| The Lightning Mapper Imager(LMI)deployed on the geostationary meteorological satellite Fengyun-4A is the first satellite-based lightning sensor in China,which is launched in December 2016.It can continuously monitor the lightning activity in China and the surrounding areas.However,the evaluation of LMI data quality and its application in severe convection monitoring and forecasting is very limited.Therefore,one purpose of our study is to quantitatively evaluate the LMI data quality by comparing with the ground-based and the other satellite-based lightning data.Additionally,the LMI data are assimilated into the Weather Research and Forecasting(WRF)Model to simulate deep convection,and the model performance are analyzed in detail.The main findings are as follows:(1)The LMI data quality is quantitatively evaluated by comparing with the other lightning datasets.The results show the LMI lightning data can reasonably reveal the decreasing trend of lightning frequency from southeast to northwest in China and the lightning radiative differences between land and ocean.It is able to detect the high lightning occurrence in South China and Southwest China,while it has poor detecting performance in regions with high lightning occurrence in East China.LMI is consistent with the other datasets in monthly variation of lightning activity.But the detecting performance of LMI is poor in the daytime due to the stronger of solar reflection.(2)The evaluation algorithm for the lightning data,which was developed based on the Bayesian theorem,is improved in our study by re-selecting the spatio-temporal matching windows and unifying the resolutions of LMI and ground-based dataset.The improved algorithm is more suitable for the evaluation of LMI detection efficiency,and has a smaller bias.The effect of diurnal variation on LMI detection efficiency in South China Sea was quantitatively analyzed by using the improved algorithm.It shows the average detection efficiency of LMI in the daytime and nighttime are 12.22% and 43.23%,respectively,indicating a diurnal difference of approximately 30%.Compared with other satellite-based lightning sensors,LMI is more sensitive to the diurnal variation.(3)A horizontal dispersion module for lightning data is developed.With our module,the LMI data can be applied in the currently available assimilation method,which was originally developed for ground-based lightning data.The 3-D graupel mixing ratio and accompanying latent heat releases retrieved by the LMI data at night were assimilated into WRF,and the discontinuity caused by the low-resolution LMI data was removed using the horizontal dispersion module.The forecast performances of LMI data assimilation in the mesoscale convective systems in the maritime and mountainous areas were analyzed.The model results show that the LMI lightning data assimilation promotes the generation and enhancement of updraft in lightning detection regions,and speeds up the start-up time of the convection.The assimilation also affects the cold-cloud precipitation microphysics,which advances the formation of cold pools and the associated gust-fronts.Most of the convective cells that are not seen in the control experiments are successfully simulated after assimilation,and the morphology,location and the vertical structure of the simulated convection are consistent with observations in the assimilation experiments.The forecasts of convective moving path and precipitation are also improved,and the most noticeable improvement is seen in the nowcast period.(4)Due to the poor performance of LMI in the daytime,a fusion algorithm is proposed to retrieve the flash rate,which improves the convection forecast effectively in the daytime.Firstly,the statistical relationships between LMI flash rate and cloud top temperature in coastal,oceanic and inland regions are established and the flash rate is retrieved preliminarily.Then,the correction coefficients are obtained based on the LMI and ground-based lightning data by using cumulative probability density matching method.Finally,the correction coefficients are applied to the primarily retrieval,and the flash rate in the daytime can be calculated.The spatial distribution and frequency of lightning in the daytime can be reasonably retrieved by using this algorithm.With the retrieved flash rate assimilated into WRF,the forecasts of the morphology,locations and moving path of convective systems are greatly improved. |
