Assessing the Effects of Wildfires Aerosols on Clouds using Satellite Observations

Aerosol-Cloud-Interactions (ACIs) still represent a major source of uncertainty on climate predictions.
Satellites have greatly contributed to better understanding these effects due to their global and
continuous observations of the atmosphere. Consequently, numeroussatellite-based estimates of the
ACI radiative forcing (or Twomey effect) as well as rapid adjustments have been obtained over the last
decades, not always in agreement with each other. Natural experiments, which correspond to specific
or controlled pollution events have been particularly helpful to assess these effects from satellite.
However, satellite retrievals are not always adapted to quantify aerosol-cloud interactions. Haywood
(2003) hasinvestigated instrumental biases due to aerosol being above cloud layers(AAC), in particular
the impact of desert dust and biomass burning aerosol (BBA), both absorbing aerosols, on cloud
effective radius(CER or 𝑟𝑒𝑓𝑓) and cloud optical thickness(COT or 𝜏). He found that when a thick aerosol
layer is situated above clouds (AAC) a passive remote sensor like MODIS will underestimate 𝜏 and,
depending on the channel’s wavelength, overestimate, or underestimate 𝑟𝑒𝑓𝑓. Such effects can be
mistakenly attributed to ACI and must efficiently be reduced or corrected.
This study focuses on the Australian wildfires from 2019/2020 to observe and try to understand how
physical effects and instrumental effects are entangled when studying ACI. Aerosol and cloud
properties are obtained from several sets of data from several instruments: MODIS, TROPOMI, AMSR2, as well as ERA5 reanalysis. We only keep non-precipitant liquid clouds and we separated AAC cases
from non-AAC ones. We chose several areas in south hemisphere (Pacific, Atlantic and Indian Oceans)
for studying the evolution of ACIs with aerosols plume transport and decoupling each effect.
We obtained encouraging results, where instrumental biases have been observed under AAC
conditions; overestimation of 𝑟𝑒𝑓𝑓 and underestimation of 𝜏 were found over Pacific during DJF
2019/2020 and over Atlantic Ocean for the same period. But, imposing a strong cloud fraction, we
observed that biases were disappearing, which lets us think that MODIS might also misjudge aerosols
as clouds. However,satellite observations are limited, and we need for Radiative Transfer calculations.
Indeed, we will combine simulations (using RTTOV) with observations to better characterise and
correct the instrumental bias and then focus on physical aerosols impacts.