Abstract:How to accurately address model uncertainties with consideration of the rapid nonlinear error growth characteristics in a convection-allowing system is a crucial issue for performing convection-scale ensemble forecasts. In this study, a new nonlinear model perturbation technique for convective-scale ensemble forecasts is developed to consider a nonlinear representation of model errors in the Global and Regional Assimilation and Prediction Enhanced System(GRAPES)Convection-Allowing Ensemble Prediction System(CAEPS). The nonlinear forcing singular vector(NFSV) approach, that is, conditional nonlinear optimal perturbation-forcing(CNOP-F), is applied in this study, to construct a nonlinear model perturbation method for GRAPES-CAEPS. Three experiments are performed: One of them is the CTL experiment, without adding any model perturbation; the other two are NFSV-perturbed experiments, which are perturbed by NFSV with two different groups of constraint radii to test the sensitivity of the perturbation magnitude constraint. Verification results show that the NFSV-perturbed experiments achieve an overall improvement and produce more skillful forecasts compared to the CTL experiment, which indicates that the nonlinear NFSV-perturbed method can be used as an effective model perturbation method for convection-scale ensemble forecasts. Additionally, the NFSV-L experiment with large perturbation constraints generally performs better than the NFSV-S experiment with small perturbation constraints in the verification for upper-air and surface weather variables. But for precipitation verification, the NFSV-S experiment performs better in forecasts for light precipitation, and the NFSV-L experiment performs better in forecasts for heavier precipitation, indicating that for different precipitation events, the perturbation magnitude constraint must be carefully selected. All the findings above lay a foundation for the design of nonlinear model perturbation methods for future CAEPSs.

CALCULATION OF SINGULAR VECTORS BY USING THE TANGENT AND ADJOINT TECHNIQUE OF GRAPES

Research on GRAPES Singular Vectors and Application to Heavy Rain Ensemble Prediction

Application and Characteristic Analysis of the Moist Singular Vector in GRAPES-GEPS

An Initial Perturbation Method for the Multiscale Singular Vector in Global Ensemble Prediction

Enhanced Grasshopper Optimization Algorithm For Modeling Multivariable Systems With Svr

Study on Optimization of Atmospheric Correction Process Based on 6S Radiative Transfer Model

Development of Moist Singular Vectors in GRAPES-GEPS and a Preliminary Evaluation

Improving Seasonal Forecast Skill of North American Surface Air Temperature in Fall Using A Postprocessing Method

Retrieval of forest growing stock volume by two different methods using Landsat TM images

Numerical Studies on Forecast Error Correction of GRAPES Model with Variational Approach

A Nonlinear Representation of Model Uncertainty in a Convective-Scale Ensemble Prediction System

Stochastic perturbations in the semi-Lagrangian advection algorithm of the SL-AV global atmosphere model

Research of GRAPES Numerical Weather Prediction Model

Analysis of the Singular Vectors of the Full-Physics Florida State University Global Spectral Model

Verification of GRAPES Unified Global and Regional Numerical Weather Prediction Model Dynamic Core

THE USE OF SHEAR GRADIENT VORTICITY IN TROPICAL CYCLONE HEAVY PRECIPITATION PREDICTION: A HIGH-RESOLUTION NUMERICAL CASE STUDY

SVQ: Sparse Vector Quantization for Spatiotemporal Forecasting

Global ionospheric maps forecasting based on an adaptive autoregressive modeling of grid point VTEC values

Analysis of the singular vectors of the full-physics FSU global spectral model

The Impacts of Time-Step Size and Spatial Resolution on the Prediction Skill of the GRAPES-MESO Forecast System

A semi-implicit semi-Lagrangian global nonhydrostatic model and the polar discretization scheme