The last two decades have seen tremendous growth in the range of
geodynamically relevant observations, including, but not limited to
detailed tomographic imaging of the deep Earth from the sublithosphere
to the CMB to provide state estimates of the convective system, noise
mitigated plate motion reconstructions that reveal temporal velocity
variations of tectonic plates much faster than the mantle transit time,
stratigraphic analysis to expose systematic changes at the continental
scale in the distribution of hiatus at the level of series and stages,
and paleomagnetic investigations that map rapid bursts in the motion
of some hotspots, all together documenting the relentless convective
activity of Earth's mantle and its impact throughout the Earth system.
At the same time powerful geodynamic inversion techniques based on the
adjoint and other optimisation methods have become available to link
models and data quantitatively through variational and sequential
assimilation schemes of observations into geodynamic simulations. The
simultaneous growth of observations and theoretical capabilities
provides us with unprecedented opportunity to test the underlying
assumptions of dynamic Earth models. This transdisciplinary session
brings together observational and theoretical scientists to discuss
the scope and format of established and nascent convection related
observables, and welcomes contributions that highlight the noisy
nature of observables while exploring methods to handle the impact
of uncertainty in the geodynamic data assimilation framework.