Air pollution in a coastal area: transport and deposition of photochemical oxidants T. Georgiadis^, A. Baroncelli", P. Bonasoni^, G. Giovanelli^, F. Ravegnani^, F. Fortezza\ L. Alberti\ V. Strocchi^ 7m **PMP - USL35 Environ. Phys. Laboratory, Via Alberoni 17, 48100 Ravenna, Italy Abstract Monitoring performed at the environmental network stations of Ravenna, a city along northern Italy's Adriatic seaboard, shows that the breeze-driven transport of air masses markedly contributes to the pollution levels. The time evolution of photochemical oxidants along the coast is strongly influenced by the breeze circulation in summer, when primary pollutants are transported offshore by the land-breeze, undergo photochemical transformations at sea and are driven back inshore by the sea breeze, resulting in high concentration values of photochemical oxidants at the ground. A mass-budget approach was applied to determine the roles played by chemistry and deposition in the ozone evolution of the survey area. The characteristics of the sea-breeze front were used to evaluate the vertical structure and the height of the ozone layer. 1 Introduction Several studies have been conducted over the last few years in coastal areas to determine the evolution and distribution of primary pollutants, their reactionproducts and the concentration levels of oxidants produced by photochemical chains. Local concentration values and their diurnal variations are highly dependent on solar radiation, oxidant precursor emissions and micrometeorological conditions. In addition, topographical discontinuities like coastlines can induce peculiar wind circulation systems that transport pollutants. Geographically and climatically part of the Po Valley, greater Ravenna is bordered on the east by the upper Adriatic Sea and features an extensive seaboard industrial belt with petrochemical and power plants that produce large amounts of chemical emissions. The time evolution of photochemical oxidants is strongly influenced by small-scale transport due to land and sea breezes,
418 Pollution Control and Monitoring especially during summer. The morning land breeze drives the factory-emitted compounds offshore, where the high intensity of solar radiation induces the formation of oxidants by photoreactions, and after mid-day the sea-breeze transports these oxidants inshore. Such varying topographical features as the sea-land discontinuity, with their differing chemical and physical properties, result in markedly differentiated deposition processes of relative intensities and often lead to fumigation episodes exceeding, for ozone concentration, national health thresholds. The present study evaluates the distribution characteristics of oxidant pollution at the survey site along with the physical processes governing their transport and deposition. 2 Climatic and atmospheric patterns of the Ravenna area The greater Ravenna is situated at the eastern part of the Po Valley on the Adriatic Sea. The area's climate is generally tempered by the Alps, a barrier that mitigates the effects of northern winds. Winters are marked by rain, high relative humidity and long foggy periods and summers by high solar radiation intensity, temperature and a local land-sea breeze circulation coincident with the anticyclonic conditions; stationary atmospheric inversions are commonly found throughout the year. The breeze system's range of action, which extends 30-40 km offshore, is reinforced by the notable shallowness of the western Adriatic Sea. Furthermore, the nearby Apennines induce a peculiar overlapping of the mountain-lowland and lowland-sea basin that strengthens the breeze system, a fact also reported by other authors [1,2]. Breeze reversal, generally occurring late between 11:00 a.m.-12:00 p.m. local time, is one of the phenomena that can produce a largescale circulation in which air masses recirculate horizontally for several dozen kilometres [3,4]. For about 40 km (half of which onshore) along the path of the locally prevailing WNW-ESE winds lie four neighbouring areas of diverse physical and chemical features: the city, the industrial belt, the shoreline with a strip of pine stand and the sea basin. These varying topographical features, along with the peculiar meteorological conditions, induce differentiated effects on the offshore-to-onshore breeze, modifying it by changes in roughness, which produce a mechanical internal boundary layer, and by temperature contrasts, which produce a thermal internal boundary layer (TIBL), with subsequent alteration of pollutant dispersion and deposition. 3 Experimental Systematic summer measurements of ozone have been conducted since 1978 to assess the evolution of photochemical oxidants in the greater Ravenna area. The environmental network managed by the Public Health Service comprises five permanent monitoring stations equipped with Philips PW9771 analysers for Cb; additional analysers of NMHC (Byron 301) and NOx (Philips PW9760) were available at the stations for shorter periods.
Pollution Control and Monitoring 419 The stations are situated along an 8-km path running from the city in a roughly perpendicular line to the shore; the occasional use of an offshore oil-rig platform for mounting an additional monitoring Oa station brought the total path to a length of 30 km. Two other monitoring stations were equipped with meteorological sensors for solar radiation intensity, wind speed and direction, temperature and relative humidity. A 60-m tower standing close to the Candiano channel in the industrial area measured wind speed, temperature and relative humidity at three heights. Vertical soundings of atmospheric parameters by balloon-borne meteorological sensors, 'pilot' balloons and electronic theodolites were conducted in 1978-1982 and in 1989 to determine the general breeze circulation at the site. The offshore ozone concentration was measured in 1981 aboard the oceanographic vessel Daphne of the Marine Biology Research Centre of the Emilia-Romagna Region. A summer 1982 monitoring survey was conducted by aircraft-borne instrumentation to assay the vertical layering of ozone concentration both onshore and offshore using a Dasibi 1003 RS ozone analyser. 4 Results of the experimental surveys and discussion The data recorded during well defined breeze circulation along the monitoring path are reported in figure 1 as typical diurnal variation of ozone concentration. They indicate that (a) the correlation of ozone concentration trends with solar radiation at the industrial and coastal sites is more marked than at the offshore station; (b) the 24-hour values registered at the offshore station were higher than at the others and the overnight ones exceeded by several-fold those at latter; (c) the rising slopes of the industrial and coastal values evince a phasedrift of about one hour; (d) the onshore ozone concentration trend shows a second peak in the late evening. These results make it possible to draw an overall picture of the diurnal evolution of ozone concentration. During the early morning hours air-masses rich in primary pollutants are driven offshore by the land-breeze while at the same time the increasing solar radiation intensity enables the formation of photochemical ozone inland. The light winds occurring between 6:00-7:00 a.m. (local time) bring the transforming air-mass from the urban and industrial areas to the coast in about one hour, thereby increasing the site's ozone concentration. During the afternoon these air-masses, now rich in photochemical oxidants are driven coastwards by the sea breeze in layers that are quasi-stable because of the underlying sea surface. Once they reach the coast, the presence of the TIBL can induce fumigation, producing high concentration values at the surface level. The overnight formation of a strong inversion induces a decoupling between the very light surface layer circulation and the upper layers, where pollutants such as ozone can be trapped without undergoing marked dilution processes. These pollutants can produce a pool at sea, where the concentration can remain high.
420 Pollution Control and Monitoring These findings suggested that the study be focused on the overall description of the processes involved in the transport, layering and deposition of chemicals at this complex site. 100-6 12 18 LOCAL TIME (h) Figure 1: Typical diurnal variation of Os and solar radiation for three monitoring points (data recorded between 18-31 July 1980). PCB (sea), AMGA (industrial area), COOP (coast), - solar radiation. 4.1 Transport phenomena An empirical trajectory model was used to simulate the horizontal transport processes of photochemical oxidants. It employs a space-time reconstruction of the field of average surface winds off the Ravenna coast during the summer. The assumed hypothesis, in tracing the path, is that the investigated air-mass has reached a state of vertical equilibrium in an ideal layer of isothermal atmosphere. The wind field was generated by linear interpolation of the wind speed data recorded along the monitoring path at the land and offshore stations. The time evolution was calculated using the diurnal variation of the wind vector along the path, which was then averaged over the period with breeze circulation [3]. A distance of 35 km offshore was set as the boundary value of the horizontal wind speed (v = 0). Tests performed with boundary value ranging between 30 and 40 km showed no significant differences in the resulting outputs. A position situated in the middle of Ravenna's industrial belt was selected as the coordinate origin and starting point of all the runs. The analysis of the sea trajectories for varying emission hours evinces that the air masses return to the shoreline or inland overnight and in the early morning hours but tend to accumulate at sea at later times (Fig. 2). An additional finding reveals an evident tendency to accumulate secondary pollutants out at sea and close to the coastline. Thus most of the recorded patterns evinced in the diurnal oxidant variations both in- and offshore can be explained by assuming a horizontal transport.
Pollution Control and Monitoring 421 Emission hour: 0 a 20 30 Figure 2: The sea trajectories of an air mass leaving the industrial area at the following times: (a) 2:00 a.m., (b) 8:00 a.m., (c) 10:00 p.m. 4.2 Deposition phenomena and the mass-budget approach While photochemical models have been employed to trace the evolution of oxidants, one of the main constraints is the great number of reactions that normally take place and the non-linearity of these processes. A mass-budget approach can simplify this study by parametrizing the component of the budget in final and initial pollutant concentration, concentration change due to chemical generation or destruction, and concentration change due to surface deposition. Although the approach is very rough, it can provide an estimation of the processes involved in the area. The dynamic material balance of a chemical can be expressed as AM/At = Si - Li, (1) where the first term is the time variation of the pollutant reservoir, Si the emission and Li both dry deposition and chemical reaction depletion. For an ozone-rich air mass transported by the sea-breeze, the time variation of the ozone mass can be represented as AM/At = -[k3 + ( Vd / H )]M, (2) where ks is the chemical rate constant of the NCh + Cb reaction, Vd the deposition velocity for a given surface, H the height of the reservoir layer and M the total mass of ozone. This equation can be used only under the assumption of very simple atmospheric-chemistry conditions; no sources of ozone must be intersected by the path of the air mass and the chemistry must be regulated only by the photo stationary cycle. While these conditions are not often found close to industrial or petrochemical plants, this method is useful to evaluate the contribution of a more complex chemistry as a residual of measured and calculated values once the deposition velocities and the reservoir layering height are defined. The deposition velocities for ozone were calculated using the inferential technique [5] over water and land by measurements performed at these sites
422 Pollution Control and Monitoring and inferred over the industrial and urban area with data reported in literature [6,7]. Table 1 reports the results of the mass-budget model as hourly differences in ozone concentrations for the varying surfaces affected by pollutant transport, along with the corresponding deposition velocities and average concentration values. Surface water land industrial area urban area AC (ppb/hr) 1.2 3.6 5.8 2.7 Vd(m/s) 0.001 0.01 0.025 0.02 c (ppb) 100 80 65 45 H(m) 300 800 1000 1200 Table 1: Results of the mass budget model. The calculated ratios between the results of the model and the values found by direct measurements are all greater than one (water ratio = 2.3, land ratio = 2.2, industrial area ratio = 1.4, urban area ratio = 1.5), implying that ozone destruction is not only attributable to deposition processes and to the photostationary state. 4.3 Ozone layering As noted above, the mass budget approach depends on the deposition velocities and the height of the reservoir layer assumed. The assumption of a stable, stratified, uniform ozone surface-layer that approaches the coastline during the re-entry of aged air-masses is only a useful approximation for a first rough description of these complex phenomena [8]. To assess the relationships between pollutant layering and breeze circulation, vertical soundings were performed to describe wind height evolution and airborne profiles of ozone concentration. The aircraft soundings reported in Table 2 indicate that the layering mechanisms operate throughout the area. Along the coast throughout day, the ozone concentration's vertical distribution and the vertical structure of wind speed and direction are closely correlated (Fig. 3 a, b, c). During the evening the highest concentration was found above 300 m. Given the increasing atmospheric stability both on sea and land, the ozone trapped at this height should not reach the surface but recirculate in the early morning under the land breeze, thereby contributing to increase the offshore pool of photochemicals. The applications of these findings to the mass-balance approach lead to differences of about 20% in the calculated vertical average coastal values. Given the mean contribution of all the layers as determined via their relative height, the resulting ratio between measured and calculated values is 1.7. The ratio value reduction from 2.2 to 1.7 suggests that the deposition processes play an increasing role when the ozone layering is taken into account. In reality the key point is a correct interpretation of the surface measurements in the presence
Pollution Control and Monitoring 423 of such a complex dynamic evolution of photochemical pollutants in the entire atmospheric boundary layer. Height (m) 400-500 300-400 200-300 100-200 0-100 400-500 300-400 200-300 100-200 0-100 400-500 300-400 200-300 100-200 0-100 Time Inshore 80-80 70 75 95 131 140 134 115 120 130 99 105 108 Coast 99 107 70 97 80 100 122 139 146 160 100 120 100 99 95 Offshore 145 125 _ 99 63 _ 120 115 139 175 127 98 108 Table 2: In situ soundings of ozone concentration for greater Ravenna wind speed 1 2 3 1400 1200'V ) 1000 Zc 800 / \ to o 600 400 200 > \ 0 " ^. i (m/s) /ind direction (degrees) 50 100 150 ozone concentration (p.p.b.) Figure 3: Vertical sounding of wind speed (solid line) and wind direction (dashed line) recorded at 08:30 (a) and at 14:30 (b) local time and vertical profiles of ozone concentration recorded along the coast from (dotdashed line), (solid line) and 18:00-19:00 (dashed line) local time. 5 Conclusions The distribution characteristics of oxidant pollution in Ravenna's coastal area are determined mainly by the presence of varying surfaces that generate
424 Pollution Control and Monitoring transport, deposition and chemical processes with markedly differentiated relative intensities. The experimental surveys here conducted and an empirical trajectory model developed to simulate the horizontal transport processes of the local air masses indicate that the diurnal pattern of ozone concentration is subject to the sea-land breeze circulation system. Offshore ozone concentration values always exceed the inshore ones and remain high also during the night because of greater atmospheric stability, limited vertical exchanges and diminished destruction processes at sea. The onshore ozone concentration trend shows a second peak in the late evening due to the re-entry of ozone-rich air masses from the sea. A rough mass-budget approach evinces the importance of a more complex chemistry than the photostationary cycle, especially at sea and in the coastal area. Measurements of ozone vertical profiles indicate a marked vertical layering, and calculations suggest that ozone layering must be taken into account when models of oxidant transport and deposition are applied in situations characterised by local circulations with a marked time-space evolution. Acknowledgements This study has been supported by the CNR-ENEL Project - Interaction of energy systems with human health and environment - Rome, Italy. References 1. Camuffo, D. & Bernardi, A. The diurnal trend in surface mixing ratio at Padova, Italy, Boundary-Layer Meteorol, 1982, 22, 273-282. 2. Kitada, T, Igarashi, K. & Owada, M. Numerical analysis of air pollution in a combined field of land/sea breeze and mountain/valley wind, J. dim. Appl Mcfeom/., 1986, 25, 767-784. 3. Fortezza, F, Strocchi, V., Giovanelli, G, Bonasoni P. & Georgiadis, T Transport of photochemical oxidants along the northwestern Adriatic coast, Atmospheric Environment, 1993, 27A, 2393-2402. 4. Lyons, W A & Cole, H.S. Photochemical oxidant transport: mesoscale lake breeze and synoptic scale aspect,./. Appl. Met., 1976, 15, 733-743. 5. Hicks, B.B., Baldocchi, D.D., Meyers, T.P., Hosker, R.P. & Matt, D.R. A preliminary multiple resistance routine for deriving dry deposition velocities from measured quantities, Water Air and Soil Pollut., 1987, 36, 311-330. 6. Wesely, ML Parametrization of surface resistances to gaseous dry deposition in regional-scale numerical models, Atmospheric Environment, 1989, 23, 1293-1304. 7. Wieringa, J. Representative roughness parameters for homogeneous terrain, Boundary-Layer MeteoroL, 1993, 63, 323-363. 8. Georgiadis, T, Giovanelli, G & Fortezza F. Vertical layering of photochemical ozone during land-sea breeze transport, // Nuovo Cimento, 1994, 17C, May-June in press.