Particle size as a state variable in the GPM combined algorithm

Abstract:

We present a space-borne radar profiling methodology based on the use of the mass weighted particle size, Dm, as a state variable. The formulation has been developed to mitigate uncertainties in the GPM radar-radiometer algorithm. Specifically, in situations characterized by low information content in the observations, e.g. light stratiform rain over land with no reliable estimates of the Path Integrated Attenuation (PIA), the GPM combined algorithm tends to resort to the default reflectivity precipitation relationships embedded in the algorithm’s scattering lookup tables. However, the current lookup tables do not account for the inverse relationships between the generalized particle size distribution intercept Nw and the mass-weight particle diameter Dm. The inverse relationships between Nw and Dm observed in ground observations are only weakly enforced in the current version of the algorithm through the application of an iterative procedure. More strongly enforced Nw-Dm relationships are possible in the current framework, but are not deemed to represent the best approach, as larger updates may affect the numerical stability of the algorithm. A better alternative is the derivation and use of scattering lookup tables that implicitly account for the relationships between Nw and Dm. As such tables may impact the continuity of precipitation fluxes through the melting layer, they are likely to work best in formulations based on Dm as a state variable, as the continuity of Dm combined with appropriate Particle Size Distribution (PSD) shape parameterizations is likely to enable appropriate preservation of precipitation flux continuity. In this study, we investigate a Dm-based radar retrieval methodology supported by scattering lookup tables that incorporate climatologic Nw-Dm relationships. A systematic comparison between the new and default combined formulation is carried out to determine the benefits of a Dm-based radar profiling algorithm methodology in the GPM combined algorithm.