameter Estimation — let the optimizer find the best-fit parameters systematically