Numerical Forecast Correction of Temperature and Wind Using a Single-Station Single-Time Spatial LightGBM Method

Achieving high-performance numerical weather prediction (NWP) is important for people’s livelihoods and for socioeconomic development. However, NWP is obtained by solving differential equations with globally observed data without capturing enough local and spatial information at the observed station. To improve the forecasting performance, we propose a novel spatial lightGBM (Light Gradient Boosting Machine) model to correct the numerical forecast results at each observation station. By capturing the local spatial information of stations and using a single-station single-time strategy, the proposed method can incorporate the observed data and model data to achieve high-performance correction of medium-range predictions. Experimental results for temperature and wind prediction in Hainan Province show that the proposed correction method performs well compared with the ECWMF model and outperforms other competing methods.

Research Progress on Precipitation Accuracy Verification and Statistical Post-Processing in Ensemble Numerical Forecast System

The temporal and spatial accuracy of precipitation of ensemble numerical forecast systems is an important factor that affects the level of meteorological and hydrological coupled forecasting. This article focuses on the current research of verification of precipitation accuracy and statistical post-processing. The verification of forecast precipitation accuracy mainly focuses on the probabilistic characteristics such deterministic accuracy, the resolution, the forecasting skills and the degree of dispersion. Some mainstream statistical post-processing methods have strong performance of spatial downscaling and error correction, but they commonly have the defect of destroying the temporal and spatial dependent structure of precipitation. A comprehensive statistical post-processing method integrated the three functions is the development direction in the future. At the same time, statistical post-processing methods to improve the certainty and probabilistic accuracy of forecast precipitation need to be systematically identified. Its impact on the spatio-temporal dependence structure also needs to be improved.

System of Multigrid NLS-4DVar Data Assimilation for Numerical Weather Prediction (SNAP)：System Formulation and Preliminary Evaluation

<p class="a"><span lang="EN-US">The system of multigrid NLS-4DVar data assimilation for Numerical Weather Prediction (SNAP) is established, building upon the multigrid NLS-4DVar assimilation scheme, the operational Gridpoint Statistical Interpolation (GSI)-based data-processing and observation operator and widely used numerical forecast model WRF (easily replaced by others global/regional model). The multigrid assimilation framework can adequately correct errors from large to small scales to achieve higher assimilation accuracy. Meanwhile, the multigrid strategy can accelerate iteration solution improving the computational efficiency. NLS-4DVar, as an advanced 4DEnVar method, employs the Gauss-Newton iterative method to handle the nonlinear of the 4DVar cost function and provides the flow-dependent background error covariance, which both contribute to the assimilation accuracy. The efficient local correlation matrix decomposition approach and its application in the fast localization scheme of NLS-4DVar and obviating the need of the tangent linear and adjoint model further improve the computational efficiency. The numerical forecast model of SNAP is any optional global/regional model, which makes the application of SNAP very flexible. The analysis variables of SNAP are rather the model state variables than the control variables adopted in the usual 4DVar system. The data-processing and observation operator modules are used from the National Centers for Environmental Prediction (NCEP) operational GSI analysis system, prominent in the various observation operators and the ability to assimilate multi-source observations. Currently, we have achieved the assimilation of conventional observations and we will continue to improve the assimilation of radar and satellite observations in the future. The performance of SNAP was investigated assimilating conventional observations used for the generation of the operational global atmospheric reanalysis product (CRA-40) by the National Meteorological Information Center of China Meteorological Administration. Cyclic assimilation experiments with two windows, which is 6-h for each window, are designed. The results of numerical experiments show that SNAP can absorb observations, improve initial field, and then improve precipitation forecast. </span></p>