Evaluating the Forecast Impact of Assimilating ATOVS Radiance With the Regional System of Multigrid NLS‐4DVar Data Assimilation for Numerical Weather Prediction (SNAP)

• Main Authors: Hongqin Zhang, Xiangjun Tian
• Format: Article
• Language: English
• Published: American Geophysical Union (AGU) 2021-07-01
• Series: Journal of Advances in Modeling Earth Systems
• Subjects: data assimilation; numerical weather prediction; multigrid; NLS‐4DVar; satellite radiation; precipitation
• Online Access: https://doi.org/10.1029/2020MS002407

Abstract The regional System of Multigrid Nonlinear Least Squares Four‐dimensional Variational Data Assimilation for Numerical Weather Prediction (SNAP) was recently established based on the multigrid NLS‐4DVar assimilation scheme, Weather Research and Forecasting numerical model, and Gridpoint Statistical Interpolation (GSI)‐based observation quality control and observation operator modules. The analysis variables are model state variables, rather than the control variables adopted in the conventional 4DVar system. The regional SNAP adopts the multigrid NLS‐4DVar, which can correct errors from large to small scales and accelerate iteration solutions, to minimize the cost function and obtain the optimal analysis. Therefore, the regional SNAP has a higher assimilation efficiency and accuracy. In addition, the assimilation performance of the regional SNAP for conventional and radar observations had been evaluated. The main goal of this study is to achieve the direct assimilation of satellite radiation data using the regional SNAP. In this study, 1‐week cycle assimilation experiments assimilating Advanced TIROS Operational Vertical Sounder (ATOVS) data were designed to fully evaluate the performance of regional SNAP compared with the GSI Ensemble Four‐dimensional Variational (4DEnVar) scheme. First, a rainstorm was selected to illustrate the performance of regional SNAP. The cumulative precipitation distribution of SNAP was closer to reality and the higher equitable threat score and lower far score indicate that the regional SNAP improves precipitation forecast. In the 1‐week numerical experiments, for the u/v wind and temperature variables, the regional SNAP outperforms GSI and there was an improvement in GSI for the humidity variable.