Ice hydrometeor profile retrieval algorithm for high-frequency microwave radiometers: application to the CoSSIR instrument during TC4

[摘要] A Bayesian algorithm to retrieve profiles of cloud icewater content (IWC), ice particle size (D_me), and relativehumidity from millimeter-wave/submillimeter-wave radiometers ispresented.The first part of the algorithm prepares an a priori filewith cumulative distribution functions (CDFs) and empirical orthogonalfunctions (EOFs) of profiles of temperature, relative humidity, threeice particle parameters (IWC, D_me, distribution width), andtwo liquid cloud parameters. The a priori CDFs and EOFs are derivedfrom CloudSat radar reflectivity profiles and associated ECMWFtemperature and relative humidity profiles combined with three cloudmicrophysical probability distributions obtained from in situ cloudprobes. The second part of the algorithm uses the CDF/EOF file toperform a Bayesian retrieval with a hybrid technique that uses MonteCarlo integration (MCI) or, when too few MCI cases match theobservations, uses optimization to maximize the posterior probabilityfunction. The very computationally intensive Markov chain Monte Carlo(MCMC) method also may be chosen as a solution method. The radiativetransfer model assumes mixtures of several shapes of randomly orientedice particles, and here random aggregates of spheres, dendrites, andhexagonal plates are used for tropical convection. A new physical modelof stochastic dendritic snowflake aggregation is developed. Theretrieval algorithm is applied to data from the Compact ScanningSubmillimeter-wave Imaging Radiometer (CoSSIR) flown on the ER-2aircraft during the Tropical Composition, Cloud and Climate Coupling(TC4) experiment in 2007. Example retrievals with error bars are shownfor nadir profiles of IWC, D_me, and relative humidity, andnadir and conical scan swath retrievals of ice water path and averageD_me. The ice cloud retrievals are evaluated by retrievingintegrated 94 GHz backscattering from CoSSIR for comparison with theCloud Radar System (CRS) flown on the same aircraft. The rms differencein integrated backscattering is around 3 dB over a 30 dB range. Acomparison ofCoSSIR retrieved and CRS measured reflectivity shows thatCoSSIR has the ability to retrieve low-resolution ice cloud profiles inthe upper troposphere.

[学科分类] 几何与拓扑