Field expeRiments for modEling, aSsimilatioN and adaptive sampLing
The coastal ocean shapes the two-way interaction between the deep ocean/ocean basins and the coastal populations and human societies. They determine how anthropogenic influences originating from the continents are redistributed, while impacting the maritime environment. Coastal ocean processes directly impact and influence how humans interact with the oceans, whether for civilian maritime needs such as fishing, recreation or extraction of minerals, or for security and dual-use needs related to monitoring and surveillance. It is critical for us to understand and ultimately predict the evolution of the different processes in this dynamic environment.
Yet our predictability and consequent understanding of this complex environment has been lagging in part because sufficient interdisciplinary studies across Biology and Physics have been lacking, in part because tools and methods have not been fully brought to bear on arguably a difficult domain to work in.
FRESNEL proposes to close the observe-assimilate-predict-sample loop by demonstrating the applicability of adaptively controlled marine robots in the aerial, surface and underwater domains, while sampling the upper water-column ‘at the right place and time,’ driven by ocean models with increasing predictive skill. In doing so, we wish to increase predictive skill of ocean models, leverage advances in Artificial Intelligence and decision-making, robotics, and bring to bear recent advances in Machine Learning for adaptation and prediction.
FRESNEL involves a diverse group of seasoned researchers working across traditional disciplinary boundaries. The tight integration between model prediction and assimilation that we foresee occurring as part of this effort will be enhanced so as to provide realistic forecasts of a range of biophysical variables, including temperature, salinity, wind, surface and subsurface currents, and bio-optical properties. These, in turn, will be used to intelligently target sampling with these multi-domain platforms in the air, ocean surface, and underwater, augmented by satellite remote sensing, including from a recently launched multi-spectral sensor on a Small Satellite.
The novelty of this proposed effort is in the integrative aspects of a field exercise, which will allow FRESNEL to leap-frog experimental design, autonomous operations, assimilation, modeling, and prediction in ways not done before. The project will outreach substantially with local authorities, subsistence fishermen, and an NGO in the domain of operation in Nazaré, Portugal, and engage locally.
FRESNEL is funded by the Office of Naval Research (ONR) under ONR award number N00014-22-1-2796
