Particulate Matter (PM10) Sources in Bizerte, Tunisia Using Logistic Regression Model

Abstract

Short-term exposure to particulate matter (PM10), the total mass of PM with a diameter of 10 micrometers or less, has been linked to increased morbidity and mortality. Most of the dust particles released to the atmosphere worldwide is emitted by Saharan Desert in North Africa. This paper evaluates the effect of dust particles from African regions on PM10 concentrations in Bizerte, Tunisia using a logistic regression statistical methodology. The 2004 and 2005 PM10 measurements were obtained by the Tunisian National Agency for Environmental Protection, ANPE, pollution monitoring station and for the same period, 6-day back trajectories were generated using the NOAA-HYSPLIT. The study found that the DLV exceeded the 50-μg/m³ limit 293 times in 2004 and 257 times in 2005. Significant associations have been found between far air sources and the measured concentrations for both 2004 (OR=1.024, 95% CI = 1.016-1.032) and 2005 (OR=1.013, 95% CI = 1.003-1.022). The model performance has significantly improved when shorter back trajectories (closer source regions) were considered. The feasibility of using a logistic regression statistical methodology to evaluate the effect of air masses from African regions on PM10 concentrations in Bizerte, Tunisia has been demonstrated.