Toward more direct evaluation of cloud resolving models with polarimetric, Doppler radar observations

Poster PDF

Description

The Tropical Warm Pool– International Cloud Experiment (TWP-ICE) provides a rich dataset for examining tropical deep convection, as well as an opportunity to compare quality microphysical and kinematic observations with cloud resolving model simulations. While numerous studies have already simulated and studied TWP-ICE cases, evaluation of model microphysics schemes with radar-derived kinematic and microphysical fields remains challenging due to observational errors and a mismatch between simulation parameters and observational quantities. To this end, we describe a new framework for comparing model simulations and radar bulk microphysics, as well as examine different methodologies for improving radar-derived kinematics and microphysics. Accurate vertical wind observations are critical for understanding strengths and weaknesses in CRM simulations. To that end, we compare two dual-Doppler methodologies for the 22-23 January 2006 monsoon case. The first is the multi-pass variational wind retrieval outlined in Collis et al. (2013), whereupon a cost function involving radial velocity, noise, and the anelastic continuity equation, is minimized. The second is direct integration of the anelastic mass continuity equation. Both methods are compared to profiler-derived vertical motion using a dual-frequency (920-MHz and 50 MHz) retrieval technique (Williams, 2012). The polarimetric CPOL radar allows for characterization of bulk microphysics occurring within convection. One shortcoming of bulk hydrometeor identification schemes, however, is that they only identify the most dominant category occurring within a radar pulse volume (or grid point), despite the likelihood of mixtures occurring in both nature and model grid points. We present a new technique that examines areas of possible mixtures by categorizing the 2nd and 3rd highest scoring hydrometeor types. Additionally, we assign a confidence to the classifications, providing new depth and insight to the traditional fuzzy logic hydrometeor identification toward the goal of more direct comparisons with cloud resolving models. In order to facilitate more direct comparisons between the cloud resolving models and radar observations, a new framework is described. The POLArimetric Radar Retrieval and Instrument Simulator (POLARRIS) takes CRM output and simulates polarimetric, Doppler radar observations using scattering matrices. These scattering matrices use model-consistent assumptions about size distributions, bulk density, and particle phases coupled with assumptions about axis ratios, particle orientations and oscillations in order to simulate polarimetric radar variables. Synthetic model polarimetric fields are then run through the same classification framework (e.g. HID) as the radar observations, allowing for more direct evaluation of cloud resolving model output.