2021-11-30[Event Report]DA Seminar on November 29, 2021
Prof. Hirose gave a talk on Green’s function for parameter tuning in an ocean model and Kalman filter-based ocean data assimilation. Although a coastal ocean DA system shows relatively high correlation coefficients relative to in-situ ADCP velocity observations, there remains room for improvement. He presented that Green’s function makes significant improvements by estimating appropriate parameters especially in tidal forcing and vertical viscosity.
To reduce the computational cost, he applies a reduced-order approximation proposed by Fukumori et al. (1995) to the Kalman filter. Sensitivity experiments on the structured and unstructured grid points concerning the coastal line demonstrated better dynamical consistency and numerical stability in the unstructured gridded experiments. He also investigated the impacts from the perturbed atmospheric and lateral boundary forcing. Finally, he presented a collaborative project between researchers, industrial companies, and fisheries. They invented a new CTD instrument for fishermen easy to handle, in-situ temperature and salinity profiles measured by fishermen significantly increase the accuracy of analyses and predictions, and the improved analyses and predictions are provided to fisheries through android application.