35 Representativeness of Urban Highest Polluted Zones for Sitting Traffic-Oriented Air Monitoring Stations in a Chinese City∗ Francesca COSTABILE∗∗, Franco DESANTIS∗∗, Weimin HONG∗∗∗, Fenglei LIU∗∗∗, Rosamaria SALVATORI∗∗, Fenjuan WANG∗∗∗∗ and Ivo ALLEGRINI∗∗ Passive sampling technique was used to preliminary assess the spatial and temporal dis- tribution of air pollution in Suzhou (P.R. China) in 2003, with the aim to determine the rep- resentativeness of zones where to monitor air quality. 100 sites were selected to measure 15-days average concentrations of SO2,NOx,NO2,O3,NH3, Benzene, Toluene and Xylene in winter and summer. The distribution trend of NO and NO2, as well as the ratio of NO2/NOx was used to identify the heavy traffic polluted zones and design a network of monitoring sta- tions (macrositing). NO2/NOx ratios varied throughout the two selected seasons and from site to site clearly demonstrating the different extent and completeness of NO → NO2 oxidation processes in time and space. The exceedance frequencies of pollutants was used to design the type of monitoring equipment to be installed at each station (micrositing). An air quality monitoring network composed by 9 monitoring stations has been finally established in 2004 according to the result of this survey and is now currently running. Key Words: Air Pollution, Air Quality, Monitoring Network, NO2/NOx Ratio, Representa- tiveness emission factors (emission rates as a function of speed and 1. Introduction operating variables) were developed by the China National Traffic-induced air pollution in China first gained at- Mobile Source Emission Laboratory(5): because of Chi- tention in the early 1980s(1). The local pollution problem nese vehicular production technology, difficult road con- is apparently becoming severe(2). China has been long ditions, and the city’s driving cycle, emission factors for suffered from serious air pollution characterized by high vehicles are much higher than those in Europe and the ambient concentrations of SO2 and particles. The main United States. As a result of these factors and the rapid reason is that more than 75% of primary energy comes economic growth, most Chinese cities have suffered from from coal combustion. Since 1980, great efforts have been serious NOx pollution. Since 1990, NOx emissions from ff put to abate SO2 and particle emissions from coal com- stationary sources in China have levelled o somewhat, bustion. Ambient concentration, especially TSP, dropped but vehicular NOx emissions have increased. An increas- gradually(3). However, in the process of economic devel- ing effort is being now there devoted to understanding traf- opment within last decades, vehicular population has in- fic characteristic, emissions, and emission dispersion pro- creased dramatically. Pollution related with traffic exhaust cesses, and the relationship between these phenomena. is getting worse in some big Chinese cities. In 1995 the to- Currently, ambient air quality is routinely measured tal amount of vehicles in China were more than 20 million by a network of fixed-site monitoring stations in most with an average annual increase rate of 15%(4). Vehicle countries. The continuous monitoring is constrained by several factors, ranging from the cost and bulk of the mon- ∗ Received 4th October, 2005 (No. 05-5104) ∗∗ itoring equipment to power supply and maintenance re- Institute for Atmospheric Pollution, Italian National Re- quirements. Thus, the number of monitoring stations must search Council, via Salaria km.29, 300, 00016 Montero- be reduced as more as possible and their spatial represen- tondo, Rome, Ialy. E-mail: [email protected] ∗∗∗ Environmental Monitoring and Research Centre, 14 Sanx- tativeness becomes a key-point providing a basis for clas- iang Road 215004, Suzhou, P.R. China sifying air quality assessment. The United States Environ- ∗∗∗∗ Department of Environmental Engineering, BeiHang Uni- mental Protection Agency defined five categories of spa- versity, Beijing 100083, P.R. China tial scales — micro, middle, neighbourhood, urban, and JSME International Journal Series B, Vol. 49, No. 1, 2006 36 regional scales — in its guidelines for sitting State, Lo- cal and National Air Monitoring Stations. The European legislation suggests having stations for the monitoring of the population exposure in heavy traffic sites (Italian type C: traffic exposure) or in urban background (Italian type A: urban background exposure), but also considering the need to monitor sites where the exposure of population may be a problem (Italian type B: residential population exposure); in addition, it suggests that any Member States should set-up a proper background monitoring network for regional or national background evaluation of atmospheric Fig. 1 Suzhou population density by district expressed in number of person per square kilometre. The numbers pollution (Italian type D: non-urban background station). above each vertical bar indicate the surface area (square The issue of representativeness, however, has been over- kilometres) of each district(6) looked in the past; the lack of a quantitative method to describe the concept of representativeness results in in- consistent comparison for monitoring data among differ- publications of this laboratory(7), (8) no further information ent sites. are here reported. Sampling was carried out by mount- This paper reports findings on the representativeness ing the Analyst® samplers at heights of approximately 2.5 of urban highest polluted zones in Chinese urban environ- meters above ground level under a protective shelter with ments where to site traffic-oriented air monitoring stations. their open ends down to exclude dust, rain-sheltered, airy, These findings are taken from a pollution assessment car- beyond the kerbside. As a quality check, for all measured ried out in 2002 – 2004 during the course of the AQMS pollutants an additional sampler used as a field blank was project in the framework of the Sino-Italian Program, held deployed during each period of exposure. At the end of by the Italian Ministry of Environment and Territory and each period of exposure, samplers and blanks were col- the Chinese State Environment Protection Agency. Dur- lected, placed in airtight containers and stored until anal- ing this project an Air Quality Monitoring System was ysis carried out by ion chromatography for SO2,NH3, established by our Institute in the City of Suzhou (P.R. O3,NO2 and NOx and by gascromatography for benzene, China). In this work we treated the issue of representa- toluene and xylenes. tiveness as a spatial time-dependent problem due to the Sampling was carried out a 15-days basis over the pe- fact that site representativeness is determined by spatial riod from August to October 2003. The sampling protocol variation in measurements around monitoring sites over a was broadly based on methods and approaches described period of time and that air measurements were reported as in the EC document on guidance on preliminary assess- time-averaged concentrations. The urban area of Suzhou ment in relation to the Framework Directive(9), (10). Essen- was studied and passive sampling technique was used to tially, a grid over the area under investigation was con- assess air pollution. Particularly, the distribution trend of structed and for each cell of the grid a location as much NO and NO2 was used to identify the heavy traffic pol- as possible representative of the pollution level in the grid luted zones where to extend the measurement of the other was chosen. The accuracy of the Analyst® samplers com- pollutants by means of automatic monitoring technique. pared to the reference techniques expressed as percent rel- ative error is around ±20%. This is in compliance with the 2. Methodology Directive 1999/30/EC requirement for uncertainty (preci- Suzhou lies in P.R. China, in the centre of Yangtze sion and bias) < 25% for NO2,SO2 and NOx and within River Delta, south of Jiangsu Province, and borders on the recommended accuracy for diffusive sampling indi- Shanghai on the east. The total population is 5.9 mil- cated in the guidance on diffusive sampling as provided 2 (11) lion persons. The urban area covering 1 649.72 km with for NH3 and BTX by EC . In total 100 stations were 2.17 million persons(6) includes 8 districts: Canglang Dis- sited and set up throughout the area to provide a spatial trict, Pingjiang District, Jinchang District, Huqiu District, scheme that indicates the extent of local variability within New&Hi-tech, Industrial Park, Wuzhong District and Xi- the investigated area. The stations were equipped with angcheng District. The population, area and population Analyst passive samplers for the measurements of SO2, (6) density per district at the end of 2003 are reported in NOx,NO2,O3,NH3, BTX (benzene, toluene and xylene). Fig. 1. Due to the monsoon climate, summer is the rainy The sampling points were sited “on spot” into the 6 city period, with plenty rainfall and high temperature. districts (Fig. 2). We divided the whole city area in ur- Analyst® diffusive monitors (Marbaglass, Rome, ban, suburban and non-urban background zones depend- Italy) were used in this study. Since the development and ing on the population density distribution. Canlang, Pin- use of the Analyst® has been fully described in previous jiang and Jinchang districts have been considered as urban Series B, Vol. 49, No. 1, 2006 JSME International Journal 37 area; New&High Tech, Huqiu and Industrial park districts hot period. This campaign was replicated two times, in as suburban area; Wuzhong and Xiangcheng districts as order to avoid possible malfunctioning due to the atmo- non-urban background area. Inside every district the sam- spheric conditions. Considering the limited time-schedule pling points were selected in order to have the most con- available for the fulfilment of the project preliminary as- tinuous distribution according to the mesh of a geometric sessment and according to Suzhou’s climate, October was grid. Mesh dimension of urban area grid is smaller than chosen as representative of the dry-low windy campaign.