The Study And Improvements On Characteristics Of Initial Perturbation Growth Rate In The Regional Ensemble Prediction System Of GRAPES

The initial and model errors lead to large amounts of uncertainty in deterministic numerical forecast models,However,Ensemble forecast method is designed to explicitly address the uncertainty,The key of which is how to describe all sources and distribution characteristics of initial errors,and reflect the evolution of initial errors with forecast lead time in a logical and correct way,So,The reasonal characteristics of initial perturbation growth rate are vitally important to ensemble forecast system.Nowadays,We have found the GRAPES-REPS(Global and Regional Assimilation and Prediction Enhanced System-Regional Ensemble Prediction System)model based on ETKF(Ensemble Transform Kalman Filter)initial perturbation method,However,The analysis and study on the characteristics of initial perturbation growth rate and perturbation structures are really little,and yet there hasn't conducted an improvement on the deficiency related to ETKF method,so as to form more reasonable initial perturbation structures and increasing rates.In order to systematically evaluate ETKF method and improve the description ability of GRAPES Regional Ensemble Prediction System(GRAPES-REPS)to the forecast uncertainty,Firstly,By analyzing ETKF characteristics of initial perturbation components,variance accuracy,kinetic energy spectrum,perturbation energy evolution,daily variation,ensemble spread and root mean square error,C haracteristics of initial perturbation structure and its growth rate in the regional ensemble prediction system of GRAPES-REPS are revealed in the present study.In addition,Based on the deficiency existed in ETKF method,We constructed a “double-components ETKF initial perturbation method” to improve ensemble forecast quality in lower layers;Secondly,This study further demonstrates the effect of model bias on the quality assessment of an ensemble prediction system(EPS),which provides ground for constructing reasonable quality assessment methods of an EPS and perturbation methods;Lastly,To solve the severe model bias in GRAPES-REPS,We developed a 3D bias damping method,trying to reduce the systematic errors and improve the ensemble forecast quality.Get the following main conclusions:(1)The perturbation field derived from the ETKF initial perturbation sche mes is mainly large-scale with a flow dependent structure.In addition,the ensemble perturbation can effectively capture the structure of forecast error;The total energy of initial perturbation and ensemble spread can keep appropriate growth rate at all forecast lead times.However,The growth rate of elements' ensemble spread and RMSE in lower layers is smaller than those in higer layers;The increase in horizontal resolution can increase the large-scale perturbation spectrum energy at the middle and high levels,and contribute to the improvement of ensemble forecast results at isobaric levels and near the surface,including the forecasts of temperature and wind.(2)The double components ETKF initial perturbation method designed in this paper can increase absolute analysis perturbation in lower layers,initial perturbation energy and ensemble spread.In addition,The research demonstrates that the perturbation total energy and ensemble spread represent the increase of forecast errors.(3)it is necessary to remove forecast biases before one can accurately evaluate an EPS.Only when an EPS has no or little forecast bias,can ensemble verification metrics reliably reveal the true quality of an EPS without removing forecast bias first.(4)A 3D bias damping method can reduce the systematic errors in model and improve the quality of ensemble forecast systems to a great extent.The spread-skill relation is improved,although the improvements in spread's spatial structure and spread's magnitude are much less,Furthermore,The probabilities become much sharper and almost perfectly reliable.