Een methode voor het afleiden van lokale luchtkwaliteit gebaseerd op grootschalige circulatie patronen en oppervlakte meteorologie.

摘要

Clean air is essential to our own health and that of the environment. Si nce the industrial revolution, the quality of the air we breathe has det eriorated considerably for some chemical species (e.g. ozone (O3) and pa rticulate matter (PM)), which can lead to serious health problems. One t he one hand, this is caused by a change in human activities, characteriz ed by e.g. rising industrial and energy production, the burning of fossi l fuels and especially a dramatic rise in traffic. On the other hand, th e effect of a changing climate itself has the potential to increase aver age ambient concentrations and the frequency of pollution episodes. At p resent, levels of PM and the photochemical pollutant O3 continue to exce ed the thresholds set by the European Air Quality Directives, albeit var ious mitigation strategies are introduced since the late 70s to reduce ( precursor) emissions. In this respect, most of the European emission red uction strategies do not take into account a change in climate condition s. As in the future O3 and PM are likely to remain of concern, this doct oral research aims to estimate the effect of future climate change on O3 (and PM) concentrations, assuming that future emissions remain at the p resent-day level. This research develops a statistical downscaling method as a combination of two key techniques: 1) the objective Lamb circulation pattern classification describing the large-s cale synoptic conditions and 2) a linear multiple regression analyses, r elating surface meteorology to ozone concentrations. First, the circulat ion technique based on sea level pressure data describes for a given day the location of the high and low-pressure centres. After the classifica tion, all members of each circulation-type class are ascribed to a linea r regression equation, this for each season separately, taking into acco unt a range of meteorological surface variables for a rural mid-la titude site in the Netherlands. The combination of the Lamb circulation pattern approach prior to a multiple linear regression is a new approach in the statistical downscaling community and analyses have shown to sig nificantly improve the observed O3 non-linear characteristics. On the ot her hand, our research shows that PM10 is less influenced by the conside red meteorological variables. Therefore, this technique is only applied on future climate scenarios for ozone. The application of this technique on various climate change scenarios fr om a single global climate model shows an overall increase of maximum 8 hourly mean O3 concentration with 2.5 to 6.5 and 6.1 to 10.9 µg/m³, for the 2051-2060 and 2091-2100 period respectively, against the presen t-day 10-year average of 55.2 µg/m³. This increase is even stronger when considering the summer season only. An increase in maximum temperature and shortwave radiation, associated with a decrease in cloud cover under the various future scenario assumptions are the main drivers of ozone increase. In order to verify whether thes e findings are physically plausible, our results are compared with the o bserved heat wave in August 2003, characterized by a poor air quality (especially in terms of O3) and excess of mortalities in large parts of Western and Central Eur ope. Here, the observed O3 concentration during August 2003 exceeded the European Air Quality Guidelines in Cabauw for 9 days, which lies in the upper range of our estimated number of exceedences for the 2050s and corresponds closely to the mean number of exceedences by the e nd of this century (9.7 days month-1). This is confirmed by the exceptio nal warm and dry weather during August 2003, which characteristics correspon ds to the climate change scenarios by the end of this century. This sugg ests, that, not only in terms of temperature, but also in terms of O3, t he August 2003 conditions could become representative for our future cli mate.