Satellite microwave measurements of cloud ice properties

Doctoral thesis, 2010

Upper tropospheric cloud ice and water vapour are two intimately connected components of the atmosphere, that influence the Earth's energy budget. The response of these components to a warming climate is not well understood. Global observations are crucial for improving our understanding of the time-varying climate, but the amount and quality of such observations of cloud ice and water vapour are limited. This thesis deals with satellite microwave measurements of cloud ice properties and humidity in several ways. The retrieval performance of a proposed passive combined millimetre and sub-millimetre wave instrument for cloud ice observation is studied. Furthermore, retrievals of cloud ice mass and humidity from Odin-SMR (Odin-Sub-Millimetre Radiometer) are presented. Additionally, Odin-SMR retrieval products are combined with correlative datasets to study diurnal variations. It is confirmed that the proposed instrument would be well suited to observe ice cloud bulk properties, such as the vertically integrated mass content and an effective cloud particle size. A novel and general retrieval methodology is developed and applied to Odin-SMR measurements of the tropical upper troposphere. The retrieval algorithm recreates vertically resolved cloud ice mass and relative humidity with respect to ice (RHi) with a precision of 65% and 5 - 17 %RHi, respectively. The retrieval algorithm takes into account of cloud inhomogeneities by a novel usage of radar data in order to reduce the systematic retrieval uncertainty. The retrieval products are shown to be in good agreement with correlative datasets from the CloudSat Cloud P rofiling Radar and Aura Microwave Limb Sounder. The diurnal cycle of water in the tropical upper troposphere is studied, by combining these datasets. Strong diurnal variations of cloud ice amount were observed primarily over land regions, with a pronounced maximum in the afternoon. It was found that climate models have a problem of correctly simulating the observed short time-scale variations.