The Geoid: Where Gravity Enters the Story
So far, every step in the processing chain has been geometric. The satellites supplied observations, the RTK engine solved a position, and the receiver expressed that position as latitude, longitude, and ellipsoid height. The geoid does not move or recompute that coordinate. It introduces gravity so the same vertical position can be described as the kind of elevation used in surveying, mapping, and engineering.
Geometry Has Taken Us as Far as It Can
At this point in the journey, the receiver has everything it needs to describe the antenna position geometrically. Latitude tells us where the point lies north or south. Longitude tells us where it lies east or west. Ellipsoid height tells us how far the antenna is above or below the reference ellipsoid. From the positioning engine's perspective, the problem is solved. The coordinate is complete, precise, and internally consistent.
Yet most engineering projects are not interested in geometry alone. Water in a storm drain does not flow relative to a mathematical ellipsoid. A roadway grade is not established by asking how far the pavement lies above a smooth reference surface wrapped around Earth. Surveyors, engineers, and mapping agencies usually need a height that is connected to gravity, because gravity determines what we experience as level and which direction water will flow.
GNSS is exceptionally good at solving a geometric problem. Most engineering projects require a gravity-based height. Those are related problems, but they are not the same problem.
A World Without Gravity
Imagine, for a moment, that gravity could be turned off while everything else remained in place. Earth would still exist, the satellites would still orbit, and a GNSS receiver could still compute latitude, longitude, and ellipsoid height. The geometric positioning process we have followed so far would continue to work because it does not need a definition of sea level to determine where the antenna is in space.
But without gravity, the phrase 'elevation above sea level' would lose its meaning. Water would not settle into oceans, rivers would not flow downhill, and there would be no natural level surface for engineers to use as a practical vertical reference. The receiver could still locate the antenna with extraordinary precision, but it could not produce the kind of elevation most people expect when they compare a rover reading with a benchmark.
That thought experiment reveals the distinction we need. Ellipsoid height comes from geometry. The elevation used for most engineering and mapping work depends on gravity. Until gravity enters the processing chain, we still have only the first kind of height.
A geoid model does not make the GNSS position more accurate. The receiver computed the position before the geoid entered the process. The model changes how the vertical component is referenced and reported.
Imagine Covering the Earth with Water
Now imagine covering the entire Earth with an endless, perfectly calm ocean. Remove the tides, waves, winds, and currents, then allow the water to settle under gravity. The surface formed by that idealized ocean is a useful way to begin thinking about the geoid. It represents a surface that is everywhere level with respect to Earth's gravity field.
Unlike the ellipsoid, that surface would not be perfectly smooth or mathematically simple. Earth's mass is not distributed uniformly. Mountain ranges, deep ocean trenches, variations in rock density, and structures inside the planet all influence the gravity field. As a result, the gravity-defined surface rises and falls relative to the smooth reference ellipsoid.
The geoid is a modeled version of that gravity-based level surface. It is not another coordinate system, and it does not replace the ellipsoid used by GNSS. Instead, it provides the missing relationship between a geometric height and an elevation that behaves the way engineers expect relative to gravity.
The geoid does not determine where the coordinate is. It provides the gravity-based reference needed to describe that coordinate as an orthometric elevation. The position remains the same; only the vertical reference and reported height change.
We now have two valid ways to describe the height of the same point: ellipsoid height and orthometric height. In the next section, we will place them side by side, introduce geoid height, and show the simple relationship that connects all three.