Uncertainties Assessment of Synthetic Radar Simulations

Forward radar operators allow to simulate synthetic radar observations 
from the state of the atmosphere simulated by a numerical weather 
prediction (NWP) model and hence offer great opportunities for 
validation of or assimilation in NWP models. Several radar operators 
have been proposed in the literature with various degrees of 
sophistication, but their evaluation remains often quite rudimentary. 
The simulated radar fields are expected to depend strongly on the 
assumptions about the geometrical and physical properties of the 
considered hydrometeors (e.g. aspect-ratio, canting angle distributions, 
permittivity, etc) but the influence of these assumptions on the 
variability of the resulting simulated radar variables is often poorly 
known, especially in the solid phase where the estimation of radar 
variables is the most difficult. This uncertainty about the error 
sources of the operator makes it difficult to relate discrepancies 
between real and simulated radar observations to actual limitations of 
the NWP model. In this work we present a new state of the art 
polarimetric radar operator and use it to perform a multi-factor 
sensitivity analysis of the simulated radar observables. The dependence 
of the radar operator outputs on its main assumptions is tested for 
several events corresponding to a variety of synoptic conditions in both 
liquid and solid precipitation. This study contributes to characterize 
the uncertainty in the simulated radar observables, which can be useful 
both for error propagation in assimilation and for model evaluation.