Improving Mesoscale Land-Atmosphere Coupled Covariance through Physics Parameter Perturbations on Dryline Convection Initiation Cases

Inadequate representation of the uncertainty in the parameterization of land-atmosphere coupling processes is an impediment to effective strongly coupled ensemble-based data assimilation on a range of space and time scales. In particular, there is a need to improve the ensemble representation of mesoscale coupled background error covariance structures related to physics uncertainty in convection-allowing data assimilation systems such as RRFS. Such improvements are intended to help optimize the strongly coupled data assimilation of screen level and soil moisture observations, which are expected to be particularly beneficial for Great Plains dryline convection initiation (CI) cases. These cases are important because the evolution of the dryline and associated CI can be strongly influenced by the coupling processes between the land and atmosphere, especially on cases with weak synoptic scale forcing. Dryline CI cases will therefore be used to conduct a series of deterministic and ensemble forecast experiments aimed at improving the representation of the ensemble-based coupled covariance through land surface model, surface layer, and planetary boundary layer parameter perturbations. Results will emphasize understanding the physical interpretation of the land-atmosphere coupled covariance structures related to model error, and suggestions for the optimal use of physics parameter perturbations within MPAS to represent these covariance structures will be presented. Initial steps towards implementing the strongly coupled data assimilation in a RRFS-like JEDI-MPAS workflow will also be discussed.