Effects of Different Factors on the Visibility in Kaohsiung Area Using Hierarchical Regression

  • Chang-Gai LeeEmail author
  • Wen-Liang Lai
Conference paper
Part of the Advances in Intelligent Systems and Computing book series (AISC, volume 834)


The main purpose of this research is to investigate the impact of temperature, humidity, wind velocity, wind direction, etc. on aerosols and visibility. It then uses inferential statistics techniques, such as principal component analysis, stepwise regression analysis, and hierarchical regression, to examine the effects of various components as part of the secondary pollutants on the visibility of the Kaohsiung area. The research encompasses data from 1999 to 2004, where air-pollution-related data were obtained from QianZhen Air Monitoring station, while meteorology-related data were obtained from the Kaohsiung Station of Central Weather Bureau. The results show, during the particular period, that the average visibility of the Kaohsiung area was 6.4 ± 3.6 km and the visibility, on average, increased by 0.33 km each year. Annual average of PM2.5 concentration dropped 0.79 μg/m3 each year. During this particular period, 68.74% of the time was under haze conditions while 69.83% of the time the visibility was lower than 10 km. The results show that low visibility of the Kaohsiung area is primarily caused by haze. Some average concentrations during the period are listed herein, PM2.5 = 35.93 μg/m3, nitrate = 4.10 μg/m3(i.e.,11.40% of PM2.5), sulfate = 9.59 μg/m3(constitute 26.68% of PM2.5), organic carbon = 8.49 μg/m3(i.e., 23.64% of PM2.5), and elemental carbon = 2.64 μg/m3(i.e.,6.58% of PM2.5). The nitrate amount and nitrate-to-sulfate ratio of PM2.5 are 11.40% and 0.43, respectively, which are lower than those found in previous researches in Kaohsiung area. During the same period, the monthly average OC/EC values were all more than 2.2, indicating there was a great potential of the formation of secondary aerosols in the Kaohsiung area, which was also confirmed by the calculated secondary organic carbon (SOC) and the monthly average of SOC/OC exceeding 85%. In addition, a model was built via multivariate linear regression and hierarchical regression to investigate the factors that can potentially affect the visibility of the Kaohsiung area. It has been found that the regional visibility is more or less inversely correlated to pollutants such as sulfate, nitrate, elemental carbon, organic carbon, NOx, NH3, PM2.5, and PM10. All the potential factors, categorized into air pollutants and climate properties, were further analyzed using principal component analysis, which reveals the two categories belonging to two different groups. Based on multivariate linear regression analysis results, as of climate properties, only humidity has a large effect on the way how air pollutants affect/interact visibility. Accordingly, a model built via hierarchical regression found that the visibility in the Kaohsiung area is mainly impacted by nitrate (−0.62), followed by elemental carbon (−0.41) and then sulfate (−0.30). This implies that the visibility attenuation in Kaohsiung area could be mainly caused by photochemical reaction as well as combustion. Through path analysis, it is also discovered that the regional visibility in summer time is higher than that in winter time, which is likely due to different photochemical reaction rates caused by temperature variation.


Visibility Aerosol Big data Hierarchical regression Path analysis 


  1. 1.
    Tsai, Y. I., Lin, Y. H., Lee, S. Z.: Visibility variation with air qualities in the metropolitan area in Southern Taiwan. Water Air Soil Pollut. 144, 19–40 (2003)Google Scholar
  2. 2.
    Chan, Y. C., Simpson, R. W., Mctainsh, G. H., Vowles, P. D., Cohen, D. D., Bailey, G. M.: Source apportionment of visibility degradation problems in Brisbane (Australia) – using the multiple linear regression techniques. Atmos. Environ. 33, 3239–3250 (1999)CrossRefGoogle Scholar

Copyright information

© Springer Nature Singapore Pte Ltd. 2019

Authors and Affiliations

  1. 1.Department of TourismTajen UniversityPingtung CountyTaiwan, ROC
  2. 2.Graduate Institute of Environmental Management, Tajen UniversityPingtung CountyTaiwan, ROC

Personalised recommendations