Continuous four-dimensional source attribution for the Berlin area during two days in July 1994. Part II: Sensitivity studies

Abstract

The multiple nested three-dimensional (3D) mesoscale Eulerian grid point model MM5 is directly coupled with a Lagrangian particle trajectory model in order to perform a four-dimensional source attribution for the area of Berlin based on the horizontal distribution of the import probability density (IPD). The technical aspects are already demonstrated in the companion paper A (Part I) including the illustration of the meteorological situation at the two consecutive days of investigation and a primarily 3D source attribution. We conducted further sensitivity studies concerning the effect of vertical mixing, the static stability of the particles/emissions and the regarded time scale on the IPD distribution which is extended to four dimensions. The main results are: • Heterogeneity and temporal variability of the wind field enhance the contributions of nearby sources (emissions) to the total import of the receptor in contrast to stationary wind fields which increase the scope of the IPD distribution in the upstream direction. • Regions of static stability, for example morning hour inversion layers, enhance the contribution of far distance sources with longer import times. • The import velocities increase, as far as long distance source-receptor transitions are concerned, because they are mostly realised via higher transport paths. • The third (vertical) dimension is not negligible for the task of a complete source attribution, as a considerable amount of elevated emissions, preferably out of 300±100 m elevation, reaches the receptor box which is only 50 m in depth. Hence, downward mixing of elevated and far distance sources is an important process and driven by the diurnal course of turbulence and low level jets within the PBL • On the short time scale (few days), the source attribution is not independent from the regarded time scale (simulation time) due to the neglecting of older emissions released before the beginning of the simulation.