| dc.description.abstract | Aerosol-cloud interactions have been studied since the 1970s. The interactions of the processes, such as aerosols’ direct, semi-direct, and indirect effects on clouds in the atmosphere, are complex. There are large uncertainties in resolving the impact of aerosols on clouds in global circulation models. One area that has been identified with some of the largest uncertainties is the ubiquitous stratocumulus over the Benguela Region. Understanding how aerosols influence the lifetime and characteristics of the stratocumulus has not been fully explored in the Benguela Region. This study seeks to evaluate the representation of simulated aerosols and their effect on stratocumulus clouds in the Benguela Region in the Conformal Cubic Atmospheric Model (CCAM). The first aspect evaluated how CCAM simulates the magnitude, spatial pattern and seasonal cycles of total aerosol optical depth (AOD), as well as compositional AOD (i.e. black carbon, organic carbon, sulphate, small dust and sea salt) against satellite observations from a Moderate Resolution Imaging Spectroradiometer (MODIS) and a reanalysis product (i.e. Modern-Era Retrospective analysis for Research and Applications (MERRA)). The second aspect evaluated how CCAM simulates the magnitude, spatial pattern, and seasonal cycle of cloud fraction and liquid water path against satellite observations. The third aspect examined the seasonal spatial correlation between simulated aerosol optical depth, and cloud fraction and liquid water path using the CCAM model. To carry out the analysis for all three aspects, numerous tests were used, including eyeball verification as well as bias, spatial root mean square error and spatial correlation (r). The time period that the study considered was 2002 to 2017.
In general, the simulated total AOD in CCAM was lower compared to MODIS total AOD across all seasons. The aerosol optical depth attributable to black carbon and organic carbon, which peak during biomass burning, was underestimated by CCAM when compared to MODIS or MERRA. During biomass burning season, there is a positive linear correlation between aerosol optical depth attributable to black carbon and organic carbon, indicating that there is a linear association with how CCAM simulates aerosol optical depth compared to MODIS and MERRA. Aerosol optical depth attributable to sulphate
was overestimated by CCAM when compared to MERRA. Dust aerosols (i.e. small dust) and marine aerosols (i.e. sea salt) were underestimated by CCAM when compared to MERRA. This study also found that aerosol optical depth attributable to black carbon and organic carbon had interannual seasonal peaks in the spring in CCAM, MODIS and MERRA. Dust aerosols (i.e. small dust) and marine aerosols (i.e. sea salt) had different interannual seasonal peaks in different seasons within CCAM and within MERRA. Furthermore, dust aerosols and marine aerosols had different interannual seasonal peaks when comparing the CCAM simulation to MERRA.
Secondly, this study found that CCAM underestimates low cloud fractions at the northern region and the coast for all seasons, and that CCAM overestimates cloud fractions in the southern region of the domain when compared to satellite observations. Furthermore, CCAM has a magnitude of error ranging between 40 and 60% at the northern region and 0 and 30% in the southern region of the domain compared to satellite observations. It was also found that when aerosol optical depth was >0.22 in CCAM, there was no correlation (r = -0.19 – 0.19) between aerosol optical depth and cloud fraction. When aerosol optical depth was <0.22, aerosol optical depth and cloud fraction were weakly positively correlated in the northern region of the domain (r = 0.20 – 0.39). This study found that CCAM underestimated the liquid water path over the ocean when compared to satellite observations. Also, CCAM produced a magnitude of error ranging between 10 and 60 kg/m2 when compared to satellite observations. Cloud fractions or liquid water path correlations between CCAM and satellite observations were more often weak positive (r = 0.20 – 0.39) than negative, although not strong. It was also found that when any correlation between cloud fraction or liquid water path and aerosol optical depth was found between CCAM and satellite observations, it was more often positive in the northern regions and negative in the southern domain.
Thirdly, this study found that aerosol optical depth attributable to black carbon and organic carbon had moderate or strong negative correlations (r = -0.79 - -0.40) with cloud fraction or liquid water path in winter within CCAM. Within CCAM, aerosol optical depth attributable to sulphate had strong positive correlation (r = 0.60 – 0.79) with cloud fraction in summer. Dust aerosols and marine aerosols had no strong correlations (r = -0.19 – 0.19) with cloud fraction within CCAM. Interannual seasonal plots for total aerosol optical
depth and compositional AOD and cloud fraction are similar, though not always within CCAM. Within CCAM, aerosol optical depth attributable to sulphate had no strong correlation (r = -0.19 – 0.19) with liquid water path. Dust aerosols had no correlation (r = -0.19 – 0.19) with liquid water path within CCAM. Marine aerosols had strong positivecorrelations (r = 0.60 – 0.79) in autumn and winter, and strong negative correlations (r =-0.79 - -0.60) in spring with liquid water path within CCAM.
The main findings of this study relative to the objectives were: (i) CCAM simulates the spatial distribution and peaks of aerosols, cloud fraction and liquid water path well, (ii) aerosols do influence cloud fraction and liquid water path, either directly proportionally or indirectly proportionally, and (iii) the influence of cloud fraction and liquid water path is not due to aerosols only, but also other parameters. | en_US |
