Analysis of aerosol-cloud-interactions over semi-arid and arid subtropical land regions from three different satellite datasets (MODIS, AATSR/ENVISAT, IASI)

Kurzfassung

Indirect aerosol effects, i.e. the change of cloud physical properties by aerosol interactions, have been identified as one of the largest uncertainties in the current understanding of the climate system. Despite the uncertainties of the representations of aerosol-cloud interactions in current climate projections, they have large impact on the climate system itself – in terms of the radiation balance, but also in terms of precipitation, and thus vegetation cover, and re-distribution of water throughout the atmosphere. Nevertheless, so far only very few studies of large-scale statistics of aerosol-cloud interactions over land are available. Moreover most studies on the topic cover liquid water clouds only. Aerosol cloud interactions over arid and semi-arid land regions have been analysed from three different satellite datasets with respect to aerosol type and cloud phase. The regions of the analysis cover Southern Africa, the Sahel domain with the influence of the West African monsoon circulation, the North-Western African Maghreb region and the Arabian Peninsula. These regions have been chosen as they are dominated by one (Maghreb, Arabia) or two (Sahel, Southern Africa) aerosol types and as mineral dust is one of the dominating aerosol types in all of them. The second dominating aerosol type is biomass burning in the Sahel and Southern Africa. These aerosol types can be discriminated by separating the aerosol information into fine mode (biomass burning) and coarse mode (desert dust) aerosol. Thus they can generally also be discriminated from satellite, although these capabilities are limited over land. Over land the diurnal cycle of convection is much stronger and aerosol interactions with deep convective cloud systems over land have been identified to be of great importance not only for precipitation in regions under pressure of desertification, but also with respect to climate change. For liquid water clouds the well-known first indirect aerosol effect (“Twomey effect”), i.e. higher cloud albedo due to smaller droplet sizes, could be confirmed for all regions, if liquid water path is held constant. Nevertheless, liquid water path has been found to be affected by aerosol presence and the aerosol effect on liquid water path dominates the net effect of aerosols on cloud optical depth. For ice phase clouds the same effects are observed with ice water path controlling the net aerosol effect on optical depth. From thermal infrared retrievals of mineral dust and ice clouds an increase of ice particle size with respect to background conditions has been detected. Together with observations at solar wavelengths the differences can be interpreted as indications for an increase of optically thicker clouds at the cost of cirrus coverage. Although the Twomey effect has been identified to be active in all cases, cloud water path and cloud phase transitions could be identified to be of predominant importance for resulting cloud propertiy changes due to aerosol presence. The second indirect aerosol effect (“Albrecht effect”) could not be identified from the statistical analysis. Although cloud cover distributions as functions of aerosol optical depth (AOD) indicate an increase of cloud cover with AOD, these could not be related to any other cloud properties including cloud droplet size. Thus the satellite observations do not support the relatively simple formulation of the second indirect aerosol effect (longer cloud lifetime due to drizzle suppression as a consequence of smaller droplets). An aerosol effect on cloud phase has been identified with respect to cloud water path. It could not be confirmed in terms of cloud coverage. The statistical analysis of cloud macro- and microphysical properties has been performed after the observations have been projected all to the same cloud top temperature distribution. This method allows correcting for effects of the temperature 4 and moisture fields (meteorological conditions), which otherwise would dominate the statistical results. It has been shown that aerosol type is important for aerosol cloud interactions in subtropical land regions. Moreover the cloud water path (liquid and ice) has been identified to be a strong constraint on indirect aerosol effects, outweighing e.g. the optical depth increase by droplet size reduction (“Twomey-effect”). It could moreover be shown that aerosol-cloud interactions are also important for ice cloud properties in subtropical land regions, which have yet not fully been addressed in statistical analyses of indirect aerosol effects and consequently in climate projections. Nevertheless, by means of the large-scale statistical analysis, also some deficits of current satellite datasets have been identified, which have to be solved in order to furthermore reduce the uncertainties of indirect aerosol effects. It has been the first attempt to quantify aerosol-cloud interactions focussed on semi-arid and arid land regions, performing the same kind of analysis to liquid water and ice clouds at the same time with the same methods, comparing results from three different independent satellite datasets, using advanced statistical descriptions of the observed deviations from background in order to account for non-linearity and multimodal or non-Gaussian probability distributions of cloud properties, applying a newly developed method to account for variations in cloud top temperature affecting cloud property observations statistically and also introducing a newly developed dataset from IASI which is sensitive to desert dust and ice clouds only, adding information about aerosol type sensitivity of aerosol-cloud interactions.