MEASUREMENTS AND ANALYSIS OF CONCENTRATIONS OF GASEOUS-HYDROGEN PEROXIDE AND RELATED SPECIES IN THE RURAL CENTRAL PIEDMONT REGION OF NORTH-CAROLINA

Ambient concentrations of gas-phase total peroxides and hydrogen peroxide were measured during mid-July to mid-August 1991, and during June and early July, 1992 at a rural site, SONIA, in the rural Central Piedmont region of North Carolina as a part of the Southern Oxidants Study. Simultaneous measurements were also made of trace gases and meteorological parameters. The measurements of hydrogen peroxide were made in order to explore the behaviour of this photochemical oxidant in rural southeast U.S., and to study its correlation with ozone and meteorology. H2O2 showed a pronounced diurnal variation with peak concentrations during the afternoon (1200-1600 EST). The maximum H2O2 concentrations observed were approximately 2.2 ppbv during the 1991 intensive, and approximately 1.6 ppbv during 1992. The mean daytime H2O2 concentrations observed were 0.64 +/- 0.56 ppbv during 1991, and 0.52 +/- 0.36 ppbv during 1992, which were found to be significantly higher (at the 95% confidence level) than the nighttime averages (approximately 0.15 ppbv). HO2 radical concentration of approximately 14 ppbv was determined from calculated H2O2 production rates. On occasion high levels of nighttime H2O2 was observed which was found to be associated with the breakdown in the stability of the nocturnal boundary layer. A 48 h back trajectory analysis performed on the air masses arriving at Site SONIA indicated that significant amounts of pollutants could be transported to this rural site from the surrounding urban areas and this could affect the H2O2 levels at this site. An observational based statistical analysis was performed. The results of correlation matrices show that H2O2 is most strongly correlated to ozone, temperature, solar radiation, and relative humidity. A multiple linear regression was carried out by regressing H2O2 on all the measured physiocochemical variables. The R2 was found to be 0.81 for the 1991 data and 0.71 for 1992. A simple linear regression between H2O2 and ozone gave R2 of 0.60 and 0.39 for 1991 and 1992, respectively, when data from the entire period of measurement was included.