基于数值预报的空气质量预测模型的研究

发布时间：2017-12-31 04:34

本文关键词：基于数值预报的空气质量预测模型的研究　出处：《浙江理工大学》2017年硕士论文　论文类型：学位论文

【摘要】：近年来,人类工业化进程的不断加快加剧了环境的恶化程度,随着区域重污染天气特别是重度雾霾天气的不断出现,空气质量问题已经成为社会热点话题之一。当前,我国的大气污染已经由过去单一类型的污染逐步转变为多种污染物、多种作用机制同时存在、而且相互影响的复合型大气污染。复杂的大气污染问题加大了空气质量预测预报的难度,传统单一的预测预报方式已经不能准确的预测空气质量状况。本文基于空气质量数值模拟预报技术,进行区域空气质量预测预报的研究。提出并且设计了一个基于数值预报的空气质量预测模型:多模式-集成修正预测模型。模型采用高度模块化设计思想,气象场资料模块根据浙江省区域气象特点设置气象模式软件MM5与WRF特定的参数方案,作为空气质量数值模式的气象场输入;污染源本地化模块定制浙江省区域的污染源清单数据,经过源排放输入文件的处理程序SMOKE处理后作为空气质量数值模式的污染源清单输入;空气质量多模式运行模块实现包括WRF-CHEM、WRF-CMAQ、MM5-CMAQ和MM5-CAMx四种空气质量数值模式的并行运行;集成-修正模块对多模式运行结果进行集成修正,该模块包括多维度动态权重更新模型与误差修正模型,多维度动态权重更新模型通过算法来动态调整各个模式的各类污染物在集成预报中权重因子,误差修正模块通过算法来计算集成预报中的误差大小,对预测结果进行修正。基于多模式-集成修正模型,构建浙江省空气质量预测预报业务系统。通过将监测数据与预测数据进行特定的参数评估分析,发现此模型在区域空气质量预报方面具有良好的准确度。实验结果表明采用算法控制模式预报结果的方式可以充分利用各个模式的优点,做到预报结果的互补,比单个模式、简单的平均集成结果的预测结果更加准确、可靠。
[Abstract]:In recent years, the accelerating process of human industrialization has aggravated the degree of environmental degradation, with the emergence of regional heavy pollution weather, especially severe haze weather. Air quality has become one of the hot topics in our society. At present, air pollution in China has gradually changed from a single type of pollution to a variety of pollutants, and a variety of mechanisms exist at the same time. The complex problem of air pollution increases the difficulty of air quality prediction and prediction. The traditional single forecasting method can not predict the air quality accurately. This paper is based on the numerical simulation and prediction technology of air quality. An air quality prediction model based on numerical prediction is proposed and designed. The model is multi-model-integrated modified forecasting model. The model adopts highly modular design idea. According to the regional meteorological characteristics of Zhejiang Province, the meteorological field data module sets up the specific parameter scheme of meteorological model software MM5 and WRF, which can be used as the meteorological field input of the numerical model of air quality. Pollution source localization module customizes the pollution source inventory data in Zhejiang province, and after processing the source emission input file by SMOKE, it is used as the air quality numerical model of the pollution source list input. The air quality multi-mode operation module includes WRF-CHEMN WRF-CMAQ MM5-CMAQ and MM5-CAMx. The integration-correction module integrates the results of multi-mode operation. The module includes a multi-dimensional dynamic weight updating model and an error correction model. Multi-dimensional dynamic weight updating model dynamically adjusts the weight factor of various pollutants in the integrated forecast by the algorithm, and the error correction module calculates the error size of the integrated forecast by the algorithm. Based on the multi-model-integrated correction model, the operational system of air quality prediction and prediction in Zhejiang Province is constructed. The monitoring data and the forecast data are analyzed by the specific parameter evaluation. It is found that the model has good accuracy in regional air quality prediction. The experimental results show that the algorithm control model forecast results can make full use of the advantages of each model, and achieve the complementary prediction results. Simple average integration results are more accurate and reliable than individual models.
【学位授予单位】：浙江理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：X51

【参考文献】