The bending angle received from low earth orbit satellites (LEO) can be used to derive the refractivity of the atmosphere. However, those products are retrieved under the assumption that the atmosphere is spherically symmetric. In general, this assumption has proved to be verified for global model with moderate resolution (>120km), and assimilating Abelian retrievals as “local” refractivity profiles measurements has seemed to be partly successful. This approach, however, is likely to fail with mesoscale models, where the existence of humidity gradients on short distances can severely affects the atmospheric refractivity field homogeneity.
In this presentation, we report a study on the validity of the spherical symmetry assumption at mesoscale. 3-dimensional refractivity fields are generated with a high-resolution version of the 5th generation of the Penn State/NCAR Mesoscale Model (MM5). Simulated observed refractivity profiles are created by Abelian inversion of bending angle distributions, results of the applications of the ray-tracing technique on the model 3-dimensional refractivity fields in realistic COSMIC satellite orbital configuration. The difference between model retrieved refractivity profiles and model local profiles is quantified. The impact of the spherical symmetry assumption at mesoscale is subsequently assessed through series of 3-dimensional variational assimilation of simulated observations. Control 3-dimensional refractivity fields are generated using MM5 high-resolution simulation of Cyclone Danny in which conventional and SSM/I observations are assimilated. Simulated COSMIC Abelian refractivity profiles are retrieved from the control runs and assimilated into a degraded version of MM5, in which SSM/I observations were not assimilated. The experiment is repeated, but this time model local and not Abelian refractivity profiles are assimilated. Differences in the assimilation results are solely due to the spherical symmetry assumption and are quantified.
Supplementary URL: