Studying Interactions Between Smoke and Clouds

ARM Mobile Facility data are used to understand interactions between smoke particles from biomass burning, clouds, and sunlight.

The Science

Smoke produced by biomass burning in continental Africa is blown westward over the southern Atlantic for approximately one third of the year, from June to October.  This smoke coincides with an extensive deck of low-lying stratocumulus clouds in this region.  Smoke typically absorbs sunlight and heats the atmosphere while low-level marine clouds typically reflect sunlight and cool the atmosphere.  Understanding how these two effects will counter each other and the impacts that will have on regional climate and precipitation patterns depends, in part, on whether the smoke and cloud particles physically interact, or whether the smoke always occurs in a layer above the clouds.  A series of field experiments in this region, including the deployment of the ARM Mobile Facility (AMF) to Ascension Island for 17 months, attempt to address these questions.

The Impact

First findings from the ARM Mobile Facility deployment to remote Ascension Island (midway between Africa and South America in the Atlantic Ocean) indicate that smoke is present much more often near the surface than was previously thought. The new measurements from the 17-month-long ARM campaign suggest that August is the smokiest month near the surface. The smoke includes other aerosols besides black carbon and is most absorptive of sunlight in June and least in October. The smoke is more present near the surface earlier in the biomass burning season, or June, while later on toward September and October, more of the smoke resides above the cloud layer. This has implications for which aerosol-cloud microphysical and radiative interactions are dominant during different times of year. The maximum aerosol loading event observed during the campaign is investigated and found to be due to an unusual direct westward flow from the continental African fire sources at low altitudes.


Observations from June to October 2016, from a surface-based ARM Mobile Facility deployment on Ascension Island (8°S, 14.5°W) indicate that refractory black carbon (rBC) is almost always present within the boundary layer. The rBC mass concentrations, light absorption coefficients, and cloud condensation nuclei concentrations vary in concert and synoptically, peaking in August. Light absorption coefficients at three visible wavelengths as a function of rBC mass are approximately double that calculated from black carbon in lab studies. A spectrally-flat absorption angstrom exponent suggests most of the light absorption is from lens-coated black carbon. The single-scattering-albedo increases systematically from August to October in both 2016 and 2017, with monthly means of 0.78 ± 0.02 (August), 0.81 ± 0.03 (September), and 0.83 ± 0.03 (October) at the green wavelength. Boundary layer aerosol loadings are only loosely correlated with total aerosol optical depth, with smoke more likely to be present in the boundary layer earlier in the biomass burning season, evolving to smoke predominantly present above the cloud layers in September-October, typically resting upon the cloud top inversion. The time period with the campaign-maximum near-surface light absorption and column aerosol optical depth, on 13-16 August 2016, is investigated further. Back trajectories that indicate more direct boundary layer transport westward from the African continent are central to explaining the elevated surface aerosol loadings.

Principal Investigator(s)

Paquita Zuidema
University of Miami


This study was supported by the Office of Biological and Environmental Research (BER) of the Department of Energy (DOE) through the Atmospheric System Research (ASR) program and the Atmospheric Radiation Measurement (ARM) user facility.


Zuidema P, A Sedlacek, C Flynn, S Springston, R Delgadillo, J Zhang, A Aiken, A Koontz, and P Muradyan. “The Ascension Island Boundary Layer in the Remote Southeast Atlantic is Often Smoky.” Geophysical Research Letters 45(9), 4456-4465 (2018). [DOI:10.1002/2017GL076926]