Objectives  To explore the trend of visibility in the winter in Beijing and its influencing factors.

Methods  The annual, monthly, daily and hourly variation of visibility was analyzed by performing Spearman's correlation analysis and partial correlation was analyzed with SPSS to investigate the relationship between the visibility with temperature, relative humidity, atmospheric pressure and wind speed. Finally, a multiple linear regression equation for visibility was constructed by using R language.

Results  The average visibility in the winter of 2008-2013 in Beijing was 16~21 km. The highest visibility was in 2010-2011 and the lowest was in 2012-2013. However, there was no obvious regular change of monthly average visibility. The visibility observed hourly in a descending order was at:6:00, 12:00, 0:00 and 18:00. The visibility was positively correlated with wind speed, temperature and atmospheric pressure with a partial correlation coefficient of 0.429, 0.085 and 0.320 respectively. The visibility was negatively correlated with PM_2.5 and relative humidity with a partial correlation coefficient of -0.379 and -0.532 respectively. The multiple linear regression equation was Y=-255.0+19.35X₁-0.26X₂+2.73X₃ (Y:visibility; X₁:wind speed; X₂:relative humidity; X₃:barometric pressure).

Conclusions  The trend of variation and key influencing factors for visibility in winter in Beijing were obtained in this study. The results provided theoretical basis for the forecast of visibility and the early warning of health-related risks.