Application of validation data tests from an on-line volatile organic compound analyser to the detection of air pollution episodes in urban areas

This paper presents the routine and validation tests developed to evaluate 62 volatile organic compound (VOC) measurements (C2–C10), in ambient air from an online VOC analyser. These tests are based on experience obtained from application on hourly data sets obtained during April–October campaigns (1997–2001) at the Air Quality Monitoring Station sited at the School of Engineering, Bilbao (SEB) (Northern Spain). Every campaign yields 4000–4800 uninterrupted running hours, which amounts to about 1,200,000 valid data over the five intensive measuring campaigns. Quality control through data validation is critical to minimise the amount of faulty data reported, because erroneous individual data values give place to serious errors in data analysis results. These tests also allow the detection of unexpected concentration values of some pollutants. Air pollution episodes detected during data validation and meteorological influence analysis are presented. VOC data analysis together with complementary meteorological data may explain the main observed trends, especially for some of the VOCs. Acetylene and benzene are two clear cases: commonly used in the literature as vehicle emission tracers, the analysis show that, in many cases, they must be attributed to urban or industrial sources.