哈密市近10 a空气质量分析及预报模型研究

Study on Air Quality Analysis and Forecast Model of Hami City in Recent 10 Years

摘要: 利用2010~2019年哈密市气象及空气质量监测数据，分析了哈密市空气质量状况及其与降水、能见度、平均气温、风速、日照时数等气象因子的相关关系，采用多元逐步回归法建立四季空气质量指数预报模型并进行检验。结果表明：哈密市空气质量以优良为主，达标率为88.4%；2011年空气质量最好，污染日数最少，2014年轻度及以上污染日数最多，为空气质量最差年。近10 a春季空气质量最差，夏季最好。在不同季节，AQI指数与降水、能见度、气温均呈负相关关系，与风速、日照时数、气压呈正相关。建立了春、夏、秋、冬四季空气质量预报模型且均通过显著性检验；经检验，夏季模型预报效果最好，秋、冬季次之，春季较低，预报准确率都超过70%，可应用于哈密市空气质量预报业务。

Abstract: Using the meteorological and environmental data of Hami City from 2010 to 2019, the air quality status of Hami City and its correlation with meteorological factors such as precipitation, visibility, temperature, wind speed, and sunshine hours are analyzed. Using multiple stepwise regression method to establish and test the four season air quality prediction model. The results show that the air quality of Hami City was excellent, and the standard rate is 88.4%. Pollution days are the least in 2011, and the air quality was best. In 2014, the number of days of mild pollution and above was the most, and the air quality was worst. The air quality is the worst in Spring and the best in Summer in past 10 years. In different seasons, AQI index is negatively correlated with precipitation, visibility, and temperature, and positively correlated with wind speed, sunshine hours, and air pressure. The air quality forecast models in Spring, Summer, Autumn and Winter were established and passed the significance test. The results show that the Summer forecast model has the best forecast effect, followed by Autumn and Winter, and the Spring model has the lowest, forecast accuracy of more than 70%, which can be applied to the air quality forecast business in Hami City.