Impact of biomass burning aerosols on the clear-sky shortwave radiative fluxes

Abstract

Aerosol radiative forcing is recognized as a key climate change component, and biomass burning is one of the largest sources of anthropogenic tropospheric aerosols. Biomass burning aerosols have both a direct radiative impact by scattering and absorbing solar radiation and an indirect radiative impact by serving as cloud condensation nuclei. Although our knowledge of the optical properties of smoke aerosols has been significantly improved during the past decade through field experiments, there still exists large uncertainty in both direct and indirect smoke radiative forcings at the Top-of-Atmosphere (TOA) and at the surface. Using satellite measurements, direct shortwave radiative forcings of smoke aerosols are estimated both at the TOA and at the surface. Optical thickness (τ) and single scattering albedo (ω0) are the two most important optical parameters of smoke aerosols in determining their radiative impact. A look-up table approach is applied to retrieve these parameters using Advanced Very High Resolution Radiometer (AVHRR) imagery from the NOAA platform. Using NOAA-14 AVHRR visible reflectances and ground-based τ measurements, the ω0 values of smoke aerosols are estimated for selected days during both the Smoke, Clouds and Radiation-Brazil (SCAR-B) experiment held in Brazil during 1995 and for selected days during the Zambian International Biomass Burning Emission Experiment (ZIBBEE) held in Africa in 1997. The retrieved average values of ω0 at a wavelength of 0.64 μm range between 0.83 to 0.91, which are in good agreement with in situ measurements and previous studies. Using the retrieved ω0 and observed τ values, surface downward shortwave irradiances (DSWI) are calculated over 4 sites in South America and Africa and are compared with the ground-based DSWI measurements. Comparisons show that for 10 out of 14 cases, the root mean square (RMS) errors between model calculations and observed DSWIs are within 30 Wm-2, with the largest error of 51 Wm-2. The results indicate that when ground-based values of aerosol τ are available, reasonable estimates of ω0 can be retrieved from AVHRR imagery which, in turn, then can be used to characterize biomass burning aerosols in radiative transfer calculations. The Angular Distribution Model (ADM), which is used to convert the measured solar radiances into fluxes at the TOA, is a key component to obtain accurate estimates of TOA fluxes. In this study, an ADM for smoke aerosols is calculated using a discrete ordinate radiative transfer model. Using collocated Visible and Infrared Scanner (VIRS) data and the Cloud and Earth Radiant Energy System (CERES) scanner data from the Tropical Rainfall Measuring Mission (TRMM) platform, instantaneous TOA shortwave (SW) fluxes are estimated using the new smoke ADM and compared with the SW fluxes from the CERES product for selected days over biomass burning regions in South America in 1998. The RMS error between the CERES SW fluxes and fluxes using the smoke ADM is 13 Wm-2. The new smoke ADM developed as part of this study can be used to estimate the radiative impact of biomass burning aerosols using satellite imagery. Surface SW fluxes can be estimated from the TOA observed SW fluxes using the radiative transfer model. In this study, a δ-four stream radiative transfer model has been modified to account for the biomass burning aerosols so that it can be used over biomass burning regions. Using in situ measurements of aerosol optical properties and ground-based measurements of aerosol optical thicknesses during the SCAR-B experiment, surface DSWI values in the presence of biomass-burning aerosols are calculated. These DSWI values are compared with broadband pyranometer measurements made at the surface. The results show that when near-coincident measurements of ω0 and τ are available, the root mean square errors between the measured and calculated DSWIs for daytime data are within 20 Wm-2. However, when assumptions about ω0 have to be made, the differences can be as large as 95 Wm-2. The results show that τ and ω0 are the two most important parameters that affect DSWI calculations. Finally, one month of VIRS and CERES data in August 1998 over biomass burning regions in South America and Africa have been processed to study the instantaneous and diurnally-averaged smoke radiative forcings both at the TOA and at the surface. The smoke τ values are retrieved from the VIRS visible channel images assuming the ω0 values of smoke aerosols. The TOA SW fluxes are calculated using the smoke ADM developed in this study, and the surface SW fluxes are calculated using the δ-four stream radiative transfer model. From the number of smoke pixels identified and from the retrieved smoke optical thicknesses, it is noted that the biomass burning season over South America is stronger than that over Africa in 1998. Analysis shows that when τ (0.64 μm) is 1.0, the direct instantaneous TOA Shortwave Aerosol Radiative Forcing (SWARF) is about –50 W-2 and the surface downward SWARF is about 210 Wm-2. The TOA SWARF per τ ranges from –35 Wm-2 to –50 Wm-2. For a τ value of 1.0 and TOA clear-sky albedo of 0.16, the diurnally-mean direct TOA SWARF ranges from –7 Wm-2 to –20 Wm-2, and the diurnally-mean surface downward SWARF ranges from 80 Wm-2 to 100 Wm-2 for ω0 values ranging from 0.80 to 0.90. The results from one month of data indicate that smoke aerosols have significant cooling effect at both the TOA and the surface.