1Define the Model Domain
Step 1 — Load a Watershed Boundary from the Data Center
Go to 'Conceptual Model Tools' → 'DrawDomain' → 'Watershed from DataCenter'. In the Server Watershed Options interface:
Uncheck 'Generalization' (preserve the detailed boundary with all vertices)
Select 'DataCenter1 (USA Only)' for NHDPlus data
Select 'Level 4 (HUC 8-digit)' as the Watershed Level
Click on the map west-southwest of Fargo, North Dakota
Click 'OK'. The Maple Creek Watershed boundary loads as the model domain — a hydrologically meaningful boundary extracted from the national hydrography database, not an arbitrary rectangle.
2Build a DataNET Workspace
Step 2 — Open DataNET and Compile Data Layers
Go to 'Analysis Tools' → 'DataNET' and follow the prompts to open the linked DataNET page. In DataNET, open the Data TreeView (Show Layers). Search for "united states" and click 'Get Layers'. Create a new Workspace and add four critical layers:
1. United States Effective Recharge 00-13 (under Groundwater Recharge)
2. United States Glacial Aquifer Thickness USGS WWDR (under Aquifer Thickness)
3. United States Glacial Aquifer Horizontal Hydraulic Conductivity (under Hydraulic Properties)
4. United States DEM 300m (under Land Elevations)
These four layers provide the essential spatial framework for a groundwater model: terrain geometry, aquifer geometry, flow properties, and water input. All from federal data services — USGS, NHDPlus — federated through DataNET.
The General Transfer Mechanism
How it works: The Data Send Interaction interface is the bridge between DataNET and IGW-NET. For raster (WCS) layers, you select the layer in the Workspace, choose a resolution (all layers are resampled to a common resolution), and click 'Send'. The raster arrives in IGW-NET where you assign it to a model property: option 1 = Top Elevation, option 2 = Bottom Elevation, option 3 = Conductivity, option 4 = Recharge.
Any data → any property: The mechanism is completely general. A recharge raster from a research product? Assign to recharge. A soil K map from USDA? Assign to conductivity. A satellite-derived DEM from ESA? Assign to top elevation. A custom contaminant concentration map? Assign to initial concentration. The user decides what each layer means — the platform provides the transfer pipeline.
Preview before you commit: Each layer can be overlaid on the map first (the "Overlay on maps only?" option). This lets you visualize the data — check spatial coverage, verify the range of values, see the spatial pattern — before assigning it to a model property. Look at the data before trusting it.
3Transfer Data Layers
Step 3 — Initiate the DEM Transfer
In IGW-NET, go to 'Analysis Tools' → 'DataNET' to transfer the model domain polygon to DataNET. In DataNET, go to 'Data Manipulation' → 'Transfer'. In the Data Send Interaction interface, select 'WCS' for Services, '300' for Data Resolution (300m — matching the DEM layer's native resolution). Select the DEM 300m layer and click 'Send'.
Step 4 — Preview the DEM
Click 'OK' to receive data from DataNET. At the prompt, choose 'OK' to overlay on maps only. A color-contour rendering of the DEM appears — red for high elevation, blue for low. This is your preview: verify that the terrain looks correct for the Maple Creek Watershed before committing it to the model.
Step 5 — Assign DEM as Top Elevation
Send the DEM again. This time, choose 'Cancel' at the overlay prompt. Select option '1' to assign the raster as Top Elevation. The file uploads to your MAGNET user folder. In Domain Attributes, the 'Import' box is checked under Top Elevation with the uploaded file selected. Click 'Save'.
Step 6 — Preview and Transfer Recharge
In the DataNET Workspace, switch to 'United States Effective Recharge 00-13' (average effective recharge from 2000–2013). Send and preview first — blue = high recharge, red = low. Then send again and assign as option '4' (Recharge). Click 'Save' in Domain Attributes.
Step 7 — Preview and Transfer Hydraulic Conductivity
Switch to 'Glacial Aquifer Horizontal Hydraulic Conductivity'. Preview the spatial pattern — K varies across the watershed reflecting glacial geology. Send and assign as option '3' (Conductivity). Click 'Save'.
Step 8 — Preview and Transfer Aquifer Thickness
Switch to 'Glacial Aquifer Thickness USGS WWDR'. Preview to see how thickness varies. Send and assign as option '2' (Bottom Elevation). In Domain Attributes, select the 'Thickness' option — because this raster contains thickness values, not bottom elevation values. IGW-NET computes bottom elevation as: Top Elevation minus Thickness. Click 'Save'.
4Simulate and Validate
Step 9 — Run the DataNET-built Model
Go to 'Simulation Tools' → 'SIMULATE'. The model solves using all four DataNET layers — terrain, recharge, K, and aquifer geometry — as spatially variable inputs across the Maple Creek Watershed. Head contours and velocity vectors appear in plan view. Draw a cross-section ('Analysis Tools' → 'X-Section') to see the vertical structure — the aquifer geometry defined by the DEM and thickness layers, with heads computed through the full thickness.
Step 10 — Compare to Observed Water Levels
Go to 'Analysis Tools' → 'Calibration'. Select 'IGW Server' as the Data Source. The Server Data Filters interface extracts static water levels from water wells in the MAGNET Data Center — real measurements from the region. Click 'OK' to load the data and draw the Calibration Chart.
Check 'Show Std' and 'Add Band-mean' to add confidence intervals (±1 standard deviation) and moving-window averages. The chart shows reasonably good agreement between simulated and observed heads — the DataNET-built model, without any manual calibration, captures the regional head distribution from federated data alone.
Key Concepts
The model was built entirely from web services: No data was downloaded to a desktop. No files were manually processed. Four raster layers from federal databases, transferred through DataNET's bridge, became a working groundwater model with spatially variable properties. This is the "algorithms come to the data" philosophy made concrete.
Validation without calibration: The comparison against static water levels (Step 10) is not calibration — no parameters were adjusted. It's validation of the data-driven approach: can you build a meaningful model from federated data alone? The reasonable agreement says yes. This is the power of the pre-processed and federated data: the initial model already captures the essential physics. Calibration then refines it further.
The general transfer is unlimited: This tutorial used four layers — but DataNET connects to hundreds. Any WCS raster from any supported service can be transferred to any model property. Soil data from USDA, climate data from NOAA, geological data from USGS, satellite data from ESA — all available through the same interface. The tutorial demonstrates the pattern; the applications are limitless.
Watershed boundaries as model domains: Step 1 used a HUC-8 watershed boundary as the model domain — not an arbitrary rectangle. This ensures the model boundary is hydrologically meaningful: the watershed divide is a natural no-flow boundary for groundwater. Combined with DataNET's spatial data, the entire model setup — domain, terrain, properties, recharge — can be assembled from server data in minutes.
5What's Next
One more tutorial completes the IGW-NET Quick Tutorial series:
Tutorial 28: Data Processing & Regression — statistical analysis, spatial interpolation, and regression tools for exploring and preparing data