Theoretical Analysis of Predictability and Sensitivity for Ocean Circulation Models Based on Primitive Hydrodynamic Equations.

摘要

This project addressed the following aspects of ocean model predictability: (i) using the first passage time conception as the theoretical tool for the quantitative analysis of model predictability skill; (ii) calculating a low-order statistics of the prediction skill for the Princeton Oceanographic Model (POM) applied for modeling of shallow-water wind driven circulation in a semi-closed basin; (iii) comparison of numerical and analytical methods, such as the ensemble prediction technique (EPT) , singular vectors (SV) and iterative analytical solution of Pontryagin Kolmogorov equation (PKE); (iv) introducing a new measure (sensitivity indexes) for the sensitive analysis of numerical oceanographic models; (v) the analysis of contributions of different kinds of uncertainties such as errors in initial data, stochastic variations of wind and normal velocity along an open boundary into the losing of model predictability skill.