Fuzzy system models combined with nonlinear regression for daily ground-level ozone predictions

This study focuses on applying a Takagi–Sugeno fuzzy system and a nonlinear regression (NLR) model for ozone predictions in six Kentucky metropolitan areas. The fuzzy “c-means” clustering technique coupled with an optimal output predefuzzification approach (least square method) was used to train the Takagi–Sugeno fuzzy system. The fuzzy system was tuned by specifying the number of rules and the fuzziness factor. The NLR models were based in part on a previously reported, trajectory-based hybrid NLR model that has been used for years for forecasting ground-level ozone in Louisville, KY. The NLR models were each composed of an interactive nonlinear term and several linear terms. Using a common meteorological parameter set as input variables, the NLR models and the Takagi–Sugeno fuzzy systems model exhibited equivalent forecasting performance on test data from 2004. For all 2004 ozone season forecasts for the six metropolitan areas, the mean absolute error was 8.1 ppb for the NLR model and 8.0 ppb for the Takagi–Sugeno fuzzy model. When a nonlinear term (which was part of the NLR model) was included in the fuzzy model, the combined NLR–fuzzy model had slightly better performance than the original NLR model. For all 2004 metropolitan area forecasts, the mean absolute error of the NLR–fuzzy model forecasts was 7.7 ppb. These small differences may be statistically significant, but for practical purposes the performance of the fuzzy models was equivalent to that of the NLR models.     

Validation of a hybrid forecasting system for the ozone concentrations over the Paris area

A simplified hybrid statistical-deterministic chemistry-transport model, is used in real time for the prediction of ozone in the area of Paris during Summer 1999. We present here a statistical validation of this experiment. We distinguish the forecasts in the urban area from forecasts in the pollution plume downwind of the city. The validation of model forecasts, up to 3 days ahead, is performed against ground based observations within and up to 50 km outside of Paris. In the urban area, ozone levels are fairly well forecast, with correlation coefficients between forecast and observations ranging between 0.7 and 0.8 and root mean square errors in the range 15–20 μg m⁻³ at short lead times. While the bias of urban forecast is very low, the largest peaks are somehow underestimated. The ozone plume amplitude is generally well reproduced, even at long lead times (root mean square errors of about 20–30 μg m⁻³), while the direction of the plume is only captured at short lead times (about 70% of the time). The model has difficulties in forecasting the direction of the plume under stagnant weather conditions. We estimate the model ability to forecast concentrations above 180 μg m⁻³, which are of practical relevance to air quality managers. It is found that about 60% of these events are well forecast, even at long lead times, while the exact monitoring station where the exceedance is observed can only be forecast at short lead times. Finally, we found that about half of the forecast error is due to the error in the estimation of the boundary conditions, which are forecast by a simple linear regression model here.     

A model for predicting maximum and 8 h average ozone in Houston

The objective of this research was to develop a statistical model to predict one day in advance both the maximum and 8 h (10 am–5 pm) average ozone for Houston (TX). A loess/generalized additive model (GAM) approach was taken to model development. Ozone data (1983–1991) from ten stations in the immediate Houston area were used in the study. The meteorological data came from the Houston International Airport. The models were developed using data for April through October for 1983–1987 and 1989–1990. Forecasts were developed for 1988 and 1991. The final model, which was multiplicative in nature, contained three interaction terms for the west/east and south/north wind components (average of hourly values from 8 pm to 5 am, 6 am to 9 am, and 10 am to 5 pm). Opaque cloud cover (averaged over the period 10  am to 5 pm), yesterday’s maximum ozone, today’s maximum temperature and morning mixing depth were also important variables in the model.Individual forecasts were generated for all ten stations in the Houston area using observed meteorology. In addition forecasts were produced for three measures of the network as a whole. The root-mean-square prediction error for the 8 h average forecasts ranged from 13.2 to 16.3 ppb (with R² ranging from 0.66 to 0.73) for the individual stations and from 18.5 to 22.0 ppb (with R² ranging from 0.61 to 0.68) for maximum ozone. A detailed examination was undertaken for a day on which the forecast was much too low.     

The variability of free tropospheric ozone over Beltsville,Maryland (39N, 77W) in the summers 2004–2007

Ozone profiles are often used to investigate day-to-day and year-to-year variability in origins of free tropospheric ozone. With this in mind, more than 50 ozonesonde launches were conducted in Beltsville, MD, during the summers of 2004 through 2007. Budgets of free tropospheric ozone were calculated for each ozone profile in the four summers using a laminar identification (LID) method and unusual episodes were analyzed with respect to meteorological variables. The laminar method showed that stratosphere-to-troposphere transport (ST) accounted for greater than 50% of the free tropospheric ozone column on 17% of days sampled, a more pronounced influence than regional convective and lightning (RCL) sources. The ST origins were confirmed with trajectories, and tracers (water vapor and potential vorticity). The amount of free tropospheric ozone from ST and RCL sources varied from year-to-year (up to 13%) and can be explained by differences in mean meteorological patterns. On average, almost 30% of the free tropospheric column was attributed to ST influence, about twice as much as RCL, although the LID method may not capture weeks-old lightning influences as in a chemical model. The prevalence of ST ozone in summertime Beltsville soundings was similar to six sounding sites in the IONS-04 campaign [Thompson, A.M., et al., 2007b. Intercontinental Transport Experiment Ozonesonde Network Study (IONS, 2004): 1. Summertime upper tropospheric/lower stratosphere ozone over northeastern North America. J. Geophys. Res. 112, D12S12; Thompson, A.M., et al., 2007c. Intercontinental Transport Experiment Ozonesonde Network Study (IONS, 2004): 2. Tropospheric ozone budgets and variability over northeastern North America. J. Geophys. Res. 112, D12S13.] and to statistics from a 30 year climatology of European soundings [Collette, A., Ancellet, G., 2005. Impact of vertical transport processes on the tropospheric ozone layering above Europe. Part II: Climatological analysis of the past 30 years. Atmos. Environ. 39, 5423–5435]. The Beltsville record also demonstrated the value of soundings for air quality forecasting in an urban area. The 22 nighttime soundings collected over Beltsville in 2004–2007 can be divided into distinct polluted and unpolluted subsets, the former 20 ppbv higher in residual layer ozone (1 km) than the latter. These distinctions propagated to daytime differences of 10 ppbv at the surface in the Washington, DC, area, with the high-ozone residual layers leading to non-attainment of the National Ambient Air Quality Standard for ozone. More frequent ozone observations aloft appear essential for better understanding ozone variability and for enabling air quality modelers to achieve more accurate ozone forecasts.     

An enhanced PM2.5 air quality forecast model based on nonlinear regression and back-trajectory concentrations

An enhanced PM_2.5 air quality forecast model based on nonlinear regression (NLR) and back-trajectory concentrations has been developed for use in the Louisville, Kentucky metropolitan area. The PM_2.5 air quality forecast model is designed for use in the warm season, from May through September, when PM_2.5 air quality is more likely to be critical for human health. The enhanced PM_2.5 model consists of a basic NLR model, developed for use with an automated air quality forecast system, and an additional parameter based on upwind PM_2.5 concentration, called PM24. The PM24 parameter is designed to be determined manually, by synthesizing backward air trajectory and regional air quality information to compute 24-h back-trajectory concentrations. The PM24 parameter may be used by air quality forecasters to adjust the forecast provided by the automated forecast system. In this study of the 2007 and 2008 forecast seasons, the enhanced model performed well using forecasted meteorological data and PM24 as input. The enhanced PM_2.5 model was compared with three alternative models, including the basic NLR model, the basic NLR model with a persistence parameter added, and the NLR model with persistence and PM24. The two models that included PM24 were of comparable accuracy. The two models incorporating back-trajectory concentrations had lower mean absolute errors and higher rates of detecting unhealthy PM2.5 concentrations compared to the other models.     

Impact of large scale circulation on European summer surface ozone and consequences for modelling forecast

In this study, we investigate the benefit for European ozone simulation of using day-to-day varying chemical boundary conditions produced by a global chemical weather forecast platform instead of climatological monthly means at the frontiers of a regional model. We performed two simulations over Europe using the regional (0.5 × 0.5°) CHIMERE CTM forced by global scale simulations based on the LMDz-INCA CTM. For summer 2005, ozone differences exceeding 20 ppb can be punctually found between these two simulations in the borders of the domain. The mean of the differences ranges between 0 and 3 ppb beyond 15° of the frontiers of the regional model.Correlations with ground-based ozone measurements at more than 400 stations are slightly increased by the use of daily boundary conditions. The simulation of the temporal variability is significantly enhanced in particular for the daily means and daily maxima. As expected, the gain is higher at the borders of the regional domain.The change of percentile distribution shows that the net impact of high temporal resolution boundary conditions is not of major concern for surface ozone peaks which are mainly due to local photochemistry. The use of daily boundary conditions is however necessary to correctly simulate concentrations in the 20–35 ppb range which are of crucial interest for human and vegetation exposure effects.     

Reactions between ozone and building products: Impact on primary and secondary emissions

Reactions of ozone on common building products were studied in a dedicated emission test chamber system. Fourteen new and unused products were exposed to 100–160 ppb of ozone at 23 °C and 50% RH during 48 h experiments. Ozone deposition velocities calculated at steady state were between 0.003 cm s⁻¹ (alkyd paint on polyester film) and 0.108 cm s⁻¹ (pine wood board). All tested product showed modified emissions when exposed to ozone and secondary emissions of several aldehydes were identified. Carpets and wall coverings emitted mainly C₅–C₁₀ n-aldehydes, typical by-products of surface reactions. Linoleum, polystyrene tiles and pine wood boards also showed increased emissions of formaldehyde, benzaldehyde and hexanal associated with reduced emissions of unsaturated compounds suggesting the occurrence of gas-phase reactions. The ozone removal on the different tested products was primarily associated with surface reactions. The relative contribution of gas-phase reactions to the total ozone removal was estimated to be between 5% and 30% for pine wood boards depending on relative humidity (RH) and on the incoming ozone concentration and 2% for polystyrene tiles. On pine wood board, decreasing ozone deposition velocities were measured with increasing ozone concentrations and with RH increasing in the range 30–50%.     

Ground-level ozone and ozone vertical profile measurements close to the foothills of the Guadarrama mountain range (Spain)

Continuous measurements of ozone vertical profiles, OVP, in the low troposphere (around 500–2400 m) using an unattended commercial ozone profiler DIAL, were conducted during June–July 2004 in Segovia, SG, a small city in the upper plateau located close to the foothills of the Guadarrama mountain range, Guadarrama, in the Central Massif. The data obtained over almost 37 complete days have enabled us to characterise the ozone vertical exchange, describe the phenomenology of the main ozone peaks, OP, recorded in the city and their relationship with ozone transport/formation from the gas precursor emissions of the greater Madrid area across Guadarrama. To achieve the last objective concurrent measurements of ground-level ozone in SG and a representative monitoring station upwind from Guadarrama, Buitrago de Lozoya, BL, have been used. 72.2% of the concurrent maximum diurnal ozone peaks exceeding the 95 percentile hourly value in SG (OPSG) and BL (OPBL) were linked to ozone transport and formation from the greater Madrid area towards Guadarrama. An estimate of the contribution of the greater Madrid area on OPSG yielded 28 μg m⁻³.The most prominent ozone vertical stratification was linked to the mixing height, MH, and a frequent nocturnal stable layer formed, NSL. Three small ozone enriched-layers were identified at mean heights of 500, 700 and 1000 m, respectively. Ozone tended to decline versus altitude. The hourly patterns of the three layers showed two peak occurrences of similar amplitude in the early morning, 7–8 h, and mid-afternoon, 14–16 h. A minimum was also observed during daytime, 10–11 h, its origin being attributed to a dilution process induced by the “chimney effect” caused by the slopes heating during this period.The comparison between OPSG, and the maximum diurnal ozone peaks in the first layer, OL1P, showed a satisfactory relationship, correlation coefficient, r, of the linear fit 0.77, and comparable mean values, 127 and 130 μg m⁻³, respectively, revealing the presence of an uniform ozone vertical distribution in the 500 m atmospheric layer above ground level during mid-afternoon.     

Operational ozone forecasts for the region of Copenhagen by the Danish Meteorological Institute

The Danish Meteorological Institute (DMI) has developed an operational forecasting system for ozone concentrations in the Atmospheric Boundary Layer; this system is called the Danish Atmospheric Chemistry FOrecasting System (DACFOS). At specific sites where real-time ozone concentration measurements are available, a statistical after-treatment of DACFOS’ results adjusts the next 48 h ozone forecasts. This post-processing of DACFOS’ forecasts is based on an adaptive linear regression model using an optimal state estimator algorithm. The regression analysis uses different linear combinations of meteorological parameters (such as temperature, wind speed, surface heat flux and atmospheric boundary layer height) supplied by the Numerical Weather Prediction model DMI-HIRLAM. Several regressions have been tested for six monitoring stations in Denmark and in England, and four of the linear combinations have been selected to be employed in an automatic forecasting system. A statistical study comparing observations and forecasts shows that this system yields higher correlation coefficients as well as smaller biases and RMSE values than DACFOS; the present post-processing thus improves DACFOS’ forecasts. This system has been operational since June 1998 at the DMI's monitoring station in the north of Copenhagen, for which a new ozone forecast is presented every 6 h on the DMI's internet public homepage.     

Ozone production from Grenoble city during the August 2003 heat wave

The heat wave from 1 to 16 August 2003 is considered in the city of Grenoble (French Alps). The modelling system (PREVALP) is based on several models operating on nested domains: MM5 for dynamics, CHIMERE for chemistry (18 km and 6 km grid size) and METPHOMOD for both dynamics and chemistry (2 km grid size). The analysis of the results shows that during the heat wave, the mixing layer is thicker, up to 3500 m agl, hence inducing transport of ozone to high altitude. Two regimes were diagnosed: (1) a freely developing convective layer, (2) a layer trapped under south wind which makes ozone precursors accumulate in the city. Local ozone production is estimated to be 40% of the plume maximum in case (2) and only 30% in case (1). Sensitivity analysis by step increase for temperature at the boundary of the inner domain shows the non-linearity of the response; in this case most of the effect comes from chemistry. By changing biogenic emission significant changes are observed in restricted areas.     

A review of surface ozone in the polar regions

Surface ozone records from ten polar research stations were investigated for the dependencies of ozone on radiative processes, snow-photochemisty, and synoptic and stratospheric transport. A total of 146 annual data records for the Arctic sites Barrow, Alaska; Summit, Greenland; Alert, Canada; Zeppelinfjellet, Norway; and the Antarctic stations Halley, McMurdo, Neumayer, Sanae, Syowa, and South Pole were analyzed. Mean ozone at the Northern Hemisphere (NH) stations (excluding Summit) is ∼5 ppbv higher than in Antarctica. Statistical analysis yielded best estimates for the projected year 2005 median annual ozone mixing ratios, which for the Arctic stations were 33.5 ppbv at Alert, 28.6 ppbv at Barrow, 46.3 ppbv ppb at Summit and 33.7 ppbv at Zeppelinfjellet. For the Antarctic stations the corresponding ozone mixing ratios were 21.6 ppbv at Halley, 27.0 ppbv at McMurdo, 24.9 ppbv at Neumayer, 27.2 ppbv at Sanae, 29.4 ppbv at South Pole, and 25.8 ppbv at Syowa. At both Summit (3212 m asl) and South Pole (2830 m asl), annual mean ozone is higher than at the lower elevation and coastal stations. A trend analysis revealed that all sites in recent years have experienced low to moderate increases in surface ozone ranging from 0.02 to 0.26 ppbv yr⁻¹, albeit none of these changes were found to be statistically significant trends. A seasonal trend analysis showed above-average increases in ozone during the spring and early summer periods for both Arctic (Alert, Zeppelinfjellet) and Antarctic (McMurdo, Neumayer, South Pole) sites. In contrast, at Barrow, springtime ozone has been declining. All coastal stations experience springtime episodes with rapid depletion of ozone in the boundary layer, attributable to photochemically catalyzed ozone depletion from halogen chemistry. This effect is most obvious at Barrow, followed by Alert. Springtime depletion episodes are less pronounced at Antarctic stations. At South Pole, during the Antarctic spring and summer, photochemical ozone production yields frequent episodes with enhanced surface ozone. Other Antarctic stations show similar, though less frequent spring and summertime periods with enhanced ozone. The Antarctic data provide evidence that austral spring and summertime ozone production in Antarctica is widespread, respectively, affects all stations at least through transport events. This ozone production contributes to a several ppbv enhancement in the annual mean ozone over the Antarctic plateau; however, it is not the determining process in the Antarctic seasonal ozone cycle. Although Summit and South Pole have many similarities in their environmental conditions, this ozone production does not appear to be of equal importance at Summit. Amplitudes of diurnal, summertime ozone cycles at these polar sites are weaker than at lower latitude locations. Amplitudes of seasonal ozone changes are larger in the Southern Hemisphere (by ∼5 ppbv), most likely due to less summertime photochemical ozone loss and more transport of ozone-rich air to the Arctic during the NH spring and summer months.     

Ozone response to precursor controls: comparison of data analysis methods with the predictions of photochemical air quality simulation models

Comparisons were made between the predictions of six photochemical air quality simulation models (PAQSMs) and three indicators of ozone response to emission reductions: the ratios of O₃/NO_z and O₃/NO_y and the extent of reaction. The values of the two indicator ratios and the extent of reaction were computed from the model-predicted mixing ratios of ozone and oxidized nitrogen species and were compared to the changes in peak 1 and 8 h ozone mixing ratios predicted by the PAQSMs. The ozone changes were determined from the ozone levels predicted for base-case emission levels and for reduced emissions of volatile organic compounds (VOCs) and oxides of nitrogen (NO_x). For all simulations, the model-predicted responses of peak 1 and 8 h ozone mixing ratios to VOC or NO_x emission reductions were correlated with the base-case extent of reaction and ratios of O₃/NO_z and O₃/NO_y. Peak ozone values increased following NO_x control in 95% (median over all simulations) of the high-ozone (>80 ppbv hourly mixing ratio in the base-case) grid cells having mean afternoon O₃/NO_z ratios less than 5 : 1, O₃/NO_y less than 4 : 1, or extent less than 0.6. Peak ozone levels decreased in response to NO_x reductions in 95% (median over all simulations) of the grid cells having peak hourly ozone mixing ratios greater than 80 ppbv and where mean afternoon O₃/NO_z exceeded 10 : 1, O₃/NO_y was greater than 8 : 1, or extent exceeded 0.8. Ozone responses varied in grid cells where O₃/NO_z was between 5 : 1 and 10 : 1, O₃/NO_y was between 4 : 1 and 8 : 1, or extent was between 0.6 and 0.8. The responses in such grid cells were affected by ozone responses in upwind grid cells and by the changes in ozone levels along the upwind boundaries of the modeling domains.     

A performance evaluation of the National Air Quality Forecast Capability for the summer of 2007

This paper provides a performance evaluation of the real-time, CONUS-scale National Air Quality Forecast Capability (NAQFC) that supported, in part, its transition into operational status. This evaluation focuses primarily on discrete forecasts for the maximum 8-h O₃ concentrations covering the 4-month period, June through September, 2007, using measurements obtained from EPA's AIRNow network. Results indicate that the 2007 NAQFC performed as well or better than previous configurations, despite the expansion of the forecast domain into the western half of the nation that is dominated by complex terrain. The mean, domain-wide, season-long correlation was 0.70. When examined over time, the domain-wide correlations exhibit a fairly consistent nature, with values exceeding 0.60 (0.70) over 90% (55%) of the days. The NAQFC systematically over-predicted the 8-h O₃ concentrations, continuing a trend established by earlier NAQFC configurations, though to a lesser degree. The summer-long mean forecast value of 53.2 ppb was 4.2 ppb higher than the observed value, resulting in a domain-wide Normalized Mean Bias (NMB) of 8.7%. Most of the over-prediction is associated with observed concentrations less than 50 ppb. In fact the model tends to under-predict when concentrations exceed 70 ppb. As with the bias, the error associated with the latest configuration was also lower. The summer-long Root Mean Square Error of 13.0 ppb (Normalized Mean Error (NME) = 20.4%) represented marked improvements over earlier forecasts. Examination of the spatial distribution of both the NMB and NME reveals that the NAQFC was generally within 25% for the NME and 25% for the NMB over a majority of the domain. Several areas of poorer performance, where the NMB and NME often exceed 25% and in some cases 50%, were noted. These areas include southern California, where the NAQFC tended to under-predict concentrations (especially on weekends) and the southeast Atlantic and Gulf coasts regions, where the model over-predicted. Subsequent analysis revealed that the incorrect temporal allocation of precursor emissions was likely the source of the under-prediction in southern California, while inaccurate simulation of PBL heights likely contributed to the over-prediction in the coastal regions.     

Measurements of fine and ultrafine particles formation in photocopy centers in Taiwan

This study investigates the levels of particulate matter smaller than 2.5 μm (PM_2.5) and some selected volatile organic compounds (VOCs) at 12 photocopy centers in Taiwan from November 2004 to June 2005. The results of BTEXS (benzene, toluene, ethylbenzene, xylenes and styrene) measurements indicated that toluene had the highest concentration in all photocopy centers, while the concentration of the other four compounds varied among the 12 photocopy centers. The average background-corrected eight-hour PM_2.5 in the 12 photocopy centers ranged from 10 to 83 μg m⁻³ with an average of 40 μg m⁻³. The 24-h indoor PM_2.5 at the photocopy centers was estimated and at two photocopy centers exceeded 100 μg m⁻³, the 24-h indoor PM_2.5 guideline recommended by the Taiwan EPA. The ozone level and particle size distribution at another photocopy center were monitored and indicated that the ozone level increased when the photocopying started and the average ozone level at some photocopy centers during business hour may exceed the value (50 ppb) recommended by the Taiwan EPA. The particle size distribution monitored during photocopying indicated that the emitted particles were much smaller than the original toner powders. Additionally, the number concentration of particles that were smaller than 0.5 μm was found to increase during the first hour of photocopying and it increased as the particle size decreased. The ultrafine particle (UFP, <100 nm) dominated the number concentration and the peak concentration appeared at sizes of under 50 nm. A high number concentration of UFP was found with a peak value of 1E+8 particles cm⁻³ during photocopying. The decline of UFP concentration was observed after the first hour and the decline is likely attributable to the surface deposition of charged particles, which are charged primarily by the diffusion charging of corona devices in the photocopier. This study concludes that ozone and UFP concentrations in photocopy centers should be concerned in view of indoor air quality and human health. The corona devices in photocopiers and photocopier-emitted VOCs have the potential to initiate indoor air chemistry during photocopying and result in the formation of UFP.     

Quantification of Saharan dust contribution to PM10 concentrations over Italy during 2003–2005

Italy is frequently affected by Saharan dust intrusions, which result in high PM₁₀ concentrations in the atmosphere and can cause the exceedances of the PM₁₀ daily limits (50 μg m^?3) set by the European Union (EU/2008/50). The estimate of African dust contribution to PM₁₀ concentrations is therefore a key issue in air quality assessment and policy formulation. This study presents a first identification of Saharan dust outbreaks as well as an estimate of the African dust contribution to PM₁₀ concentrations during the period 2003–2005 over Italy. The identification of dust events has been carried out by looking at different sources of information such as monitoring network observations, satellite images, ground measurements of aerosol optical properties, dust model simulations and air mass backward trajectory analysis. The contribution of Saharan dust to PM₁₀ monthly concentrations has been estimated at seven Italian locations. The results are both spatially (with station) and temporally (with month and year) variable, as a consequence of the variability of the meteorological conditions. However, excluding the contribution of severe dust events (21st February 2004, 25th–28th September 2003, 23rd–27th March 2005), the monthly contribution of dust varies approximately between 1 μg m^?3 and 10 μg m^?3 throughout year 2005 and between 1 μg m^?3 and 8 μg m^?3 throughout year 2003. In 2004 the dust concentration is lower than 2003 and 2005 (<5 μg m^?3 at all sites). The reduction in the number of daily exceedances of the limit value (50 μg m^?3) after subtraction of the dust contribution is also calculated at each station: it varies with station between 20% and 50% in 2005 and between 5% and 25% in 2003 and 2004.     

Surface ozone background in the United States: Canadian and Mexican pollution influences

We use a global chemical transport model (GEOS-Chem) with 1° × 1° horizontal resolution to quantify the effects of anthropogenic emissions from Canada, Mexico, and outside North America on daily maximum 8-hour average ozone concentrations in US surface air. Simulations for summer 2001 indicate mean North American and US background concentrations of 26 ± 8 ppb and 30 ± 8 ppb, as obtained by eliminating anthropogenic emissions in North America vs. in the US only. The US background never exceeds 60 ppb in the model. The Canadian and Mexican pollution enhancement averages 3 ± 4 ppb in the US in summer but can be occasionally much higher in downwind regions of the northeast and southwest, peaking at 33 ppb in upstate New York (on a day with 75 ppb total ozone) and 18 ppb in southern California (on a day with 68 ppb total ozone). The model is successful in reproducing the observed variability of ozone in these regions, including the occurrence and magnitude of high-ozone episodes influenced by transboundary pollution. We find that exceedances of the 75 ppb US air quality standard in eastern Michigan, western New York, New Jersey, and southern California are often associated with Canadian and Mexican pollution enhancements in excess of 10 ppb. Sensitivity simulations with 2020 emission projections suggest that Canadian pollution influence in the Northeast US will become comparable in magnitude to that from domestic power plants.     

Responses of two birch (Betula pendula Roth) clones to different ozone profiles with similar AOT40 exposure

Saplings of two clones of European white birch (Betula pendula Roth) were exposed to three different ozone profiles resulting in same AOT40 value of 13–14 ppm h in a chamber experiment. The sensitive clone 5 and the more tolerant clone 2 were growing (1) under filtered air (=control), or (2) were exposed to 70 ppb ozone for 24 h d⁻¹ (=profile 1), (3) to 100 ppb ozone for 12 h d⁻¹ at 8:00–20:00 (=profile 2), or (4) to 200 ppb ozone for 4.5 h d⁻¹ at 9:30–14:00 (=profile 3) for 20 d. The saplings were determined for growth, visible leaf injuries, stomatal conductance, and concentrations of Rubisco, chlorophyll and carotenoids. Growth responses and induction of visible foliar injuries under different ozone profiles were variable, resulting in 4–17% lower dry mass of shoot, 16–46% reduction in stem height increment and 11–43% increase in visible injuries in clone 5, which was accompanied by higher leaf turnover rate under profile 3 indicating compensation growth. In clone 2, ozone-induced responses ranged from slight stimulation in stem height growth to 13% decrease in dry mass of shoot and 2–16% increase in visible injuries. Daytime stomatal conductance rates were lowered by 14–54% in clone 5 and 9–74% in clone 2, depending on profile. The additional power-weighted analyses revealed that high peak concentrations and exposure shape were important for induction of visible injuries in both clones and reduction in stomatal conductance in clone 5, whereas growth reductions were rather related to total cumulative exposure. The results indicate that profile of ozone exposure, night-time stomatal conductance (24 h flux), and recovery time for defence and compensations reactions should not be ignored in plant response and ozone flux modelling.     

A Comparison of Nonlinear Regression and Neural Network Models for Ground-Level Ozone Forecasting

ABSTRACT

A hybrid nonlinear regression (NLR) model and a neural network (NN) model, each designed to forecast next-day maximum 1-hr average ground-level O₃ concentrations in Louisville, KY, were compared for two O₃ seasons—1998 and 1999. The model predictions were compared for the forecast mode, using forecasted meteorological data as input, and for the hindcast mode, using observed meteorological data as input. The two models performed nearly the same in the forecast mode. For the two seasons combined, the mean absolute forecast error was 12.5 ppb for the NLR model and 12.3 ppb for the NN model. The detection rate of 120 ppb threshold exceedances was 42% for each model in the forecast mode. In the hindcast mode, the NLR model performed marginally better than the NN     

Evidence of impact of aerosols on surface ozone concentration in Tianjin,China

In this study, we will present evidence that aerosol particles have strong effects on the surface ozone concentration in a highly polluted city in China. The measured aerosol (PM10), UV flux, and O₃ concentrations were analyzed from 1 November (1 Nov) to 7 November (7 Nov) 2005 in Tianjin, China. During this period, the aerosol concentration had a strong day-by-day variation, ranging from 0.2 to 0.6 mg m⁻³. The ozone concentration also shows a strong variability in correlation with the aerosol concentration. During 1 Nov, 2 Nov, 6 Nov, and 7 Nov, the ozone concentration was relatively high (about 30–35 ppbv; defined as a high-ozone period), and during 3 Nov to 5 Nov, the ozone concentration was relatively low (about 5–20 ppbv; defined as a low-ozone period). The analysis of the measurement shows that the ozone concentration is strongly correlated to the measured UV flux. Because there were near cloud-free conditions between 1 Nov and 7 Nov, the variation of the UV flux mainly resulted from the variation of aerosol concentration. The result shows that higher aerosol concentrations produce a lower UV flux and lower ozone concentrations. By contrast, the lower aerosol concentration leads to a higher UV flux and higher ozone concentrations. A chemical mechanism model (NCAR MM) is applied to interpret the measurement. The model result shows that the extremely high aerosol concentration in this polluted city has a very strong impact on photochemical activities and ozone formation. The correlation between aerosol and ozone concentrations appears in a non-linear feature. The O₃ concentration is very sensitive to aerosol loading when aerosol loading is high, and this sensitivity is reduced when aerosol loading is low. For example, the ratio of Δ[O₃]/Δ[AOD] is about −16 ppbv AOD⁻¹ when AOD is less than 2, and is only −4 ppbv AOD⁻¹ when AOD is between 2 and 5. This result implies that a future decrease in aerosol loading could lead to a rapid increase in the O₃ concentration in this region.