Monitoring Of Atmosphere Background Bioaerosol By Fluorescent Spectrometry And Its Evaluation

Bioaerosol is widely distributed in the atmosphere and plays a very important role in ecosystems,climate change and public health security.The monitoring,analysis and forecasting of atmosphere background bioaerosol concentration is a significant basic research project.In this research,the Fluorescent Data Acquisition Instrument(FDAI)was used to background bioaerosol concentration’s long-term monitoring in Changsha(capital of Hunan province,China).Based on the monitoring study,the relationship between bioaerosol concentration and meteorological factors was analyzed.A BP neural network model was used to forecast bioaerosol concentration by meteorological factors.After then,the bioaerosol concentration forecasting model was optimized by wavelet transform and particle swarm optimization algorithm.The main results of this study were listed as below:(1)The background bioaerosol concentration was monitored by FDAI in Changsha for the first time,monitoring result displayed that atmosphere background bioaerosol was with a concentration range of 5-173 pcs/L and an average value of 42.35 pcs/L,which was slightly higher than bioaerosol concentration of most other urban areas and rural areas in other reported studies.The ambient measurements illustrated that bioaerosol concentration varied dramatically in Changsha showed strong daily and seasonal cycle characteristics.The highest bioaerosol concentration occurred at 7:00,and lowest bioaerosol concentration was during daytime,between 13:00 and 15:00.Bioaerosol concentration was higher in autumn and lower in winter.(2)There are strong correlation between bioaerosol concentration and meteorological factors,such as wind direction,temperature,relative humidity,vapor pressure,dew temperature,air pressure,particulate matter et al.The influence of wind direction and wind speed to bioaerosol is mainly due to the affection of aerosol diffusion,and the wind speed can affect the release of spores;The influence of temperature to bioaerosol is mainly due to the phenomenon of advection inversion,which makes bioaerosol not easy to diffuse;The influence of humidity to bioaerosol is mainly due to the increasing of humidity will promote spores release;The influence of air pressure to bioaerosol is mainly due to the air pressure affects the atmospheric circulation,and the diffusion of bioaerosol is accompanied by atmospheric circulation.(3)A BP neural network model was used to forecast bioaerosol concentration and the error of model reached the application accuracy demanded.It can be found that there were obvious internal correlations among meteorological factors.Principal Component Analysis can eliminate the error caused by multicollinearity of meteorological factors in BP neural network.The BP neural network model has achieved good results in forecasting bioaerosol concentration.The forecasting value curve and the actual measured curve fit very well,which almost overlap each other.The average relative error is 10.55%,the average absolute error is 2.80 pcs/L,the grade of forecast accuracy is 84.01.The error of the model are acceptable,which means the model is promising for forecasting bioaerosol concentration.(4)Wavelet De-noising-based Back Propagation(WD-BP)neural network model and Particle Swarm Optimization-based Back Propagation(PSO-BP)neural network model were used to forecast bioaerosol concentration,with an average relative error of8.75% / 6.03%,an average absolute error of 1.22 pcs/L / 0.72 pcs/L,and a grade of forecast accuracy 89.21 / 89.53.The forecasting performance of WD-BP / PSO-BP neural network is higher than that of single BP neural network.In a word,the monitoring of atmosphere background bioaerosol and its evaluation were investigated in this study.It provides a foundation for monitoring and forecasting of abnormal changes of bioaerosol concentration,from which the public capabilities of biosecurity risk prevention can be sharply improved.