Impact of Assimilation of Radar Data on the Simulation of Squall Line Event

Accurate prediction of squall lines that accompany thunderstorms is a challenging task. A squall line was recorded in Dayyer port station over the southwest of Iran in Bushehr province, on 19 March 2017. In this study, the properties associated with the mentioned squall line including the time of formation, growth and destruction of the convective cells in terms of intensity and vertical growth, as well as the associated precipitation and other meteorological features are simulated using the WRF-ARW model with 3-dimensional variational (3DVAR) assimilation and control experiment (CTRL) for 18UTC 18 March 2017 with two domains of 27 and 9-km horizontal resolution. Radial winds and reflectivity of Bushehr Doppler Weather Radar (DWR) along with surface and upper-level observational synoptic data are used to simulate the above mentioned squall line event with the aim of updating initial and boundary conditions through 3DVAR assimilation in WRF model. In order to verify the simulated properties associated with the squall line event, the horizontal wind speed, mean sea-level pressure, surface temperature and surface relative humidity, as well as time series of reflectivity and vertical growth in the squall line location on Bushehr DWR were compared with the corresponding observational data. To assess the performance of accumulated precipitation forecasts, the fraction skill score (FSS) curves are plotted for different rainfall thresholds 0.5, 5, 10 and 15 mm/day. In general, the results showed that the radar data assimilation has a significant effect on the simulation of the characteristics accompanied with the squall line event such that without data assimilation, the WRF model is not capable of simulating the squall line thoroughly. The results of 3DVAR simulation are also much closer to the observational data in predicting the features along the squall line. The absolute value of the mean errors in simulations with assimilation for surface horizontal wind speed, surface temperature, mean sea level pressure, and surface relative humidity were decreased by 27%, 7%, 18%, and 10%, respectively, compared to those without assimilation. The formation time and growth of convective cells, their horizontal distributions, vertical structure, and their destruction time in 3DVAR simulation are closer to the verifying observational data. The 3DVAR simulation also achieved significant success in predicting 6-h accumulated precipitation with a threshold value of above 10 mm. Also, the value of error in 3DVAR simulation in 6-h accumulated precipitation at synoptic stations in Bushehr province was decreased by 33%, compared to those without assimilation.