Can precipitation suppression due to anthropogenic aerosol pollution be detected using back trajectory methods?

The WMO statement on the status of weather modification suggests that there is ample evidence that human activities such as biomass burning, agriculture and industry have the capacity to modify local and regional weather conditions. A claim to such an effect is that "air pollution must be an important factor in determining precipitation amounts in the [Australian] Snowy Mountains" (Rosenfeld, 2000). This claim is made with reference to aerosol pollution sources that are quite remote from the region, based on satellite observations of cloud droplet effective radius, and perhaps unsurprisingly, has sparked a degree of controversy (Ayers, 2009). It should be noted, however, that the post-frontal airmass typically originates in the Southern Ocean and is among the most pristine in the world, so any suppression effect will be most clearly seen in this region.

To address the question of whether precipitation has been suppressed by aerosol pollution can be identified in surface precipitation records, back trajectories arriving at Automatic Weather Station (AWS) sites within the Australian alpine region during precipitation events have been calculated. Back trajectories for air parcels arriving at three hour intervals were calculated using the Hybrid Single Particle Lagrangian Integrated Trajectory Model (HYSPLIT; Draxler and Hess 1998), driven by 12.5 km resolution Meso-LAPS forecast meteorological data. Trajectories are classified according to synoptic type and an airmass-precipitation climatology for each arrival point is generated by convolving the trajectories with carefully quality controlled sub-hourly precipitation records from a high density network of pluviometers in the Snowy Mountains region.

The precipitation associated with these trajectories is to first order dominated by the effect of orography, in particular showing signs of suppression where there is elevated terrain upstream, and enhancement when flow aligns with the local gradient of the orography of the Great Dividing Range. The effect is pronounced for higher precipitation rates, as shown in figure 1. The effect of a hypothetical upstream pollution "source" is investigated to characterise the sensitivity of the detection procedure. During periods where precipitation is primarily driven by orographic forcing, the hypothesis that pollution from the sites identified by Rosenfeld (2000) is suppressing precipitation in the Snowy Mountains region is directly tested.

Position probability distribution for all trajectories (lower right) and difference compared to trajectories associated with threshold amounts of 0.2, 1.0 and 3.0 mm over the preceding three hours. The number of trajectories used for each composite is denoted by n, and terrain elevation shown by contours at 500 and 1000 m.