Evaluation of satellite and radar cloud retrieval methods during IMPROVE-2 icing events

The presence of supercooled liquid cloud droplets poses a serious hazard to aircraft operations. Current techniques for detecting and forecasting the presence of aircraft icing conditions rely on radar, satellite, and surface observations as well as model predictions. Meteorological conditions related to aircraft icing are well-documented; both passive and active remote sensing techniques are useful for characterizing a subset of those conditions. Satellite-based methods using combinations of visible reflectance and infrared emittance detect supercooled liquid water near the tops of opaque clouds, and hence can provide useful information for icing detection schemes. Radar retrieval methods that rely on fuzzy logic techniques yield particle habit classifications, providing further detail about cloud microphysical characteristics. In this paper, we investigate the capabilities of satellite-based cloud phase retrieval schemes based on data from the Geostationary Operational Environmental Satellite (GOES-12) imager and the NOAA Polar Orbiting Environmental Satellites (POES) Advanced Very High Resolution Radiometer (AVHRR/3). Radar-derived particle classifications from the NCAR S-band dual-polarization Doppler radar (S-Pol) are also examined.

In situ measurements of cloud top properties in icing conditions are essential for evaluating the efficacy of satellite- and radar-derived cloud and icing products. The Improvement of Microphysical Parameterization through Observational Verification Experiment (IMPROVE) provided an array of in-situ observations suitable for assessment of remote sensing techniques. Here we examine three cases from IMPROVE-2, conducted in west central Oregon during the fall of 2001. Satellite, aircraft, and radar measurements are employed for detection and characterization of aircraft icing conditions. In two of the cases, supercooled liquid was encountered by the research aircraft. A third case, in which a completely glaciated cloud was observed, was selected in order to test the retrieval algorithms in a non-icing situation.

Results show general consistency between the remote and in situ measurements. Cloud thermodynamic phase derivations from GOES-12 and AVHRR/3 algorithms suggest a combination of mixed- and liquid-phase conditions in the two cases where icing conditions were observed the aircraft. Radar-derived particle identification (PID) fields are also compared with aircraft in situ and satellite measurements. Ice crystals and irregular ice crystals are the primary particle type detected in the two cases where icing was observed by the aircraft. Due to the mixed-phase conditions, only a few radar pixels indicate the presence of liquid droplets. In the third case, satellite and radar retrievals indicate ice-phase particles exclusively, as was observed by aircraft in situ measurements. Detailed results and algorithm descriptions will be provided at the conference.