A063-0015
Smoke in the Amazon: isolating aerosol-radiation-interactions (ARI) and aerosol-cloud-interactions (ACI) on large scales using convection permitting resolution (ARI affects thermodynamics and convection, ACI affects radiative properties)

Wednesday, 9 December 2020
Poster
Ross Herbert, Philip Stier and Guy Dagan, University of Oxford, Department of Physics, Oxford, United Kingdom
Abstract:
The Amazon rainforest covers a vast region and plays a globally important role in generating oxygen and storing carbon. This region regularly experiences sustained (weeks to months) plumes of absorbing smoke that extend for thousands of kilometres during the annual dry season. The large spatial scale and duration of the perturbation makes modelling the response of clouds and precipitation to smoke challenging. Thus, the impact of wildfire smoke on the hydrological cycle and energy budget over the Amazon remains uncertain.

In this study we use the ICON regional model with a large 3000 x 2000 km domain and convection permitting resolution of 1.5 km to study the impact of smoke over a period of 8 days. To isolate the cloud response, smoke horizontal and vertical distribution is prescribed using the MACv2-SP aerosol plume model and observational measurements are used to produce a realistic perturbation to the cloud droplet distribution. The semi-idealised design allows us to identify the key processes that act to modify the cloud field and precipitation by isolating the radiative effects (ARI) and cloud microphysical effects (ACI) and to separate the cloud response from uncertainties in the aerosol field itself.

Strong diurnal cycles in the response to smoke across the domain are found: ARI predominantly impacts thermodynamics and deep convective processes, whereas ACI predominantly impacts cloud radiative properties. The large spatial and temporal scale used allows us to make robust conclusions that help us to understand the role that aerosol (both absorbing and scattering) has on regions dominated by strong diurnal cycles in convection. Our results also highlight limitations of current polar-orbiting observing systems and examine potential issues for models that parameterise convection and fail to reproduce the observed timing of its diurnal cycle.