The application of reduced-rank retrospective optimal interpolation and ensemble retrospective optimal interpolation to WRF

Abstract:

Retrospective optimal interpolation (ROI) algorithm assimilates observations over the analysis window for variance-minimum estimate of an atmospheric state at the initial time of the analysis window, suggested in Song et al. (2009). The assimilation window of ROI algorithm is gradually increased, which is similar with that of the quasi-static variational assimilation (QSVA). Reduced-rank retrospective optimal interpolation (reduced-rank ROI) was developed with an improvement by incorporating eigen-decomposition and covariance inflation (Song and Lim, 2011). In addition, ensemble retrospective optimal interpolation (EnROI) is derived using the Monte Carlo method form ROI, developed by Song et al. (2011). In this study, both of reduced-rank ROI and EnROI are applied to Weather Research and Forecasting model (WRF) for validating the algorithm and investigating the capability of data assimilation in the framework of 3-dimensional model. One-profile assimilation experiment and Observing System Simulation Experiment (OSSE) have been performed. A total energy norm is used for the normalization. The characteristics and strength/weakness of ROI algorithm are identified with comparing the results from the experiment with other data assimilation method.