Abstract: Based on the atmospheric chemistry transport model CAMx, its adjoint model is constructed. The adjoint model code is hand-generated and numerical experiments are designed to test and validate the code. CAMx and its adjoint model are applied for simulation and sensitivity analysis of air pollution in the Pearl River Delta. The sensitivities of high ground level sulfur dioxide and ozone respect to pollutant sources are calculated, and the applicability of linear sensitivity coefficients is discussed through numerical experiments, which provide important information for extensive analysis of pollution mechanisms and control strategies. Using the adjoint model, the sensitivities of an object function respect to thousands of input variables can be calculated efficiently. With the introduction of the adjoint method, the function of CAMx is considerably extended, and provides a powerful tool for inverse problems on emission parameters as well as for management of atmospheric environment.

Key words: CAMx model, adjoint method, atmospheric pollution, sensitivity analysis

摘要： 在大气化学传输模式CAMx的基础上，用人工编写代码的方法，构建其伴随模式，并通过数值试验，对伴随模式程序进行了检验。应用CAMx及其伴随模式，针对珠江三角洲地区空气污染进行数值模拟和敏感性分析，计算了地面二氧化硫和臭氧关于污染物排放源的敏感性，并用数值方法探讨了线性敏感系数的适用范围，可为深入研究该地区污染机理和控制策略提供重要依据。用伴随模式可以高效率地计算目标函数关于多个输入变量的敏感性，显著地扩展了原模式的功能，为大气环境优化控制和大气化学数据同化等科学问题提供了有力的研究工具。

关键词: CAMx模式, 伴随方法, 大气污染, 敏感性分析