Geoid Models Using Combinations of Airborne and Ground Gravity Data - a Case Study From the Kimberley Region, Australia

Most national jurisdictions have moved towards defining their height reference systems in terms of a gravimetric geoid or quasi-geoid, in order to allow vertical positioning using GNSS measurements in combination with knowledge of the local geoid. Airborne gravity measurements have proven to be very useful in geoid computation, both as a sole source of data in difficult areas, and as an addition to ground gravity data. Many areas of the world have only sparse and unreliable ground gravity coverage, and benefit greatly by the acquisition of airborne gravity data. In this presentation, airborne gravity data using Sander Geophysics' AIRGrav system is used to compute geoid models, which are then compared with GNSS/leveling data and current geoid models such as AUSGeoid2020.

In this presentation we focus on the Kimberley region of Western Australia, having a rough gravity field with widely spaced ground gravity data, and where some reasonably accurate GNSS/leveling data are available. The airborne data used are based on two surveys flown for the Geological Survey of Western Australia and Geoscience Australia in 2016 and 2018, both at a 1 mGal accuracy or better. The survey area used in this study covers an area approximately 500 x 500 km, with flight lines flown at a spacing at 2.5 km. The lines in the north extend up to 60 km offshore in some places.

Results from the computation of the geoid using different combinations of gravity data, including airborne-only (variable spacing of vertical only and with horizontal components), and airborne with ground data, will be presented and inter-compared, using least squares collocation and Fourier methods for downward continuation, data transformation and rigorous error estimation. This highlights the possibility to approach the 1 cm geoid accuracy for such a remote region, when combined with GOCE satellite data. One of the difficulties often encountered in verifying the accuracy of precise gravimetric geoid models is the quality of GNSS/leveling data that can be used as a ground truth; for Australia, the height datum is tied to multiple tide gauges, and leveling therefore does not correspond to a geoid surface. We illustrate this by the geoid inter-comparison to available GNSS/leveling data, and also with comparisons to satellite altimetry data offshore.