Anisotropism-based sea surface temperature forecasting method for high-current area

Abstract

The invention discloses an anisotropy-based sea surface temperature forecasting method for a high-current area. The method comprises the following steps: (1) giving an initial guess value of a controlvariable omega (x, y); (2) T is set as a research variable, and TS is obtained from a 'time' t = 0 forward integral advection diffusion model to t = S; obtaining a diffusion coefficient a by using aparameterization method; (3) performing integral tangent on the linear adjoint model to obtain a gradient g (omega) of the target function J (omega); (4) minimizing an algorithm optimization control variable omega (x, y); (5) circulating the steps (2) to (4) until the control variable omega (x, y) enables the target function J (omega) to reach the minimum value, and outputting the TS at the moment; and (6) establishing an analysis field according to the TS obtained in the step (5), wherein the established analysis field provides an initial field for forecasting the sea surface temperature of the strong current area. Aiming at the characteristic of isotropy of original diffusion filtering, a method of adding an advection term to a diffusion equation and parameterizing a diffusion coefficient is adopted, and the influence of ocean current is added to the whole assimilation process, so that an assimilation result is closer to a true value.

Description

technical field [0001] The invention relates to marine environmental engineering, in particular to a method for forecasting sea surface temperature in a strong current region based on anisotropy. Background technique [0002] Ocean numerical forecasting refers to solving the equations of fluid dynamics and thermodynamics describing the ocean evolution process through large-scale computer numerical calculations under certain initial and boundary value conditions based on the actual conditions of the ocean, and predicting the state of ocean movement for a certain period of time in the future. and ocean phenomena, that is, means of making ocean forecasts using current ocean conditions as input data. The most important aspect of how to obtain more accurate forecast results is the initial value problem. At present, the data assimilation method can obtain a more accurate initial field and improve the forecast accuracy. Usually, the standard three-dimensional variational data assim...

AI Technical Summary

Problems solved by technology

However, the determination of the diffusion coefficient in this process is relatively empirical, and this method is isotropic in nature, and the assimilation results in the strong current area are not ideal, which leads to poor forecasting results in the numerical prediction of the strong current area

Images