Characteristics of the Invariant Imbedding T-Matrix Method for Hydrometeor Scattering Calculations

In the pursuit of an accurate and efficient computational method for obtaining the single-scattering properties (SSPs) of inhomogeneous hydrometeors, we have implemented, parallelized, and optimized the Invariant Imbedding T-Matrix (IITM) solution method. We report the computational characteristics of IITM as compared to the Discrete Dipole Approximation (DDA) method, when applied to inhomogeneous hydrometeor models.

The key quantities in the retrieval of particulate matter by remote sensing means are the SSPs of the particles involved. Being able to compute the SSPs efficiently is thus of great interest and value. For a given applicable solution method, the computational cost is highly dependent on the complexity of the hydrometeor shape and the nature of any inhomogeneities. Although the shape of the hydrometeors are in general irregular, they can sometimes be adequately approximated with simplified shapes, such as spheroids.

We therefore compare the computational efficiency of IITM and DDSCAT, an open source DDA code, primarily by using inhomogeneous spheroids. Two models of inhomogeneity are examined: 1) multilayered, concentric spheroids with slow varying dielectric constant between layers and 2) similar to those in 1) but with abrupt discontinuity in dielectric constant at the outermost layer. The first kind is often used to approximate snow or other icy hydrometeors having lower and varying apparent density (known as fluffy spheroids) than solid ice, whereas the second kind is a valid model for melting graupels having a layer of water around a porous ice core. In addition to their respective strengths and weaknesses in computational efficiency, another emphasis of the comparison is the backscattering cross section, the most important SSP in simulating radar reflectivity.

Finally, we will discuss the potential of IITM, such as the feasibility of applying it to obtain SSPs for general, irregular hydrometeors.