Liquid Water Cloud Retrievals - A Bayesian Approach

We developed a new algorithm to retrieve properties of non-precipitating liquid water clouds from millimeter wave radar and Microwave Radiometer (MWR) data using Bayes’ theorem of conditional probability. Bayes’ theorem relates the inverse problem (retrieving cloud properties from remotesensing observations) to the forward problem (modeling remote-sensing observations given a set of cloud properties). It also formally includes prior information about cloud microphysics, with this information explicitly modeled by a probability distribution function in the parameter space, not hidden in assumptions within the algorithm. The Bayesian algorithm does not make any assumptions about the shape of the cloud particle sized distribution (PSD), it is not limited to stratus-type clouds, and it provides uncertainties on each retrieved quantity.