摘要
空气质量预测在国内的关注度日益提高,传统的空气质量预测系统通常运用数值化学传输模型,利用物理方程来计算污染物的扩散、沉降和化学反应。而化学传输模型的预测准确性很大程度上需要依赖详细的污染源排放信息和气象模型的输出结果。基于统计模型的OPAQ空气质量预报业务系统,采用人工神经网络算法,可预测各污染物的日均值或日最大值。并对北京空气质量预报的结果进行了评价,OPAQ空气质量预报业务系统对空气质量预测的准确性较高,能够利用较低的计算资源得到较为准确的预测结果。与数值预报相比,OPAQ空气质量预报业务系统不需要大量的基础数据作为输入,可弥补数值预报的不足,并成为数值预报的有力补充。
Air quality forecast has been and continues to be a growing concern in China. Traditional air quality forecasting systems employ deterministic chemical transport models in which pollutant dispersion,deposition and chemical reactions are computed from the governing physical equations. Detailed pollutant emission information and meteorological model output areneeded to run the chemical transport models. An alternative statistical forecasting system OPAQ is discussed that is based on the BP neural network algorithm. The results of the OPAQ forecast system for a nuh Paer of monitoring stations in Beijing show that forecast accuracy of OPAQ is high and it can take advantage of lower computing resources to predict air quality. Compared with deterministic chemical transport model,it doesn't need multiple input data and can be a complementary approach of traditional method.
出处
《中国环境监测》
CAS
CSCD
北大核心
2016年第3期13-20,共8页
Environmental Monitoring in China
基金
环保公益性行业专项“京津冀地区大气重污染过程应急方案研究”(201309071)
“京津冀城市大气边界层过程对重污染形成的影响研究”(201409001-03)
科技部科技支撑计划环境领域项目“大气复合污染区域联合预测预报关键技术研究”(2014BAC22B04)
“京津冀空气监测预报及防控技术研究与示范”(2014BAC06B04)
中科院先导项目“大气灰霾追因与控制专项数值模式与协同控制方案课题”(XDB05030200)