大气气溶胶和它的气体前体之间的非线性关系：用神经网络分析长期空气质量监测数据
Nonlinear relationships between atmospheric aerosol and its gaseous precursors: Analysis of long-term air quality monitoring data by means of neural networks
摘要：
城市环境气溶胶与挥发性有机物（VOC）和氮氧化物（NOx）浓度关系的非线性特征直接来源于NOx、VOC和总悬浮颗粒物质（PM）的长期常规测量数据。用于分析的方法的主要思想是创建基于人工神经网络的特殊经验模型。这些模型本质上是常用的线性统计模型的非线性扩展，被认为是在不同的大气条件下提供真实（非线性）PM-NOx-VOC关系的最佳拟合。据信，这种模式可能是有用的背景下，有关大气气溶胶的各种科学和实际问题。该方法是证明了两个经验模型的例子，在南海岸空气盆地，加利福尼亚州的两个空气质量监测站收集的独立的数据集。结果表明，尽管监测站之间的距离相当大（超过50公里），从而大大不同的环境条件，经验模型表现出几个共同的定性特征。具体地说，在一定条件下，NOx或VOC水平的降低可能导致气溶胶质量浓度的增加。据认为，这些功能是由羟基自由基的非线性依赖VOC和NOx。
英文摘要：
The nonlinear features of the relationships between concentrations of aerosol and volatile organic compounds (VOC) and oxides of nitrogen (NOx) in urban environments are derived directly from data of long-term routine measurements of NOx, VOC, and total suspended particulate matter (PM). The main idea of the method used for the analysis is creation of special empirical models based on artificial neural networks. These models which are in essence the nonlinear extension of commonly used linear statistical models are believed to provide the best fit for the real (nonlinear) PM-NOx-VOC relationships under different atmospheric conditions. It is believed that such models may be useful in context of various scientific and practical problems concerning atmospheric aerosols. The method is demonstrated by the example of two empirical models created with independent data-sets collected at two air quality monitoring stations at South Coast Air Basin, California. It is shown that in spite of considerable distance between the monitoring stations (more than 50 km) and thus substantially different environmental conditions, the empirical models manifest several common qualitative features. Specifically, it is found that, under definite conditions, the decrease of the level of NOx or VOC may lead to the increase of mass concentration of aerosol. It is argued that these features are caused by the nonlinear dependence of hydroxyl radical on VOC and NOx.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
