The Czech legislation takes over general approaches of air quality assessment and potential exceedances of the set limit values in the zones from the EU directives for air quality management with the aim to reach, in the set deadlines, air quality complying with the limit values and target limit values. The legislation specifies that the assessment of air pollution level is carried out by measurements in agglomerations and areas where the level o air pollution reaches or exceeds the upper assessment threshold, and by measurements in the areas where the level of air pollution caused by ozone exceeds the long-term objectives (during the recent 5 years); further it is carried out by modelling or experts estimates in the areas where the level of air pollution by a pollutant does not exceed the lower assessment threshold; and finally by the combination of measurements and modelling in the areas where the level of air pollution reaches or exceeds the lower assessment threshold and simultaneously is lower than the upper assessment threshold.

Air pollution levels determination must cover the whole assessed area not only the nearest surroundings of the monitoring station. The air quality assessment in zones and agglomerations – particularly identifying and locating areas in which limit values may be exceeded, based on measurements – therefore becomes a problem of estimating the spatial distribution of air pollution extent; it consists in how to generalise �point� measurements, given the particular density and distribution of monitoring stations and an acceptable error of the estimate, to the entire territory under review. The spatial coverage of measurements can be increased by validation measurements. However, the ambient air quality directive and consequently, the national legislation, do not stipulate measurements any longer as the only tool for determining levels in a zone, and envisages – depending on pollution levels – the use of modelling techniques and expert estimates and their combinations. An advantage of modelling is that in comparison with point measurements it better reflects the coverage of the area under review; nevertheless, models are generally regarded as less accurate than measurements. Under modelling mainly causal dispersion and transport models are understood, including chemical transformations of the pollutants. An important role is played also by empirical, mathematical-statistical models of the estimate of time or spatial distribution of air pollution characteristics.

The maps of air pollution characteristics and atmospheric deposition are constructed by integrating the GIS system, ISKO relational database of the measured air pollution values and chemical composition of atmospheric precipitation, and the results of modelling based mainly on emissions, which is possible by using the high-performance hardware and the latest software. The important role is also played by supplementing and correcting the objective calculations on the basis of expert estimates made by the authorised institution. Using these methods we are able to carry out air pollution assessment in a very good quality and to create adequate user-friendly visualizations and presentations, both for administrative bodies and for specialists and general public.

In addition to the results of direct measurements of air pollution concentrations the results obtained from modelling are also used. For the territory of the Czech Republic the Gaussian dispersion model SYMOS 97 is used which calculates the concentrations on the basis of detailed emission inventories and data on meteorological conditions relevant for the assessed calendar year. Geomorphologically, the territory of the Czech Republic is divided into 47 areas which have different meteorological conditions. Each of the area is characterized by a wind rose, one of the inputs into the model. The calculation includes the latest available information on air pollution sources from the ISKO emission database and information on emissions from line sources. Apart from the sources on the territory of the Czech Republic the calculation includes also the available information on emission from sources abroad which plays an irreplaceable role in calculating concentrations in border areas but can be applied in the regions located further from the borders as well.
In addition to the dispersion model in some cases (e.g. ozone) the empirical model, using the quantities showing the regression dependence of the measured concentrations (such as altitude), is applied.
One of the important preconditions for creating fields of concentrations is a careful selection of the measuring stations included in the assessment, from the perspective of their use, classification and representativeness.
When preparing charts and maps of air pollution and deposition loads on the country�s territory, geostatistical procedures and map algebra tools of the GIS system are applied to estimate the fields of air pollution and deposition characteristics derived from point (station) measurements.

For the creation of the result maps assimilation of the measured and modelled data (or further supplementary data) is applied with the use of linear regression dependence of the respective quantities (measurement and model, or altitude) with subsequent interpolation of residues of this regression. In interpolation the modified version of IDW is applied (interpolation by a weighted mean of the values measured around the interpolated point, where the weight is a function of inverse distance between the interpolated point and the point of measurement) with the determination of its representative surroundings of the stations, or the interpolation kriging method (interpolation by a weighted mean of the values measured around the interpolated point, where the weight is a function of a statistic structure of the air pollution, resp. the deposition characteristics). In this case the whole method is applied separately for the urban areas and for rural areas (with the subsequent merging with the use of the map of population density). Both of the above mentioned interpolation methods enable to estimate the value of the monitored characteristic in every point of the field. If the field is statistically homogeneous [1], the estimation by means of the kriging method is optimal in that sense, that it is unbiased and its mean square error is minimal. When the kriging method is applied, the GIS software makes it possible to calculate errors of the estimation. Values of these errors show, among others, the efficiency of the enhancement of the density of the monitoring stations network and vice-versa.

The basic approach to determine the degree of representativeness is station classification. Background stations (�rural� or �urban background�) with a high degree of representativeness (dozens of kilometres) are stations affected only by remote sources; to describe local conditions stations exposed to traffic and industry (�traffic� and �industrial�) with the least area of representativeness directly affected by local sources are taken into account.

The creation of the basic geographic and topical layers in standardised projection (conform Gauss-Kr�ger projection) was launched in 1994. The DMÚ 200, DMR-2 and newly DMÚ25 digital layers are used to form the basic layers of the GIS: orography, the most important watercourses, water areas, settlements, administrative borders of districts, highway networks, and the vegetation cover.