Date of Award

5-2018

Degree Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Earth, Atmospheric, and Planetary Sciences

Committee Chair

Dr. Harshvardhan

Committee Member 1

Peter Colarco

Committee Member 2

Yutian Wu

Committee Member 3

Mike Baldwin

Committee Member 4

Greg Michalski

Abstract

Optically thick smoke aerosol plumes originating from biomass burning (BB) in the southwestern African Savanna during the austral spring are transported westward by the free-tropospheric winds to primarily overlie the vast stretches of stratocumulus cloud decks in the South-East Atlantic. We evaluated the simulations of long-range transport of BB aerosol by the Goddard Earth Observing System (GEOS-5) and four other global aerosol models using Cloud-Aerosol Lidar with Orthogonal Polarization (CALIOP) observations over the complete South African-Atlantic region. Models in general captured the vertical distribution of aerosol over land but exhibited some common features after long-range transport of smoke plumes that were distinct from that of CALIOP. Most importantly, the model simulated BB aerosol plumes quickly descend to lower levels just off the western coast of the continent, while CALIOP data suggests that smoke plumes continue their horizontal transport at elevated levels above the marine boundary layer. The levels to which the aerosol plumes are subsided, and the steepness of their descent vary amongst the models as well as amongst the different sub-regions of the domain. Investigations into possible causes of differences between GEOS-5 and CALIOP aerosol transport over the ocean revealed a minimal role of aerosol removal processes in the model as opposed to model dynamics.

It has been well established that the cloud adjustments to aerosol-radiation interactions are strongly dependent on the relative location of the aerosol layer with respect to the clouds. Consequently, we evaluated the GEOS-5 sensitivity to changes in aerosol vertical distribution by constraining the model smoke aerosol vertical profiles using CALIOP observations. A climatology of CALIOP retrievals of smoke aerosol extinction profiles were obtained using ten years of data (2006-15). An aerosol redistribution methodology was then developed to vertically adjust the smoke aerosol mass in the model to resemble the CALIOP extinction profile, such that the column aerosol mass is conserved. Three sets of experiments, each containing five ensemble members were designed for GEOS-5 AGCM in free-running mode by prescribing aerosols at each model time-step using an aerosol climatology i.e. MERRAero (2003-14). First is the control case (CTL), where aerosol vertical distributions were based on the default MERRAero climatology. Second set used the redistributed aerosol climatology (RED) and for the third set, smoke aerosols were simply removed from the oceanic parts of the domain (NOA). There was an increase in cloud fraction by about 40% for RED compared to NOA at areas of high aerosol loading near the coast. Between RED and CTL, there was an increase in cloud fraction near the coast by ~35% with respect to RED and a decrease in cloud fractions by ~25% for areas away from the coast and warmer sea-surface temperatures. Overall, the absolute magnitudes of changes in cloud fractions are small, but the percentage changes are large because the model simulated cloud fractions for the CTL case are much smaller and spatially displaced compared to the cloud fractions retrieved from MODIS to begin with. Probable mechanisms for the observed changes in cloud amounts and MBL properties were investigated. Aerosol effects on TOA all-sky radiative forcing showed close resemblance to the pattern for cloud fraction change, wherein increase in cloud cover led to enhanced cooling and vice versa. Aerosol impacts on surface radiative forcing however, suggested a strong cooling of the ocean surface irrespective of where the aerosol layer is placed in the atmosphere.

Share

COinS