LAUSR

Abstract The comparison of two methods in calculating transboundary concentrations of NO2 and PM10 using CMAQ chemical transport model is presented. The total mean annual concentrations for the pollutants were computed from the hourly outputs of the CMAQ model. The first method for calculating the transboundary concentrations is based on switching off the emissions from the studied country, while the remaining emissions are the same as in the full model run. The second method is based on running the model with the emissions from the studied country only. The resulting concentrations are then subtracted from the full model run concentrations. The result of this subtraction represents the transboundary air pollution together with pollution originating from the interaction of national and foreign sources. The pollution which cannot be attributed unambiguously to national or foreign sources is caused by the non-linear processes included in the model. It is evaluated as the difference between the two methods. It is shown that the non-linearity effect is more expressed for PM10 than for NO2 annual mean concentrations. It is estimated that the non-linearity effect for PM10 can reach values up to 2.7 μ g / m 3 in absolute value, or up to 16% of the total annual mean concentrations and up to 25% of the total estimated transboundary concentrations in the studied area. It is also demonstrated that this non-linearity effect is more important for both pollutants during some episodes than in the annual mean. The method of removing the bias from the calculated transboundary concentrations is presented, together with the proposed method of evaluation of the uncertainties of the transboundary concentration calculations. The estimated transboundary contribution is evaluated at the locations of Slovak air-quality monitoring stations. The impact of several emission reduction levels on the estimation of the transboundary contribution is also presented. The simulations are performed for the reference year of 2015 for Slovakia, but the proposed methods can by applied universally.