Typhoon Simulation Evaluation And Cumulus Convection Parameterization Optimization Of MPAS Model

In numerical typhoon simulations,comparing with the regional numerical model,the high-resolution global numerical models can better capture large-scale circulation and avoid lateral boundary effects,which is a strong support for fine-scale typhoon dynamics research.The MPAS global model based on SCVT grid not only avoids the abrupt change of physical quantity at the boundary of a regional model,but also increases the density of grid in the interested simulation area.At the same time,it also saves the computational resources.It is therefore a potential numerical model in typhoon research.For the application of MPAS model in typhoon simulation over the Northwest Pacific Ocean,firstly,the simulation performance of the MPAS model for the atmospheric environment field was evaluated with reference to the GFS model.Based on this,the effects of the physical parameterization scheme derived from the regional numerical model WRF on typhoon simulations are studied.The influence mechanism of cumulus convection parameterization scheme on typhoon path and intensity,and the influence mechanism of cloud microphysical parameterization schemes and grid resolution on the variability of typhoon precipitation simulation effects are analyzed.The optimization method of cumulus convection parameterization scheme based on genetic algorithm in "grey zone" is further explored and the improvement of the optimized cumulus convection parameterization scheme on the simulation effect of typhoon precipitation is verified.Specific research contents and conclusions are as follows:Using stationary satellite radiation brightness temperature observations to evaluate the humidity field simulations from numerical weather model,the results show that using ERA Interim data and RTTOV mode for AGRI full disk clear sky observation simulation evaluation,the radiation values of AGRI observations and RTM model simulation results are in good agreement,after the quality control process the AGRI observations can be used as the reference data for model evaluation.The simulation performance of MPAS model and GFS model for the clear sky humidity field in the selected typhoon Hato simulation area was evaluated.The humidity field of MPAS model was closer to the actual value,and the humidity field deviation was correlated with the typhoon forecast intensity error to a certain extent.MPAS model has a reliable effect on the simulation of typhoon humidity field,and it has a good application value for the simulation of the Northwest Pacific typhoon.The influence of cumulus convection parameterization scheme on the simulation of typhoon path and intensity in MPAS model is shown as follows: The simulation effect of typhoon path and intensity is sensitive to the selection of cumulus convection parameterization.The average error of the New-Tiedtke scheme is less than that of the Kain-Fritsch scheme.The simulated intensity of Kain-Fritsch scheme is closer to the measured intensity when the typhoon is landing.Although the Kain-Fritsch scheme accounted for a large proportion of the best cases in the simulation of typhoon intensity,there was little difference in the intensity error between the two schemes.Considering the simulation effect of path and strength,the New-Tiedtke scheme is preferred.Take typhoon Nesat,which has a larger path and strength simulation error as an example.By analyzing the simulated circulation field,heating rate and PV trend,it is clear that the form of the subtropical high is the decisive factor affecting the path of the typhoon.The analysis of heating rate and PV trend helps to explain the potential mechanism of path deflection.The analysis of heating rate and PV trend helps to explain the potential mechanism of path deflection.The vertical heating structure reveals that the difference in the height of convective instability simulated by different schemes leads to the difference in the latent heat transport and release,which leads to the difference in the typhoon intensity simulated by different schemes.The study on the influence of cloud microphysics scheme and grid resolution on typhoon precipitation simulations in MPAS model has drawn the following conclusions:Differences in the related thermal and dynamic values of various cloud microphysical parameterization models affect the simulation results of precipitation.Under the same physical parameterization scheme,the precipitation intensity and spatial distribution of simulations with different grid resolution are significantly different.The simulation effect of the high-resolution grid was the best.Under the same grid resolution,the precipitation intensity and spatial distribution simulated by the two schemes were not significantly different,but the precipitation intensity simulated by the Thompson scheme was weak.The TRMM hydrometeor profile was used as a reference to analyze the physical mechanism of precipitation differences.It was found that the difference in the content of solid particles in the simulated results contributed greatly to precipitation differences,while the difference in the content of solid particles in the convective clouds under different physical parameterization schemes was the main cause of precipitation differences.Cumulus convection parameterization scheme was optimized by using micro genetic algorithm to improve the precipitation simulation and the studies show that coupling micro-GA with KF scheme based on WRF model can effectively optimal KF parameter value and improve the applicability of cumulus convection parameterization in the "gray zone" of high-resolution models,so as to improve the accuracy of the ETS score and spatial distribution of precipitation forecast.The physical parameterization scheme of MPAS model is mainly derived from WRF model.The optimized KF scheme is placed into MPAS model,and the same configuration scheme as WRF model is used to simulate typhoon precipitation.The results show that the simulation effect of MPAS model is improved compared with that of the original KF scheme.Micro-GA algorithms can be used as an effective method for optimizing physical parameterization schemes in numerical models.