DataNET-based Groundwater Model — DataNET Quick Tutorial 9

1Overview & prerequisites

This tutorial demonstrates how to use web data service layers in DataNET to build a data-enabled groundwater model in IGW-NET. Although the "bridge" between the MAGNET4WATER platforms is general — you can transfer and flexibly apply any point, line, zone (WFS), or raster (WCS) layer — this tutorial focuses on four spatial framework data layers that are critically important in groundwater modeling:

Aquifer top / land surface geometry — via a DEM raster
Aquifer thickness / aquifer bottom geometry — via a glacial aquifer thickness raster
Hydraulic conductivity (and its spatial variability) — via a regional glacial aquifer horizontal-K raster
Natural recharge — via a regional effective-recharge raster (2000–2013 average)

The example also compares simulated heads to regional Static Water Levels (SWL) from water wells integrated into the MAGNET4WATER Data Center — a quick, realistic calibration check built right into the modeling workflow.

⭐ What makes the DataNET route special

Any data → any model → any field. Federated. LiDAR-ready. Hierarchical.

1. The DataNET route is flexible. This tutorial happens to use four layers — DEM, recharge, conductivity, aquifer thickness — from the MAGNET global base model. But the route is fully general. Any WMS image, any WFS vector (points, lines, polygons), any WCS raster can be mapped to any IGW-NET model field via the same Send Data dialog. Swap the 300 m DEM for a 1 m LiDAR surface. Override global recharge with a regional study product. Add a bedrock surface from state geology. Your model improves as your data improves.

2. Users improve the global base model. The base is the starting point — pre-assembled from high-quality global datasets (SRTM, SSURGO, NLCD, NLDAS, NHDPlus, USGS NWIS, and more) so that 80–90% of a typical model is already built when you zoom in. When a user contributes a better local dataset through DataNET — a refined conductivity field, a locally calibrated recharge, a high-resolution geologic surface — it feeds back into the global database. Next user zooming to that location starts from a better foundation. The base is living, not frozen.

3. The data services are federated — data stays with its owner. DataNET is not a warehouse that copies everyone's data. It's a federation of WMS/WFS/WCS endpoints. Every dataset stays hosted by its authoritative owner — NASA, USGS, NOAA, ESA, USDA, EPA, FEMA, NRCan, state agencies, universities, private providers. DataNET routes the request, subsets to your analysis area, and streams exactly what the model needs. No bulk downloads. No stale copies. No licensing re-distribution. No governance handoff.

4. Particularly handy for sub-meter LiDAR DEMs. The U.S. Geological Survey's 3D Elevation Program (3DEP) now provides seamless sub-meter LiDAR coverage across essentially all of North America — a single consistent, authoritative source replacing the patchwork of state, county, and agency tiles that used to be the norm. This is a landmark dataset for hydrologic modeling. But raw, it's impractical to use: each state's portion is terabytes; a single county can run tens to hundreds of gigabytes; no desktop GIS wants to open it. DataNET makes it efficient and effective to use. Through the WCS endpoint:

At regional scale, the same LiDAR source is auto-subsampled on the server to whatever coarser resolution the model needs — 30 m, 10 m, 3 m — and streamed in seconds. One consistent DEM source from continental scale down to meter scale.
At local scale, the full sub-meter detail is delivered to the submodel that actually needs it — no downsampling, no loss of the micro-topography that drives the answer.

That matters in a wide range of groundwater applications where resolution isn't a cosmetic choice — it's the physics:

Headwater streams and ephemeral channels. Their location and connectivity with the water table are controlled by meter-scale terrain features that coarse DEMs wash out entirely. Get the DEM wrong here and you miss the streams.
Groundwater-dependent ecosystems (GDEs). Fens, seeps, peatlands, and phreatophyte communities sit at subtle topographic thresholds — a few tens of centimeters of discharge-zone relief can decide whether the ecosystem is sustained or dewatered.
Riparian discharge, seepage faces, and springs. Location and flux depend on the intersection of the water table with micro-topography. Coarser DEMs smear discharge across whole cells; LiDAR resolves it to the feature.
Shallow-water-table and near-surface flow systems. Drainage density, depression storage, and wetland connectivity are first-order controls on recharge partitioning — all DEM-resolution-limited.

In all of these cases, DEM resolution isn't window dressing — it's the difference between a model that represents the real physical system and one that averages away the features that matter. DataNET turns USGS 3DEP from a dataset you admire into a dataset you actually use, at the exact resolution each part of your model needs.

5. Works best together with hierarchical modeling. The real power of the DataNET route emerges when it feeds a hierarchical IGW-NET model — the signature MAGNET pattern of nested parent/child submodels. A coarse regional parent pulls DEM/recharge/conductivity at continental resolution; child submodels pull the same underlying datasets at medium resolution over their smaller extents; grand-child local models pull them at full native resolution — 3DEP LiDAR at 1 m, SSURGO soils at source polygon detail, and so on. Each child inherits boundary conditions automatically from its parent. DataNET delivers the right data, from the same consistent sources, at the right resolution to each tier, on demand. Detail only where it matters. That combination — federated data services feeding a hierarchical solver — is what makes continent-scale groundwater modeling with sub-meter local detail practical for the first time. See IGW-NET hierarchical modeling for the full pattern.

What you'll need

A free MAGNET4WATER account (sign up at magnet4water.net)
Familiarity with the Data Library & Workspace (Tutorial 1) and Transfer Data to Modeling Platforms (Tutorial 2)
A modern browser (any OS; no install required)

2Load the Maple Creek Watershed as the model domain

We'll use the Maple Creek Watershed near Fargo, North Dakota — a manageable glacial aquifer system with good monitoring-well coverage.

In IGW-NET, open Conceptual Model Tools › DrawDomain › Watershed from DataCenter.
In the Server Watershed Options interface, uncheck the box next to Generalization. This tells IGW-NET to extract the original, detailed watershed boundary with all vertices (rather than a smoothed / generalized version with fewer effective vertices).
Use the DataCenter1 (USA Only) option to utilize NHDPlus data, and select Level 4 (HUC 8-digit) in the Watershed Level dropdown.
Click on the map a bit west-southwest of Fargo, ND, to update the extraction coordinates.
Click OK. After a few moments (and a few prompts), the Maple Creek Watershed boundary will be added to the map display as the model boundary.

Figure 1 — Loading the Maple Creek Watershed boundary from DataCenter at HUC-8 level — Figure 1 — Server Watershed Options dialog and extracted Maple Creek Watershed boundary (red outline).

3Open DataNET and compile a layer workspace

Now we move into DataNET to assemble the four data layers we need.

In IGW-NET, open Analysis Tools › Data Explorer › DataNET. Follow prompts to open the linked DataNET page and log in to your MAGNET4WATER account.
Back in IGW-NET, again go to Analysis Tools › Data Explorer › DataNET — this step transfers the model domain polygon into DataNET so DataNET knows the spatial scope for clipping transfers.
Within the DataNET window, open the Data Layer Library via Data Layers › Search Library.
Click Workspace. This creates a Workspace to which you'll add layers from the Library.
Select the region to search:
- Enter united states in the search bar and click find region
- Check the box next to United States
Select the format to search: uncheck all boxes except WCS (raster).
Click Get Layers. After a few moments, national-level WCS layers for the United States will load into the bottom portion of the Library.
Sort results by Category and add the following four layers to the Workspace (click the + button next to each layer name):
- United States Effective Recharge 00-13 — under Groundwater Recharge
- United States Glacial Aquifer Thickness Based on Water Well Records — under Groundwater Aquifer Elevations and Thickness
- United States Glacial Aquifer Horizontal Hydraulic Conductivity — under Groundwater Aquifer Properties
- United States DEM 300m — under Land Digital Elevations

Figure 2 — DataNET Data Layer Library with the four required WCS layers added to the Workspace — Figure 2 — DataNET Data Layer Library — four WCS layers added to the Workspace.

4Transfer the DEM — use as Top Elevation

The DEM raster serves two purposes: first, as a visual background overlay to orient the model; second, as the aquifer top elevation (TopE).

4.1 Initiate the transfer in DataNET

Select the DEM 300m layer in the Workspace (check the box next to the layer title).
Open Transfer Modeling Data › Send Data Layer within Analysis Area.
In the Send Data interface, set format to WCS.
Set Data Res. to 100. Rasters at different original resolutions will be resampled to 100px size.
Click Send. After a few moments, a Data Transferring prompt will appear in IGW-NET.

Figure 3 — Initiating the DEM layer transfer in DataNET — Figure 3 — Send Data interface — DEM 300m, WCS, 100px resolution.

4.2 Overlay the DEM as a map background in IGW-NET

Click OK to receive data from the DataNET page.
In the first prompt (…Overlay on maps only?), choose OK.
Choose OK to rendering from server side.
Accept default draw options and contour options.
Initially a grayscale image will appear, but after a few moments a color+contour rendering of the DEM in the model bounding box will appear. Areas of very high DEM are red; very low DEM are blue.

Figure 4 — DEM overlay on the IGW-NET map background for the Maple Creek Watershed — Figure 4 — DEM color+contour overlay — red = high, blue = low.

4.3 Import the DEM as Top Elevation (TopE)

Back in DataNET, click Send once more in the Send Data interface.
This time, for the first prompt (…Overlay on maps only?), choose Cancel.
In the next prompt, choose option 1 to use the raster layer as TopE (Top Elevation).
After a few moments, the file uploads to your MAGNET4WATER user folder and the Domain Attributes menu opens. The box next to Import is checked under Top Elevation, and the uploaded file is selected.
Click Save in Domain Attributes.

Figure 5 — Importing the DEM raster as the aquifer Top Elevation (TopE) in the IGW-NET Domain Attributes — Figure 5 — Domain Attributes — Top Elevation imported from the DEM raster.

5Transfer natural recharge

The United States Effective Recharge 00-13 raster represents the 2000–2013 average effective recharge.

5.1 Overlay recharge to the map background

In the DataNET Workspace, uncheck the DEM 300m layer and check United States Effective Recharge 00-13.
Click Send in the Send Data interface.
Select OK to the …Overlay on maps only? prompt.
Use 1 to apply the invert contour color-map option for rendering; click OK.
After a few moments a color+contour of recharge appears in the Map Display. Note: very low recharge is red, very high recharge is blue (inverted from DEM).

Figure 6 — Effective recharge overlay on the IGW-NET map background — Figure 6 — Recharge overlay with inverted color map (red = low, blue = high).

5.2 Import recharge as a model input

Send the data again; choose Cancel to the Overlay on maps? prompt.
In the next prompt, choose option 4 to use the raster layer as Recharge.
After a few moments the file uploads to your MAGNET4WATER user folder, and in Domain Attributes the Import box is checked under Recharge.
Click Save.

Figure 7 — Importing the effective recharge raster as the Recharge input in IGW-NET Domain Attributes — Figure 7 — Domain Attributes — Recharge imported.

6Transfer hydraulic conductivity

6.1 Overlay hydraulic conductivity

Uncheck the recharge layer in the Workspace; check United States Glacial Aquifer Horizontal Hydraulic Conductivity.
Click Send; select OK to the …Overlay on maps only? prompt; proceed through the rendering prompts.
A color+contour of horizontal hydraulic conductivity of the glacial (surficial) aquifer appears.

Figure 8 — Horizontal hydraulic conductivity overlay for the glacial aquifer — Figure 8 — Glacial aquifer horizontal hydraulic conductivity overlay.

6.2 Import conductivity as a model input

Send the data again; choose Cancel to the Overlay on maps? prompt.
In the next prompt, choose option 3 to use the raster layer as Conductivity.
After a few moments the file uploads and the Import box is checked under Hydraulic Conductivity in Domain Attributes.
Click Save.

Figure 9 — Importing the hydraulic conductivity raster as the K input in IGW-NET Domain Attributes — Figure 9 — Domain Attributes — Hydraulic Conductivity imported.

7Transfer aquifer thickness → Bottom Elevation

7.1 Overlay glacial aquifer thickness

Uncheck the hydraulic conductivity layer; check United States Glacial Thickness Based on Water Well Records.
Click Send; select OK to the …Overlay on maps only? prompt.
A color+contour of the glacial aquifer thickness appears.

Figure 10 — Glacial aquifer thickness overlay on the Maple Creek domain — Figure 10 — Glacial aquifer thickness (from water-well records).

7.2 Import thickness as Bottom Elevation (BotE)

Send the data again; choose Cancel to the Overlay on maps? prompt.
In the next prompt, choose option 2 to use the raster layer as BotE (Bottom Elevation).
After a few moments the file uploads and the Import box is checked under Bottom Elevation.
Important: in the Domain Attributes, make sure to select the Thickness option — the raster consists of aquifer thickness values, not aquifer bottom elevation values. IGW-NET will subtract thickness from the top elevation to produce the bottom surface.
Click Save.

⚠️ Thickness vs. bottom elevation. The glacial aquifer raster we just transferred contains thickness values — meters of saturated sediment — not absolute bottom elevations. IGW-NET handles both cases, but you must tell it which kind of raster you're importing via the Thickness checkbox in Domain Attributes. Miss this step and your model will fail or produce nonsense bottoms.

8Simulate and visualize

With all four data layers imported, we have a complete first-cut model — domain geometry from the HUC-8 watershed, top and bottom elevations from DEM and aquifer thickness, hydraulic conductivity from the regional K raster, and recharge from the 00–13 effective recharge product. Time to run it.

Open Simulation Tools › SIMULATE.
After a few moments, the plan-view results appear in the Map Display — head contours (color shading) and velocity vectors.
For a cross-section view: Analysis Tools › X-Section › draw a section on the map › SaveShape (in Conceptual Model Tools) › wait for cross-section plots to load.

Figure 11 — Plan-view simulation results: head contours and velocity vectors across the Maple Creek Watershed — Figure 11 — Plan-view simulation results — head contours and velocity vectors.

Figure 12 — Cross-section results showing water table and aquifer geometry along a user-drawn transect — Figure 12 — Cross-section view along a user-drawn transect.

9Compare simulated heads to regional Static Water Levels

The MAGNET4WATER Data Center integrates regional Static Water Level (SWL) records from water wells. We can pull these observations and chart them against our simulated heads — a quick reality check on the model.

Open Analysis Tools › Calibration. In the Calibration Chart interface (Figure 13), select IGW Server as the Data source.
Use default options in the Server Data Filters interface and load the extracted data into the Calibration Data Input chart.
Click OK to finish extraction and draw the Calibration Chart (Figure 14).
Check the boxes next to Show Std and Add Band-mean to add confidence intervals (1 standard deviation) and moving-window averages to the plot, respectively.

Figure 13 — Calibration Chart interface with IGW Server data source selection and Server Data Filters — Figure 13 — Calibration Chart interface — IGW Server data source + Server Data Filters.

Figure 14 — Calibration chart comparing simulated heads against observed Static Water Levels with Show Std and Add Band-mean enabled — Figure 14 — Simulated vs. observed heads with ±1σ band and moving-window mean.

Note the reasonably good agreement between simulated heads and observed static water levels — a strong indication that the framework data layers (DEM, thickness, conductivity, recharge) together produce a physically credible first-cut regional groundwater model.

DataNET-based Groundwater Model — Maple Creek Watershed, ND

This tutorial covers