The 1 km retrievals, based on well-known solar reflectance techniques, include cloud optical thickness, effective particle radius, and water path, as well as thermodynamic phase derived from a combination of solar and infrared tests. Being both global and of high spatial resolution requires an algorithm that is computationally efficient and can perform over all surface types. Collection 5 additions include the development and use of spatially-complete global spectral surface albedo maps derived from MODIS, pixel-level uncertainties incorporating fundamental error sources, and the use of additional spectral and spatial tests to help eliminate false cloud detection (e.g., dust, snow/ice, sun glint) and partly-cloudy pixels. Results from a research-level multi-layer cloud detection algorithm are also included.
We will summarize the Collection 5 changes and discuss example retrievals obtained from the archive. The emphasis will be on pixel-level (Level-2) retrievals, but implications for the gridded (Level-3) product will also be discussed.