Evaluation of three-dimensional variational assimilation on reducing the sensitivity of Tropical Cyclone GONU simulation to selective domains

The dependence of numerical models on the selected domain, in turn, affects the accuracy of prediction and simulation of Tropical Cyclones (TCs) and is considered as a very serious challenge. In the first part of this study, the WRF model was used to determine the sensitivity of the track and intensity of TC GONU to a selective domain. In the second part, the performance of assimilation 3D in order to reduce the dependence of the sensitivity of the TC Gonu simulation to the selective domains was evaluated. The Gonu was the strongest TC occurred over the Arabian Sea. The peak intensity of TC Gonu was estimated 140 knots and 130 knots by Joint Typhoon Warning Center (JTWC) and India Meteorology Department (IMD), respectively. Four domains were separately selected. All of the simulations in this study were initialized at 00 UTC of 2 June for six days. In all simulations, authors used the data from NCEP global final analysis (FNL) on a 1.0°×1.0° grid to provide initial and boundary conditions. Despite the little difference in selective domains, the results in the first section showed the simulated tracks differed compared with each other, considerably. For performing simulations in the second part, the QSCAT, BUOY, METAR, SHIPS, SONDE, and SYNOP data to number 2064, 30, 63, 18, 37, and 208, were used, respectively. The results in the second part showed that assimilation of the satellite and synoptic data at the time of the start of the model, lead to improving quality of the first guess data. Therefore, the accuracy of the simulated tracks in all selected domains was enhanced and reduced the sensitivity of TC Gonu simulation to the selected domain. Regardless of the great difference in simulated tracks, especially in the case of no use of assimilation, which in turn influences the intensification of the TC, in all of the simulations, the simulated intensity during the intensity peak of the TC is higher compared with the IMD reference data and is less compared with JTWC reference data. Since during the simulations, the sea-surface temperature  has been used constantly and on the other hand, the exact values of sea-surface temperature have a significant impact on the intensity of the TC simulation, the WRF model coupled with an ocean model for accurate determination of sea-surface temperature during simulation can improve the accuracy of the results of this study. There is, of course, another way to improve the quality of the results, when results depend on the selective domains. For every domain, one simulation is performed and the average of the simulations is considered (ensemble forecast). The high amount of time spent in this method is considered as serious trouble. It should be noted that in regional models, the sensitivity of simulations to the selected domains is also highly dependent on the boundary conditions, which should be considered.