Introducing hydrometeor orientation into all-sky millimeter/sub-millimeter assimilation

Abstract:

Oriented non-spherical ice hydrometeors are known to cause the observed polarization difference (PD) at millimeter and sub-millimeter wavelengths. This signal is driven by the different scattering properties in the vertical (V) and horizontal (H) polarizations. Currently, only the GPM (Global Precipitation Mission) microwave imager (GMI) measures ice with its dual-polarization capabilities at 166 GHz. The upcoming Ice Cloud Imager, with two channels (243.2 and 664.4 GHz) in both V and H will provide further insights. It aims at improving the representation of ice in models and it will extend the scope of ice cloud assimilation.

In assimilation systems, polarization due to oriented hydrometeors is currently ignored; only randomly oriented hydrometeors are considered, but this fails to explain such PDs. The Radiative Transfer model for TIROS Operational Vertical Sounder (RTTOV) that accounts for multiple scattering (RTTOV-SCATT), applies only “scalar” radiative transfer; one calculation deals with either V- or H- polarization, while in nature hydrometeors cause a blending of both. Moreover, it employs the same scattering properties for both simulations.

Herein, an effort is made to improve the physical representation of polarised scattering in RTTOV-SCATT. To model the effect of oriented ice hydrometeors in recreating the observed PDs from microwave radiometers, their optical thickness is increased in H and decreased in V channels by a chosen factor. The resulting ratio of extinction in the H and V channels is known as the polarisation ratio. Assimilation experiments were run using the Integrated Forecast System of ECMWF (European Centre for Medium-Range Weather Forecasts). To identify the ratio that induces the best fit between simulations and observation, passive monitoring experiments were run to simulate GMI radiances. Based on simulations of a month, the best fit was found at a ratio of 1.4, higher than found in earlier work. This reduced errors by 5-10K in high-frequency observations in cloud-affected scenes. To assess the impact of this ratio on the forecast, cycled data assimilation experiments were run using polarized scattering simulations for the Advanced Microwave Scanning Radiometer-2, GMI, and Special Sensor Microwave Imager/Sounder for a period of 6 months. Initial results are presented.