Common GPS Elevation Determination Methods in Surveying Applications

Accurate elevation data is just as important as horizontal positioning in many surveying and engineering projects. While GPS and GNSS systems can provide highly accurate three-dimensional coordinates, converting satellite-derived heights into usable elevation values often requires additional processing methods.

In practical surveying work, several approaches are commonly used to determine elevations from GPS measurements. The choice of method depends on terrain conditions, required accuracy, available reference data, and project scale.

Below are some of the most widely used methods.

One of the traditional approaches involves using geoid separation maps or height anomaly contour maps.

Surveyors first obtain the geoid separation or height anomaly value for a given location from a contour map. These values can then be combined with GPS-derived ellipsoidal heights to calculate either:

• Orthometric height
• Normal height

Although the process itself is relatively straightforward, several practical considerations should be kept in mind.

The contour map used must correspond to the same coordinate reference system as the GPS observations.

If the elevation model and GNSS measurements are based on different coordinate systems, calculation errors may occur.

The final elevation quality is heavily influenced by the accuracy of the contour map itself.

Even if GNSS positioning data is highly precise, inaccurate or low-resolution contour information can reduce the reliability of the final results.

For this reason, the contour map method is generally suitable only when reliable elevation reference data is available.

A geoid model can be considered a digital version of a contour-based approach.

Instead of manually reading values from maps, mathematical Earth models are used to estimate geoid separations across a region.

Several international geoid models have historically been used, including:

• OSU91A
• EGM series models
• Regional geoid models

These models simplify elevation conversion and improve efficiency during data processing.

However, one practical challenge is that global models do not always perform equally well in every region.

Local terrain conditions and geodetic characteristics often require country-specific or regional geoid models to achieve better results.

For this reason, many countries maintain their own localized geoid solutions for higher-precision applications.

In real-world projects, especially for local surveying work, elevation fitting is frequently used.

Elevation fitting is based on the observation that within relatively small areas, height anomalies often follow predictable spatial patterns.

Using known reference points and mathematical fitting techniques, surveyors can estimate:

• Height anomalies
• Orthometric heights
• Normal heights

The method essentially establishes a mathematical relationship between GPS-derived ellipsoidal heights and known elevation values.

Elevation fitting is fundamentally a geometric approach.

As a result, it generally performs best in areas where height anomalies change gradually, such as:

• Flat terrain
• Plains
• Low-relief regions

Under favorable conditions, fitting accuracy can often remain within a few centimeters to one decimeter.

In mountainous or highly variable terrain, performance may decrease significantly because elevation changes become more complex and difficult to model.

The quality of the fitting model depends heavily on the reference points used.

Known height anomaly values are typically obtained by combining:

• Precise leveling measurements for normal heights
• GPS observations for ellipsoidal heights

In practical field operations, surveyors commonly:

• Establish GPS points at benchmark locations
• Connect GNSS observations with leveling networks

For better fitting performance, reference points should:

• Be evenly distributed
• Cover the entire survey area whenever possible
• Surround the GNSS network rather than clustering in one location

Poor point distribution can lead to unstable fitting results.

The required number of known points depends on the fitting model being used.

Typical examples include:

For larger projects, a single fitting model may not adequately represent the entire survey area.

In these situations, surveyors often divide the project into several smaller zones.

Each region is fitted independently using local control points.

Boundary control points can be shared between neighboring regions to maintain consistency.

This partition approach often provides better results for large-scale GPS networks, particularly when terrain characteristics vary significantly across the project area.

Determining elevations from GPS measurements is not simply a matter of reading coordinates from a receiver.

The process requires appropriate transformation methods and careful consideration of terrain, reference data, and project requirements.

Whether using contour maps, geoid models, or fitting techniques, selecting the right approach can greatly improve elevation accuracy and overall surveying efficiency.

As GNSS technology continues to evolve, combining high-quality positioning data with reliable elevation models remains a key factor in achieving accurate survey results.