Mikael Kuusela (Department of Statistics and Data Science, Carnegie Mellon University)

Advances in Spatio-Temporal Analysis of Global Oceanographic Data from Argo Profiling Floats


 Argo floats measure seawater temperature and salinity in the upper 2,000 m of the global ocean. Statistical analysis of the resulting spatio-temporal data set is challenging due to its complex structure and large size. I propose mapping these data using locally stationary Gaussian process regression where covariance parameter estimation and spatio-temporal prediction are carried out in a moving-window fashion. This yields computationally tractable nonstationary anomaly fields without the need to explicitly model the nonstationary covariance structure. I also investigate a way to account for non-Gaussian heavy tails in Argo data using a Student-t distributed nugget effect. Cross-validation studies comparing the proposed approach with the existing state-of-the-art demonstrate clear improvements in point predictions and show that accounting for the nonstationarity and non-Gaussianity is crucial for obtaining well-calibrated pointwise uncertainties. I will then describe an uncertainty quantification method based on local conditional simulations which enables an extension from pointwise to spatially correlated uncertainties without the need for a global covariance model. These methods are valuable for obtaining improved observational estimates of ocean climate and dynamics. As an example, I will describe ongoing work on producing Argo-based estimates of key properties of the global climate system, including the ocean heat content, ocean circulation and ocean thermal response to tropical cyclones.

Refreshments at 3:45pm in Snedecor 2101.