Cloud property retrievals from satellite data using thermal wavelengths in daytime and nighttime

Algorithms that routinely retrieve macro- and microphysical cloud properties from various passive satellite instruments have been routinely employed for almost three decades. These techniques, particularly those employed during daylight hours, have been well-documented and -validated, proving their worth in a range of applications, e.g., weather forecasting, aviation safety, field experiments and model assimilation studies. Cloud properties, particularly optical and microphysical retrievals, derived from nighttime imagery have been less rigorously validated and, due to the lack of information in the visible portion of the spectrum, the development of new techniques has been limited. This paper will present results from and enhancements to NASA Langley's Solar infrared-Infrared-Split window Technique (SIST). The SIST has been applied operationally to imagery from both geostationary and near-polar orbiting instruments for many years, but the ability of SIST to perform appropriately during daylight hours has not been explored nor have the inconsistencies of retrieved cloud properties across day-night boundaries been addressed. Other groups, e.g. the Moderate Resolution Imaging Spectroradiometer (MODIS) science team, have developed robust thermal-only techniques that perform well both day and night, but these algorithms generally derive only cloud temperature and height by taking advantage of information in CO2 channels; they do not simultaneously produce optical and microphysical properties as SIST does due to its usage of the reflected portion of the solar infrared channels. The ability of SIST to retrieve accurate and self-consistent cloud optical depths, particle sizes and water contents both day and night will be explored as will results from sensitivity studies that judge the technique's accuracy when presented with a priori cloud temperatures, variable surface properties and calibrations. Early indications are that SIST is quite sensitive to instrument calibration which can substantially impact the accuracy of retrieved properties. Results from SIST's application and sensitivity to different narrowband channels, e.g., from the MODIS, GOES and SEVIRI instruments, will be also shown.