DFT System and Its Applications on Variational Four-dimensional Data Assimilation
by
Walls
LASG,Institute of Atmospheric Physics,Chinese Academia Sinica
Coauthors: Wang Bin (LASG, Institute of Atmospheric Physics,Chinese Academia Sinica), Xie Xing (Department of Computer Science, University of Science and Technology of China, Hefei, Anhui,China, 230000)

Abstract

The technique of variational 4D data assimilation which couples data with physical model in the process of numerical integration is fairly concerned in meteorological and oceanographical observations in recent years. To construct such a data assimilation system, the tangent linear (TL) model and adjoint model should be accurately produced. Here by use of object-oriented design methodology, we have developed a successful code-to-code system that can automatically transform Fortran source code into its correspondent TL code according to the chain rules. This system is divided into three major parts: Software system, Users system and Evaluation system. Besides its high productivity and easy-to-use property, the reliability is well tested in several typical applications such as GPS/MET rayshooting model, IAP model, SSI model, etc. The twin system, the adjoint code generator (ADG), is still under its way. In addition, we take statistical accuracy rate (SAR) is one of the most important factors in evaluating such a class of AD tools.

Keywords: data variational assimilation, tangent linear model, adjoint model, object-oriented design methodology, SAR evaluation

Date received: December 26, 1999

Copyright © 1999 by the author(s). The author(s) of this document and the organizers of the conference have granted their consent to include this abstract in Atlas Conferences Inc. Document # cads-25.