Research and Projects

The geoid is the Earth's true reference surface for height, the shape the oceans would take if they covered the whole planet, following gravity alone. Determining it accurately is one of the oldest and hardest problems in geodesy, and it is the foundation of everything else we do.

Our modelling is built on the Stokes–Helmert approach, including the one-step integration method that solves for the geoid directly using gravity observations from ground, air, and satellite missions. We focus on the details that determines a rigorous geoid model from a merely acceptable one: how the error of the topography impacts geoid heights, how noise and data gravity data gaps propagate into the final surface, and how close we can get to a sub-centimetre geoid depending on the existing datasets. 

Our work is implemented in OSVGeo (One-Step integration Vector Geoid Determination), our fully developed software for regional and national geoid modelling. We support geoid determination for nations and regions worldwide, and offer full training on the software using real test cases such as Auvergne in France and Colorado in the USA. [Learn more about OSVGeo and training by clicking the link below]

We are changing what gravity field models are built for, from determining height to serving as a position reference where satellite navigation cannot be relied on. GNSS signals can be jammed, spoofed, or simply unavailable underground, underwater, and at high latitudes. Gravity offers an alternative that cannot be switched off and can be used for resilient positioning, navigation, and timing (PNT) in GNSS-denied areas.

Quantum sensors, measuring the gravity and magnetic fields of the Earth, have now become accurate enough to make this practical. The reference maps they must match against have not kept pace, and that is where our work sits: building reference gravity and magnetic models good enough to position with, and understanding what such models need to deliver. Existing models were built to describe a surface, not the three-dimensional space that aircraft, vessels, and vehicles actually move through, so part of the task is extending them upward and downward and identifying what the available data can and cannot resolve? Answering it sets the requirements in both directions, what a reference model must achieve, and what accuracy and stability a sensor and its platform must reach to be useful.

We pursue this with partners across government, defence, and industry, and are training students with expertise in both geodesy and positioning.

Countries are rebuilding the frameworks that define where things are and how high they sit, moving from networks of physical survey marks to reference frames tied to satellites and gravity. It happens roughly once in a generation, and the science behind it is not settled.

We work on the transition itself: re-realizing legacy networks in modern geocentric frames, building the transformation models that let old and new coordinates coexist, and validating geoid-based height systems against GNSS and levelling where they are hardest to trust, mountains, coastlines, and land still rebounding from the last ice age. Modern frames are four-dimensional: the ground moves, so every coordinate carries a date, and keeping a network correct means modelling that motion rather than assuming it away. We also ask a question few have answered: how long a realization stays valid before it must be observed again.

Canada is a live case. NAD83 is being replaced by NATRF2022, the height system renewed on a new geoid model, and each province is responsible for its own adoption. We work with federal and provincial agencies through that change, and build methods that transfer to anywhere facing the same transition.