This algorithm is being run operationally at the University of Utah on the available data. The plots (see an example below) show the radar reflectivity of the cirrus layer in the height-time domain. The derived cloud properties are shown in the remaining panels. The ice water content (IWC) in mg/m3 is shown on a semi-log plot in the second panel. The third panel shows the effective radius (R_eff, black) and total cloud particle concentration (Concen, blue). The units are microns and number/liter for the effective radius and concentration, respectively. The fourth panel shows the layer emittance (black) and visible optical depth (blue). The emittance is a derived property of the algorithm while the visible optical depth is computed from the ice water path and effective size using the radiation parameterization of Fu and Liou (1993, see the references in the algorithm description link, above). The same radiative parameterization is used to determine the asymmetry parameter and single scattering albedo. These quantities are then used in a radiation transfer algorithm (Toon et al., 1988) to compute the surface flux. This computed surface flux is scaled by the a clear sky flux. These results, compared to observations are shown in the bottom panel with the calculations in black and the observations in blue. The algorithm of Long, (1997) are used to estimate the observed clear sky flux. Finally, just right of the bottom panel are means and standard deviations of the calculated and observed scaled fluxes.

This algorithm is implemented only in specific cases. The cirrus must be optically thin (emittance less than 1.0) and no lower clouds can exist below the cirrus layer. Practically, the information content of the downwelling infrared radiation used in the scheme decreases substantially as the emittance of the layer becomes much greater than 0.8. Therefore, events that have an emittance greater than 0.8 should be used cautiously. This scheme combines the infrared radiance data from a vertically viewing spectrometer (the AERI) with the radar reflectivity. Since the AERI views the sky for approximately 3 minutes out of every 8 minutes, the retrieval is valid for a three minute period during which the AERI observes the sky. For output, the 3 minute averaged data is interpolated to the MMCR data resolution of 30 seconds. Thus, the plots show constant values during the 3 minute dwell time of the AERI followed by 5 minutes of no retrievals. In plots that show short-lived cirrus events, the symbols for each individual retrieval extend at a constant value during the three minute period. In plots of cirrus events that are longer than a few hours, the symbols appear to be point measurements at a particular time although they still represent 3 minute averages followed by 5 minutes of no values.

The output files are written in netcdf format and are available upon request to Jay Mace (mace@atmos.met.utah.edu or at 801-585-9489). The period of data availability can be found by checking the data availability page.