The mixing layer height (MLH) is a measure for the vertical turbulent exchange within the boundary layer, which is one of the controlling factors for the dilution of pollutants emitted near the ground. Based on continuous MLH measurements with a Vaisala CL51 ceilometer and measurements from an air quality network, the relationship between MLH and near surface pollutant concentrations have been investigated. In this context the uncertainty of the MLH retrievals and the representativeness of ground-based in-situ measurements are crucial. We have investigated this topic by using data from the BAERLIN2014 campaign in Berlin, Germany, conducted during June and August 2014. To derive the MLH three versions of the proprietary software BL-VIEW and a novel approach COBOLT were compared. It was found that the overall agreement is reasonable if mean diurnal cycles are considered. The main advantage of COBOLT is the continuous detection of the MLH with a temporal resolution of 10 minutes and a lower number of cases when the residual layer is misinterpreted as mixing layer. We have calculated correlations between MLH as derived from the different retrievals and concentrations of pollutants (PM<sub>10</sub>, O<sub>3</sub> and NO<sub><i>x</i></sub>) for different locations in the metropolitan area of Berlin. It was found that the correlations with PM<sub>10</sub> are quite different for different sites without showing a clear pattern, whereas the correlation with NO<sub><i>x</i></sub> seems to depend on the vicinity of emission sources in main roads. In case of ozone as a secondary pollutant a clear correlation was found. We conclude that the effects of the heterogeneity of the emission sources, chemical processing and mixing during transport exceed the differences due to different MLH retrievals. Moreover, it seems to be unrealistic to find correlations between MLH and near surface pollutant concentrations representative for a city like Berlin, in particular when traffic emissions are dominant. Nevertheless it is worthwhile to use advanced MLH retrievals for ceilometer data, e.g. for the validation of chemical transport models.