Ozone forecasts

Pagowski M, Grell GA (2006) Ensemble-based ozone forecasts skill and economic value. 

The statistical comparisons between ozone forecasts and measurements for all stations in eastern Austria show correlation coefficients between 0.4 and 0.7 and standard errors around 12 ppbv for 2006. 

Daily ozone forecasts for Austria have been run during summer 2005 and 2006. The results of the forecasts have been evaluated with measurements of the Austrian air quality network for eastern parts of Austria. Generally the observed exceedances of the information threshold are re-produced well by the model. Days with exceedances of the information threshold were predicted by the model with a probability of 88% during the summer period in 2006. The daily maxima above the alert... 

Because one of the major impacts of increased UV at the earth s surface is expected to be an increase in skin cancer, several countries now include UV forecasts as part of their weather reports (e.g., see Kerr, 1994), and estimates of skin cancer increases due to ozone depletion have been made. Figure 13.20, for example, shows the estimated number of excess cases of skin cancers for the United States and for northwestern Europe if no controls had been imposed on CFCs and halons and those expected with the Copenhagen Amendments (Slaper et al., 1996). Clearly, there is expected to be a major impact of the control strategies on the incidence of skin cancer, although the number of excess cases will still be about 33,000 per year in the United States and 14,000 per year in northwestern Europe at the projected worst-case year of 2050. 

Solomon, S and D. L. Albritton, Time-Dependent Ozone Depletion Potentials for Short- and Long-Term Forecasts, Nature, 357, 33-37 (1992). 

Ziemke J.R., Herman J.R., Stanford J.L. and Bhartia P.K., Total ozone/UVB monitoring and forecasting Impact of clouds and the horizontal resolution of satellite retrievals, J. Geophys Res., 103, 3865-3871, 1998. 

The altitude effect (Sec. 3) and the radiation amplification factor (Sec. 4) were derived from UV-ERY measurements made simultaneously at two locations in the Czech Republic. The value of RAF obtained from the present data agrees with previous studies of other authors. The value of the amplitude effect agrees with the value used by National Weather Service and EPA [10] but is lower than the values obtained by other authors [2, 9]. The statistical model relating UV-ERY irradiance with total ozone and solar zenith angle was developed (Sec. 5 Fig. 2). Although the information on the total ozone does not satisfactorily improves accuracy of the UV-ERY forecast (further variables should be incorporated into the model to improve its accuracy), the model may be used to estimate annual and daily cycles of sun-visible UV-ERY irradiance for various total ozone levels. The results obtained show variability of the model UV-ERY irradiance related to variability of total column ozone. Specifically, it is demonstrated that the UV-ERY irradiance may exceed the annual/daily normal-ozone maxima during non-negligible portion of the year/day (about 214 months/hours) if the total ozone... 

A more detailed overview of the main components of the GEOS-DAS system the forecast model, the input data (total ozone observations from Total Ozone Mapping Spectrometer /TOMS/ and vertical ozone profiles from the Solar Backscatter Ultra Violet instrument /SBUV/, the analysis scheme and its implementation could be easy found in the paper of Riishojgard [19]. 

Fig. 3.1 Correlation coefficients for various model runs, comparing model forecasts of 8-h averaged peak ozone mixing ratios with those observed by surface monitoring stations. The statistics include a month worth of forecast runs...

The AQ forecasting system results have been compared with observations over a 8 month period from June 2006 to January 2007. The predicted concentration data have been divided in two time series obtained selecting the first (last) 24 h of each daily forecast cycle that covers the 48 h period. Comparison of these two time series with observations (Eig. 9.7) showed that the 48 h forecast generally obtains higher concentrations and better fits with observations in comparison to the 24 h forecast. This behaviour was common to all pollutants except ozone (which was overestimated for the 24 h simulation). The comparison of initial and 24 h concentration fields showed that differences were due to the influence of initial conditions on the first simulated day. The resolution difference between CHIMERE (50 km) and FARM (4 km) background domains did not allow obtaining a proper initialization. In CHIMERE topography, the city of Torino is located on the slope of the western Alps, at about 800 m asl. This feature clearly favours ozone overestimation and... 

These modelling systems are developed in order to fulfil several duties at DMI, e.g. (1) smog and ozone preparedness, (2) nuclear, veterinary, and chemical emergency preparedness and risk assessment, (3) birch pollen forecasting, and (4) research and development activities. 

Examples of the surface ozone and nitrogen dioxide forecasts vs. observations are shown in Fig. 16.8. The produced plots are forecasts (top panel) of O3 and NO2 at 00 UTC on 19 Mar 2010 together with near-real-time observations (bottom panel) of these species. The simulations resulted in a relatively medium air pollution levels over the Central and Eastern Europe with a few hot-spots. For O3 it is an episode observed in the east of the Northern Italy, Austria and Slovenia and... 

Friday 19 March 2010 OOUTC GEMS-RAO Forecast t+000 VT Friday 19 March Friday 19March 2010 OOOTCGEMS-RAQ Forecast t 000 VT Friday 19 March 2010 OOUTC Medal CAC Height level Surface Parameter Ozone [ ag/m ] 2010 OOUTC Model CAC Height level Surface Parameter Nitrogen Dioxide ( gfm ]... 

Fig. 16.8 GEMS Regional Air Quality Forecast by the DMI modellmg system for ozone and nitrogen dioxide top panel) forecasts vs. bottom panel) observations on 19 Mar 2010, 00 UTC...

The boundary conditions of the coarse grid were estimated from the forecast of the previous day. This method is compared with constant boundary conditions using average summer values. Total ozone column data was obtained from ECMWE data. 

Figure 18.2 shows the predicted maximum concentrations for ozone region 1. The course of ozone concentration from 1 day (Tag 1) and 2-days (Tag 2) model forecasts as well as the results from a backup run which considers constant boundary conditions only are compared to measurements (at 43 air quality stations). 

According to the station measurements, the information threshold has been exceeded on 17 days in the period depicted in Fig. 18.2a. Note that 16 of these exceedances were forecasted correctly. The alert exceedance predicted for July 21 was not observed but the measured values were just below the threshold value. The daily maxima above the alert threshold which were observed on 27-28 July 2006, were not predicted by the model. On these days, local peak emissions obviously caused a sudden and local increase of ozone concentrations around noon for 1-2 h. 

Figure 18.2b shows a scatter diagram of observed vs. modelled ozone in region 1. Depicted are daily maxima for the two forecast days and the backup run. Information (90 ppbv) and alarm threshold (120 ppbv) are marked by magenta lines (horizontal and vertical). Most of the values lie above the 1 1 line which means that the model tends to predict higher concentrations than observed. Table 18.1 summarizes the metrics of some selected stations in ozone region 1. The correlation and the standard deviation between observations and model prediction of the two forecast days as well as the backup run are depicted. The last line shows the values for all 43 stations in ozone region 1. 

The highest value of the observations as well as the model forecasts are bellow the information threshold (90 ppbv) in ozone region 1... 

---