Evaluation of satellite-derived cloud properties in preparation for data assimilation

Clouds play a crucial role in the dynamics and thermodynamics of the atmosphere, yet in many aspects, they are poorly specified in numerical weather prediction (NWP) models. Satellite remote sensing can provide a real-time characterization of cloud properties with sufficient accuracy to improve the initial conditions for NWP. The NOAA Rapid Update Cycle (RUC) model is a key component of the FAA Aviation Weather and Research Program (AWRP) focused on improved aviation safety and flight planning over the continental USA (CONUS). The RUC cycles five cloud microphysical species (cloud water, cloud ice, rain, snow and graupel) at full resolution and has the capability for updating these fields from observations every hour. Currently, the RUC assimilates a variety of data to analyze cloud parameters, including a NOAA/NESDIS cloud top pressure product derived using a CO2 slicing technique applied to Geostationary Operational Environmental Satellite (GOES) sounder data. A new effort is underway to assimilate cloud products being derived operationally at NASA Langley Research Center. The NASA products are derived at 4 km resolution from the GOES imager using visible, near-infrared, and infrared window radiances. They include cloud height, phase and effective radius at or near cloud top, as well as cloud optical thickness and water path. A primary objective of this assimilation effort is to exploit the information contained in the satellite-derived cloud water path (total column integrated cloud water density), which offers the possibility to test, improve and constrain the model cloud physics. Another objective is to synthesize the NOAA/NESDIS and NASA products to produce a more accurate cloud cover and top height product for the model initialization. In order to develop the appropriate logic for these efforts, we perform an accuracy assessment of the NASA products. A comparison between the NASA and NESDIS cloud top height products is presented. The analysis utilizes cloud boundary information derived from surface-based cloud radar data to evaluate their respective uncertainties, strengths and weaknesses. Satellite-derived cloud water path uncertainties are summarized from recent intercomparisons with surface-based microwave radiometer estimates and estimates using radar/radiometer techniques. A particular emphasis is placed on comparisons with cloud radar estimates in order to pin down the satellite sensitivities for clouds with large optical and geometric thickness. Finally, the NASA cloud products are assessed in terms of how accurately they produce radiative fluxes at the surface and top-of-atmosphere (TOA) when incorporated as inputs into a radiative transfer model. A 3-year dataset is analyzed that consists of Fu-Liou radiative flux calculations, with NASA cloud properties as input, and co-located radiative flux measurements at the surface (obtained from ARM and SURFRAD sites) and TOA (from CERES observations). Flux biases are summarized for a wide range of cloud conditions.