Gravity observations at or above the Earth’s surface are typically reduced by a reference normal gravity for geoid modelling and geophysical prospecting. Depending on where the normal gravity is calculated, two main types of anomalous gravity values are defined: ‘gravity anomaly’ and ‘gravity disturbance’. Direct ‘observation’ of the actual gravity on the geoid or ellipsoid is not feasible thus necessitating the downward continuation of the gravity observations to the appropriate reference surface for the computation of gravity anomalies or disturbances. With the emergence of space-based positioning systems such as GNSS, the heights of the observation points above the reference ellipsoid are now accessible, making gravity disturbances more readily available to measure than gravity anomalies. Different kinds of topographic corrections are applied to gravity observations for downward continuation and/or to extract information on interior and exterior features of the Earth, e.g., Bouguer gravity anomalies. Theoretically, using either type of anomalous gravity value should not matter; for such purposes, however, we argue that given the complexity of the definition of the gravity anomalies and the commission error in the orthometric heights, gravity disturbances are more accurate for geophysical and geodetic applications. We illustrate our findings by showing the magnitude of the additional topographic corrections required for effectively utilizing gravity anomalies in geoid modelling or geophysical prospecting, contrasting them with the scenario in which gravity disturbances are utilized. Our computations are performed in the challenging topographical and geological area of the Colorado region in the USA.