Quantifying Uncertainty in the Response of Convective Precipitation to Aerosol

Abstract:

The effect of aerosols on cloud microphysics, dynamics and precipitation is important for weather-scale phenomena as well as cloud radiative forcing and climate change. However, because numerical models of clouds are complex and therefore slow to run, our understanding of aerosol-cloud interaction mostly relies on individual case studies and a very limited exploration of model uncertainties. Although we know that aerosols can substantially affect the thermodynamic and microphysical properties of convective clouds, we have little idea how robust our understanding is.

Here we show how the accumulated precipitation from a mixed-phase convective cloud responds to changes in aerosol concentration, accounting for uncertainties in the cloud microphysical processes. We do this using the statistical techniques of Gaussian process emulation and variance-based sensitivity analysis that enable the interacting effects of many model uncertainties to be simulated. As expected, the mean cloud response is an increase in precipitation with aerosol when aerosol concentrations are low and a decrease in precipitation when aerosol concentrations are high. However, when we account for the uncertainties in the model processes, the variability in the precipitation response to aerosols is sufficiently large that even the direction of response can no longer be defined with confidence, except in extremely polluted conditions.

The results of this research provide information on where efforts should be directed to reduce the uncertainty in modelling aerosol-cloud-precipitation interactions.