变分伴随方法应用于大气化学初值和参数反演研究

Retrieval of the Initial Values and Parameters in Atmospheric Chemistry by the Variational Adjoint Method

将结合正则化思想的变分伴随方法应用于大气平流层中臭氧破坏催化反应过程的浓度初始值和反应速率参数的反演问题,推导了在整体资料和末端时刻局部资料反演的伴随模式和泛函梯度,并进行了一组理想数值试验。试验结果表明了变分伴随方法运用于大气化学过程的反演研究的有效性,特别是局部资料联合反演中可以运用正则化思想克服反演的不适定性。

The variational adjoint method in combination with Tikhonov regularization principle is applied to retrieve the initial gas concentrations and reaction rates in chemical reactions about ozone destroying in the stratosphere. The adjoint models and functional gradients are deduced for the complete and incomplete（terminal time） observations of the gases＇ concentrations, respectively. Finally, a series of ideal numerical experiments are preformed to show the validity of the variational adjoint method. Especially,Tikhonov regularization principle can be used to overcome the ill-posedness in the case of combinational retrieving for incomplete observations.