Sensitivity Analysis of Multidimensional Air Qualtiy Models Development and Application of a Direct Technique to the Mexico-Southwestern United States Border Region

该提案的主要重点是开发一种计算有效的灵敏度分析技术，用于多维空气质量模型，并将该技术应用于美国和墨西哥之间的边境地区。作为该应用的一部分，该研究将量化源影响并开发各种模型参数（例如沉积速度和化学速率常数）的敏感性系数。由于墨西哥和美国西南部边境地区的气候以及快速的经济增长和工业化，该地区的臭氧含量超过了两国的国家标准。解决该地区的空气质量问题特别令人感兴趣，原因有几个：首先，问题的国际性质，以及预期的跨界污染物流动，以及最近的贸易协定（例如北美自由贸易协定和马奎拉多拉斯）；二是人口增长过快；第三，不同地区的控制水平差异很大。如果能够通过应用敏感性分析充分证明源与空气质量的关系，就有可能为该区域的空气污染控制制定更具成本效益的综合战略。在直接解耦灵敏度方法的基础上，提出了一种计算上具有吸引力的多维空气质量模型灵敏度分析技术。由此产生的方法将以这样一种方式进行设计和构造，以便在目前普遍应用的空气质量模型之间易于传输，并正在开发以供将来使用。此外，该系统将利用并行计算环境。该方法将应用于沿墨西哥-美国西南边境的空气质量建模，以更好地了解该地区的污染物动态，并量化模型对输入和模型参数的响应。最后，作为这种技术的力量和用途的一个例子，该方法将用于解决该地区感兴趣的空气质量工程中三个重要的具体问题：探索使用形式敏感性分析进行模型评估和相互比较，量化臭氧浓度如何响应各种化合物和来源的排放，并进一步探索灵敏度分析在研究挥发性有机化合物（VOC）反应性尺度中的应用。该项目的预期结果是：开发一种计算效率高的空气质量建模灵敏度方法；敏感性分析方法在墨西哥-美国沿岸的应用。为了更好地了解该地区空气污染物的动态，对边界地区应用的模型敏感性和不确定性进行评估，以及佐治亚理工学院和蒙特雷高等理工学院（ITESM）之间日益增长的关系，这不仅是两位主要研究人员之间的直接合作，也是由于佐治亚理工学院一名墨西哥学生（来自ITESM）的高等教育。

9613729 Russell The major focus of this proposal is to develop a computationally effective sensitivity analysis technique for use in multidimensional air quality models, and to apply that technique to the border region between the US and Mexico. As part of that application, the research will quantify source impacts and develop sensitivity coefficients for various model parameters (e.g. deposition velocities and chemical rate constants). Because of its climate and rapid growth and industrialization, the Mexico-southwestern United States border area experiences elevated levels of ozone exceeding the national standards of the two countries. Addressing this air quality problem in the region is of particular interest for several reasons: first, the international nature of the problem, and the expected trans boundary pollutant flow, along with recent trade agreements (e.g. NAFTA and Maquiladoras); second, the high population growth in the area; and, third, the level of controls which differ widely between the various regions. If one can adequately demonstrate the source-air quality relationships through the application of sensitivity analysis, it is possible to develop much more cost effective integrated strategies for air pollution control in the region. It is proposed to develop a computationally attractive sensitivity analysis technique for use in multidimensional air quality models based on the direct decoupled sensitivity method. The resulting method will be designed and constructed in such a way as to be easily transportable between air quality models that are in common application today, and are being developed for future use. Furthermore, the system will make use of parallel computing environments. The method will be applied to air quality modeling along the Mexico-southwestem U.S. border to better understand pollutant dynamics in that region, and to quantify the model responses to inputs and model parameters. Finally, as an example of the power and uses of such techniques, the method will be applied to address three specific issues of importance in air quality engineering of interest in that region: explore the use of formal sensitivity analysis for model evaluation and intercomparison, quantify how ozone concentrations respond to emissions of various compounds and sources, and to further explore the use of sensitivity analysis in studying volatile organic compound (VOC) reactivity scales. The expected results from this project are: development of a computationally efficient sensitivity method for air quality modeling; application of the sensitivity analysis method along the Mexico-U.S. border to better understand the dynamics of air pollutants in that region, assessment of model sensitivities and uncertainties in the application along the border region, and a growing relationship between Georgia Tech and Instituto Technologico Estudios Superiores de Monterrey (ITESM) due both to the direct collaboration between the two principal investigators, but also due to the advanced education of a Mexican student (from ITESM) at Georgia Tech.