Assimilation Of Lightning Data Combined With Satellite And Radar Observations

With the frequent occurrence of meteorological disasters in China,improving the ability to predict High-Impact Weather(HIWeather)causing meteorological disasters is both an urgent need for disaster prevention and mitigation and one of the challenging international frontier research topics.With the extensive establishment of ground-based and space-based lightning observation networks,assimilation of lightning data to reduce the initial field errors is an effective way to improve the ability of numerical models to forecast HIWeather such as severe convective events.With this background,firstly,this thesis evaluates the assimilation effect of the pseudo-water vapor lightning assimilation scheme for two typical convective events,namely convective precipitation in arid areas and landfall typhoon “Mangkhut”,and establishes a sparse quality control scheme for satellite lightning observations to improve the assimilation lightning data for typhoon forecasting.Secondly,to address the problems in the pseudo-water vapor lightning assimilation scheme,a new method for improving the pseudo-water vapor lightning assimilation scheme applicable to convective processes in different climatic zones is proposed.The new method combined lightning observations with satellite cloud-top height information.Thirdly,a new idea of combining and assimilating lightning pseudo-water vapor with Geostationary Interferometric Infrared Sounder(GIIRS)inversion profiles and radar observations is explored.Fourthly,the advantages of the hybrid assimilation method of ensemble variational(3DEn VAR)with flowdependent background error covariance in the combination assimilation of lightning and radar observations are clarified.Finally,a new way of organically blending extrapolated forecasts with numerical prediction through data assimilation techniques is developed,and a regional filter pre-processing scheme to improve the assimilation effect of extrapolated reflectivity factor is established,further improving the forecasting capability of the numerical model for severe convective events.The main conclusions of the paper are as follows:(1)The lightning assimilation scheme based on pseudo-water vapor significantly improves the forecast capability of convective precipitation on leeward slopes in arid areas,and improves the track and precipitation forecast of typhoon “Mangkhut”.In case of convective precipitation on leeward slopes in arid regions,assimilating pseudo-water vapor observations driven by ground-based lightning data,a large increment of relative humidity at the location of the lightning occurrence was obtained in the analysis field.Positive and negative potential vorticity anomalies and corresponding upward and downward motions were found at the leeward slope of the forecast field.With the adjustment of the thermal and dynamic fields in the forecast field,the cloud-water particle content in the assimilation experiments increases significantly at the leeward slope.Lightning assimilation experiments significantly improve precipitation forecasting,and cyclic analysis experiments are better than single analysis experiments in terms of improvement.In the case of the landfall of the typhoon “Mangkhut”,assimilation of pseudo-water vapor observations driven by the Lightning Imager(LMI)lightning event data from the Fengyun-4A geostationary satellite(FY-4A)improves water vapor conditions at the lightning occurrence in the outer spiral cloud belt prior to typhoon landfall.The lightning assimilation experiments improved the forecast of the typhoon's moving track and did not improve the typhoon's intensity significantly.The assimilation experiments have improved the accumulated precipitation from 0-12 hours.A satellite lightning thinning quality control(QC)scheme has been developed to reduce errors introduced by excess lightning pseudo-water vapor in the analysis field.Compared with the original satellite dense lightning observations from satellite,the assimilation of thinning lightning data improves typhoon track forecasts more significantly,reducing typhoon track forecast errors and improving typhoon precipitation forecasts.(2)The proposed new scheme of pseudo-water vapor lightning assimilation by combining lightning data with the cloud top height of FY-4A can obtain more accurate pseudo-water vapor,which are tested by two convective events with different characteristics,and the new scheme can effectively improve the forecast skill of numerical models for strong convective precipitation.Lightning assimilation schemes based on isothermal layers of background field and within an assumed fixed depth of 3km above LCL obtain the smallest water vapor increments and the least improvement in precipitation.The lightning assimilation scheme between LCL and the fixed height(15 km)obtains the largest water vapor increments in the analysis field,with stronger false precipitation forecasts.Combining cloud top height with lightning data yields more accurate pseudo-water vapor observations,with more reasonable water vapor increments obtained in the analysis field.The new lightning assimilation scheme of lightning combined with cloud top height not only provides better forecasts of heavy precipitation that cannot be resolved in experiments with the adjustment of isothermal layers and fixed depth of LCL above,but also suppresses false precipitation forecasts in experiment with adjustments between LCL and fixed heights.The new scheme for lightning assimilation is able to avoid the effects of different convective intensities and is applicable to convective events in different climatic regions.(3)The FY-4A GIIRS temperature profile has a considerable amount of data and acceptable observation error.The combined assimilation of lightning data and GIIRS temperature profiles improves the water vapor conditions in the convective region of the analysis field and the thermal environment at the clear sky region,effectively improving the forecasts of heavy precipitation.The FY-4A GIIRS temperature profile is relatively stable in clear sky conditions between 700 h Pa and 100 h Pa with a root means square error(RMSE)of less than 2 K and close to 1 K.After multiple cycles of assimilating the temperature profile,the error between the analysis field and the observation is significantly reduced.In the 24-h accumulation precipitation forecast,the assimilation experiments suppress spurious precipitation that occurred in the control experiment,and had the best forecasting skill with a 15-minute assimilation window at the strong precipitation threshold.Combined assimilation of temperature profiles with pseudo-water vapor driven by FY-4A LMI and cloud top height under a 15-minute assimilation window effectively updated water vapor conditions at the location of lightning occurrence,reducing dry bias in the control experiment and cold bias in the clear sky region.The combined assimilation experiment improves the water vapor conditions in the convective region compared to the assimilation-only temperature profile experiment;the combined assimilation experiment not only updates the temperature environment of the analysis field compared to the assimilation-only lightning experiment,but also improves the wet bias of the analysis field.Combining assimilation of the lightning data and temperature profiles not only better forecasting skills at the beginning of the forecast due to the updating of water vapor conditions,but also the updating of the temperature field improves the forecast after 24 hours.(4)The advanced ensemble variational assimilation system combined assimilated lightning and radar observations to improve the water vapor,cloud-water information and dynamical conditions in the analysis field,obtain a thermodynamically balanced analysis field,reduce the model spin-up period and significantly improve the numerical model's skill to forecast severe convective precipitation.Based on a new scheme of lightning assimilation that combination of cloud top heights,using a dual-resolution hybrid 3DEn VAR assimilation method to combine FY-4A LMI lightning data and radar observations(radial velocity and reflectivity factor)in a single analysis and a highfrequency cyclic analysis,respectively.The results of the single analysis experiments show that the lightning assimilation experiments had the largest increment of water vapor,but was unable to update the hydrometeor information and dynamical conditions;the radar assimilation experiments had larger increments of hydrometeor variables and vertical velocity,but little update of water vapor;the lightning and radar combination assimilation had the water vapor properly updated and also produced larger increments of water condensate and vertical velocity.Overall,the forecasts from the high frequency cyclic analysis experiments were better than those from the single analysis experiments.The combined assimilation of lightning and radar observations effectively improves the cloud-water information and dynamical fields,and water vapor in the convective region of the analysis field,reduces the spin-up period,provides significant improvements in forecasting 6-h accumulation precipitation,and provides a better suppression of spurious convection present in the control experiment and lightning assimilation experiments.The hybrid 3DEn VAR method with flow-dependent background error covariance is not only able to influence the assimilation variables,but through the hybrid covariance is able to diffuse increments of the assimilation variables to other control variables,whether assimilation of lightning and radar data alone or combined,the hybrid 3DEn VAR method has advantage over the three-dimensional variational(3DVAR)method.(5)The extrapolated reflectivity factor assimilation scheme based on pseudo-water vapor improves the water vapor conditions in the convective region of the analysis field.The regional filtering of the extrapolated reflectivity factor can filter out the larger extrapolation errors and retain relatively reliable convective information,which further improves the assimilation effect of the extrapolated reflectivity factor and provides a new idea for blending the extrapolated forecast with NWP.In the 3DVAR assimilation framework.Assimilation experiments improve the water vapor environment in the convective region of the analysis field,and improve the forecast of reflectivity factor and accumulation precipitation.Assimilation of pseudo-water vapor driven by extrapolated reflectivity factor is more effective in improving convective systems that move slowly;for fast moving and evolving convective systems,large extrapolation errors lead to large wet deviations in the forecast of reflectivity factor and precipitation.A new method of regional filtering of the extrapolated reflectivity factor is proposed,which can filters out the large extrapolation error of the extrapolated reflectivity factor to retain relatively stable and observationally consistent convective information.Compared to the assimilated unfiltered extrapolated reflectivity factor,the assimilated filtered extrapolated reflectivity factor effectively reduces the wet bias in the analysis field and prevents spurious reflectivity factor and precipitation forecasts in the forecast field.In addition,the extrapolated reflectivity factor from the 15-minute interval cycle assimilated to 60 minutes,while having a high skill score,but faces greater problems of wet bias and false forecasts if it is not filtered.Therefore,in practice,the 30-minute extrapolation reflectivity factor has a smaller extrapolation error and the 15-minute interval cycle assimilating the 30-minute extrapolation reflectivity factor may improve the forecast less,but introduces less error and more reliable results.