The Study Of Microphysics Scheme Parameter Perturbation In Storm-scale Ensemble Forecast

Microphysical parameterization scheme greatly influences grid-scale deep-convection simulation in high-resolution numerical models,cloud and precipitation forecasts are very sensitive to small changes of parameters in the schemes.Storm-scale ensemble forecast is an effective method to estimate the uncertainty of numerical forecast.In this paper,based on Weather Research and Forecasting(WRF)Version 4.0 model,the parameters(intercept parameter and density)which strongly influence precipitation particles' size distribution in single-moment microphysical scheme are studied in “7.20” North China torrential rain process.Firstly,sensitive experiments of each single parameter are performed to analyze its influences on atmospheric hydrometeor mixing ratios and surface rainfall.Next,multi-parameter method is adapted to conduct grouped-parameter perturbation ensemble forecast experiments in the reference of sensitivity analysis.Groups' forecast effect on atmospheric variables and precipitation are compared and perturbation's spatio-temporal influences on forecast are explored in chapters.The main conclusions are as follows:(1)Atmospheric hydrometeor contents has good responses to parameter changes.The hydrometeor vertical profiles show that precipitation particle' content has positive correlation with its intercept value,graupel content has negative correlation with its density,decreasing graupel intercept increase snow content and increasing graupel density helps increase cloud water.Varying raindrop intercept helps solve the problem of storm clouds moving out of the area early in late forecast time thus increasing hydrometeor contents in middle and low levels.Raindrop intercept performs best in improving rainfall forecast among parameters.(2)Four-parameter group improves ensemble Spread skill of 500?925h Pa meridional wind and near-surface temperature a lot and has positive skill on rainfall forecast especially in decreasing torrential rainfall unnecessary forecast probability.Multi-intercept scheme has better forecast skill on middle-low level water vapor mixing ratio when graupel-parameter scheme increasing ensemble Spread.Groups have different aspects in improving rainfall forecast,single-intercept perturbation ranks second to four-parameter scheme in torrential rainfall forecast when double-intercept scheme performs worst.(3)Negative perturbation of raindrop/snow intercept are more sensitive to forecast effects than positive perturbation,as parameter categories increase,the overall impact of group perturbation on forecast continues to increase,in torrential rain region,double-intercept group's perturbation energy is weaker than single-parameter scheme's energy.Group-perturbation sensitivity has a good consistency with its Spread-RMSE skill on atmospheric variables and are reflected in its precipitation forecast on a certain degree.