On the spatial variability of the regional aerosol distribution as determined from ceilometers

Abstract. Measurements of the vertical distribution of aerosol particles are typically only available at selected sites leaving the question of their representativeness for urban and regional scales unanswered. As a contribution to solve this problem we have investigated ceilometer signals from two testbeds in Munich and Berlin, Germany. For each testbed measurements of 24 months from 6 ceilometers were available. This constitutes a unique data set, in particular as the same type of instruments are deployed and the same data evaluation schemes applied. Two parameters are discussed: the mixing layer height (MLH) as an indicator for the vertical distribution and the integrated backscatter as a proxy for the amount of aerosols in the mixing layer. The MLH was determined by the COBOLT algorithm, the integrated backscatter from the Klett (backward and forward) inversion scheme. It was found that the mean difference of the MLH at two sites within a testbed typically only varies by less than 50 m, slightly increasing with the distance of the corresponding sites. Almost 60 % of all intercomparisons agree within ±100 m. MLHs are typically correlated with R > 0.9 in particular for the Berlin-testbed. With respect to the integrated backscatter the correlation is in the range of 0.7 < R < 0.9. This is expected from the diversity of local aerosol sources within a given testbed. We conclude from our data that the MLH determined from a single ceilometer is applicable for a whole metropolitan area. However, the integrated backscatter of particles within the mixing layer exhibits a variability of 15–25 % suggesting that one ceilometer is not representative, especially if atmospheric processes shall be investigated.