Assimilation Of Cloud-to-ground Lighting Data Using 3DVar And SVD-EN3DVar Methods

Because of the characteristics of strong suddenness and destructivity, the study of severe convection weather has an important significance, and is one of the most difficult problems in the weather prediction. As an indicator of convective activity, lightning data have several unique merits and can be used as a supplement of other observational data in tracking and forecasting of strong convective storms. Some studies show that assimilation of lightning data can adjust the initial field of forecast model and improve the quality of weather forecasting. In order to make a further acquaintance about the application of lightning data in numerical simulation, a severe convection weather that happened on June 5, 2009 in Anhui Province is chosen to have a discussion.In the first place, a two-step variational method was used to retrive the radial velocity field based on Hefei Doppler radar data. The derived winds, reflectivity and conventional observation parameters are used to analyse the development of this convection activity and examine the early warning condition of the catastrophic weather such as hail, strong winds and lightning happening.In the next place, the lightning data, which is abtained from Lighting location system in Anhui Province, were converted into three-dimensional lightning-proxy radar reflectivity using an assumed simple relationship between flash density and radar reflectivity included in the Gridpoint Statistical Interpolation system. The result shows that the lightning-proxy radar reflectivity was almost same as radar observation in showing the site, size and shape. Thus, the lightning data can be used as a supplement of radar reflectivity in research on strong convection weather forecasting.Fianlly, the experiments of cloud-to-ground lightning flashes data assimilation using WRF-3DVar system and SVD-En3 DVar method were made. The results indicated that:(1)Assimilation with WRF-3DVar, the lightning data has supplemented the hydrometeors and adjusted the dynamic and thermal terms of the initial field. The assimilation of cloud-to-ground data can improve the forcasting maximum reflectivity closer to the radar observation and enhance the ability of reflectivity prediction over a five-hour period. Besides, assimilation can improve the forcast accuracy of rainfall location and provide some references for the short-range precipitation prediction. But the main problem is that the rainfall amount is systematically overestimated.(2) Assimilating lightning data by SVD-En3 DVar method through the assimilation cycle not only improved the contiguous two hours reflectivity prediction forecast, but also advance the rainfall predication quality.