Interpretation, assumptions, and limitations • potentiomap

potentiomap organizes a reproducible spatial workflow. It does not determine whether observations represent one hydraulic system or whether an interpolated surface is adequate for a particular decision.

Before interpolation

Prefer a synoptic measurement round when the objective is a spatial snapshot. Record date, time, time datum, weather or recharge context, and known pumping.
Confirm the measuring point, its offset from land surface, land-surface elevation, corrections, units, and vertical datum. Do not mix NAVD 88, NGVD 29, local datums, or assumed elevations without a documented conversion.
Confirm that positive depth-to-water values mean depth below land surface. Flowing wells may have negative depth-to-water values under that convention.
Review aquifer assignment, open or screened interval, well depth, nested-well construction, and confining units. Nearby wells can measure different heads because they represent different depth intervals or hydraulic units.
Flag pumped, recently pumped, obstructed, dry, flowing, or otherwise non-static wells. Decide whether each measurement represents the mapping objective; retain the decision trail.

The USGS Groundwater Technical Procedures cover standardized measuring-point establishment and water-level measurement. USGS furnished-record guidance also identifies site status, measuring-point offsets, corrections, well construction, and measurement timing as important documentation.

Network support and boundaries

Sparse or clustered networks constrain spatial detail unevenly. A fine grid does not resolve that limitation.
Extrapolation beyond the outer wells is weakly supported. A rectangular raster should not be mistaken for a defensible aquifer boundary.
Surface-water features may represent recharge, discharge, or neither, depending on hydraulic connection and timing. Their role is not inferred from map proximity alone.
Faults, aquitards, divides, pumping centers, recharge zones, and domain edges can alter the shape of the potentiometric surface. potentiomap does not add these conceptual boundaries automatically.
Nested wells should not be collapsed into one horizontal location without first resolving vertical hydraulic differences.

USGS describes groundwater hydrology as interpretive because the resource is not directly observable everywhere; monitoring networks provide point evidence from which spatial understanding is interpolated and extrapolated.

Interpolation and validation

TPS, IDW, ordinary kriging, and universal kriging encode different assumptions. No method is universally best. Select parameters with the monitoring geometry, conceptual hydrogeologic model, and intended use in mind. When results support consequential decisions, use an appropriate validation design, inspect residuals and geostatistical diagnostics, evaluate plausible alternatives, and document uncertainty.

Smoothing changes the modeled surface. It can support cartographic generalization, but it is not an accuracy correction. Retain the original surface and record the smoothing statistic, window, weights, and iteration count. Excessive smoothing can erase local variation or move contours.

Hydraulic gradients are not velocities

A hydraulic gradient is change in head per unit distance. ps_flow_arrows() uses the interpolated surface to infer the direction of decreasing modeled head. Its line length can be scaled for display and its density follows raster sampling, not monitoring density.

Groundwater velocity additionally depends on hydraulic conductivity and effective porosity, and flow paths can be affected by heterogeneity, anisotropy, vertical gradients, sources, sinks, and boundaries. The USGS discussion of Darcy’s law distinguishes hydraulic gradient from average linear velocity.

Hydraulic-gradient arrows do not represent groundwater velocity, travel time, particle paths, or contaminant transport. The package is not a process-based groundwater-flow model and does not replace hydrogeologic judgment.

Review record for a defensible map

At minimum, retain:

original measurements and qualifiers;
aquifer and screened-interval selection;
horizontal CRS and vertical datum;
measuring-point and land-surface corrections;
interpolation method and every parameter;
template geometry, mask, and extrapolation limits;
smoothing and contour settings;
gradient-arrow settings;
warnings, diagnostics, and validation results; and
package and dependency versions.

An interpolated surface is a model conditioned on data and assumptions. Treat it as one line of evidence within the conceptual hydrogeologic interpretation.

Review table

Use a compact review table to keep the result and its limiting evidence together in a project record.

product	review_evidence	does_not_establish
Observation set	Timing, aquifer, screened interval, datum, corrections, and qualifiers	One hydraulic system or error-free head values
Modeled surface	Method, parameters, grid, mask, diagnostics, and validation	A uniquely correct potentiometric surface
Contours	Interval or levels, source surface, and omitted-level manifest	Observed groundwater elevations between wells
Support classes	User-defined thresholds, hull rule, resolution, and reason fields	Statistical confidence without a suitable uncertainty model
Hydraulic-gradient arrows	Source surface, density, scale, endpoint validation, and direction check	Velocity, travel time, particle paths, or contaminant transport