83rd Annual

Thursday, 13 February 2003
Cloud Liquid Water Profile Information in AMSU-B Water Vapor Profile Retrievals
Clay B. Blankenship, NRL, Monterey, CA
Poster PDF (564.2 kB)
A method to simultaneously retrieve cloud liquid water and water vapor profiles using data from the Advanced Microwave Sounding Unit-B (AMSU-B) radiometer is described. Results of simulated retrievals for simple liquid and mixed-phase clouds and global maps of retrieved cloud parameters are presented. The AMSU-B, with channels at 183.31±1, ±3, ±7, 150, and 91 GHz, can be used to retrieve water vapor profiles, taking advantage of the 183.31 GHz water vapor absorption line. The retrieval, a physical inversion using an optimal estimation technique, is extended to retrieve limited cloud liquid water profiling information as well.

A pseudohumidity variable RH' is defined, equal to relative humidity when RH<100% and linearly related to cloud liquid water density when clouds exist. (Saturation is assumed over the entire scene in this case.) An iterative retrieval of RH' is performed using the Jacobian with respect to RH'. A plane parallel, non-precipitating liquid or mixed-phase cloud is assumed with an upper limit of 0.5 mm of cloud liquid water. Scattering is neglected; the presence of precipitation will generally cause the retrieval to fail due to high TB error.

Retrievals were performed using observations simulated from a database of ECMWF analysis temperature, humidity, and cloud profiles. The cloud profiles were adjusted to be plane parallel by removing layers of low cloud fraction and assuming that the remaining layers had 100% cloud cover. Simulated retrievals are able to reproduce cloud top, bottom, and liquid water amount much of the time, with cloud top being slightly underestimated, cloud base slightly overestimated, and total cloud liquid water somewhat underestimated on average. Parameters retrieved from AMSU-B observations are mapped and compared with geostationary infrared images and the output of a visible/infrared-based cloud-classification algorithm.

Supplementary URL: