Satellite microwave measurements of ice cloud properties

Licentiate thesis, 2007

Ice clouds play an important role in Earth's radiation balance and climate because they reflect incoming sunlight and trap outgoing infrared radiation. There is still an uncertainty about how ice clouds will respond to the effects of global warming. Many climate models use prognostic cloud mass schemes to represent cloud feedbacks. The problem is that there exist little observational data of cloud ice mass to validate this part of climate models. Microwave radiometry is a promising technique for cloud ice mass observations since the measured signal is more proportional to the cloud ice mass, compared to other conventional techniques. A first retrieval scheme and results of cloud ice mass observations from the first satellite instrument operating in the sub-mm part of the electromagnetic spectrum (Odin Sub-Millimetre Radiometer) are presented. A comparison to climate models show important differences in corresponding data. The retrieval of ice cloud properties in a Bayesian regression framework is the main focus of this study. This methodology has been studied with the aim to determine the retrieval performance of a suggested sub-mm instrument, dedicated to measure cloud ice properties. Bayesian regression methods require an a priori database of realistic cloud scenarios and corresponding simulated measurements, where relevant instrument characteristics are incorporated. Inversion of a measurement can be performed by interpolating between the cases in the database that approximately match the measurement or by applying a neural net to perform the mapping from measured data to cloud ice properties. The main achievement here is that a general method to combine prior information from different sources to generate retrieval databases has been developed.