Numerical Simulation And Data Assimilation Research Of Two Kinds Of Typical Disaster Weather Events Based On WRF Model

Every year, meso-and small-scale disaster weather events cause large damages on human’s lives and property all around the word. So far, however, the accurate numerical weather forecast of the disaster weather events remains a difficult and frontier issue. Further understanding of the mechanisms of the formation, evolution and damages of disaster weather events is essential for improving the forecasting techniques of them. With the rapid development of high-performance computers, the application of high-resolution numerical models plays more and more important roles in atmospheric research, especially in studying the mechanisms of the formation and evolution of disaster weather events. In this study, two typical disaster weather events, which happened in America in different seasons, are simulated and analyzed using high-resolution Weather Research and Forecasting (WRF) model developed by National Centers for Atmospheric Research (NCAR). They are a wintertime persistent inversion that happened over Salt Lake City which is the capital of the Utah States and a convective process including several convective initiation events and subsequent storms evolution over Oklahoma and Northern Texas in late spring.In wintertime, persistent inversions are widely distributed weather phenomena in mid-latitude regions. The relatively stable and cold layer result in extreme temperature minimum, fogs and accumulation of air pollutants near the surface, which can cause serious damages on industry, agriculture, traffic and human’s health. So far, however, accurate forecast of the formation, evolution and dissipation of persistent inversions still remains a challenge. Most of the previous studies focused on the formation or dissipation of an inversion using numerical simulation or observations. In this study, a complete episode of wintertime persistent inversion over Salt Lake City is simulated using high-resolution WRF model with three nested domains method. Firstly, use the Skew-T plots of the observation soundings in the International Airport of Salt Lake City (KSLC) to determine the lifetime of the persistent inversion to be from1200UTC30November to0000UTC7December2010, and select0000UTC29November to1200UTC8December2010as the simulation period. Secondly, use the Northern American Mesoscale (NAM) model analysis data to provide the boundary and initial conditions to ran WRF model with three nested domains method. The horizontal resolutions of the three domains are respectively12km,4km and1.33km, and the vertical resolution is60σ levels. After comparing the simulation results and the observation soundings, it is found that temperature, wind speed and wind direction from the simulation agree well with those from the observation in the atmospheric boundary layer and above, although there are some differences near the surface due to the influence of complex terrain over the area, and the trend of the simulated inversion top is consistent to the observation. In addition, the results also show that the trend of the simulated inversion depth is similar to the observation, although there are some differences between the simulated inversion types and strength and the observed ones during the persistent inversion.In this study, the simulation results and the observations are used to analyze the physical processes and mechanisms of the formation, duration and dissipation of the persistent inversion. It indicates that the inversion formed mainly because of the interaction between the "heating" effect from a high-pressure ridge in mid-atmosphere and a near-surface cold pool. The cold pool resulted from large albedo and melting of snow cover on the ground, weak cold advection from the south opening of the Salt Lake Valley and cold flow from the slope to the east of Salt Lake City. After that, the inversion was persistent under the conditions of the high-pressure ridge aloft and very weak vertical motion. Finally, the "cold" effect from a low-pressure trough at mid-atmosphere, combined with the mixing effect due to vertical motion, led to the dissipation of the persistent inversion.Lanzhou, a city in the west of China, which is at the same latitude band and almost the same altitude and in the same geographical environments as Salt Lake City, suffers wintertime persistent inversions and serious air pollution every winter. It is believed that the results of this study can provide some good reference for the research on persistent inversions over Lanzhou.Besides the persistent inversion described above, strong convective storms are very common weather phenomena over the center of America in late spring and early summer. Strong storms can result in plenty of rainfall leading to floods and huge damages on industry, agriculture and human’s lives. Hence, the accurate forecast of timing, location and strength of convective storms is important for disaster prevention and mitigation. However, it still remains a challenge for the lack of observation stations and inadequate understanding of the physics processes in the model. So far, the effective forecasting method is still stuck in the use of radar in tracking and short-term forecasting of storms. If the mechanisms of convective initiation are further understood, and observations of some selected physics fields are assimilated into data assimilation system to improve the initial condition, it is believed that the forecast of convective initiation and subsequent storms evolution can be greatly improved. In this study, a convective process including several storms initiation and subsequent evolution over the center of America in2011is simulated using the high-resolution WRF model and its3DVAR data assimilation system. Firstly, a sensitivity experiment including three simulations with different options (Microphysics, Surface Layer and Planetary Boundary Layer options) is conducted to choose a relatively good simulation to be the control experiment. It is found that the simulation with Lin et al. scheme (Microphysics option), Eta similarity (Surface Layer option) and Yamada-Janjic scheme (Planetary Boundary Layer option) is relatively good at simulating the five convective initiation events and three squall-lines, and selected to be the control experiment. Three of the five convective initiation events are primary initiation, and the other two of them are secondary initiation.In this study, surface observations including wind and temperature are assimilated into WRF three-dimensional variational data assimilation (3DVAR) system. It indicates that assimilation of surface observations can really improve the location and timing forecast of primary initiation, secondary initiation and subsequent storms evolution through improving the wind, temperature and humidity analysis and forecast fields. The combined analysis of the3DVAR experiment results and the surface observations show that primary initiation is apt to occur in the intersection of surface boundaries, such as cold front, dryline and outflow boundary, and there are several common conditions for primary initiation, such as moist and warm circumstances and cyclonic wind field. However, secondary initiation takes place under the influence of existing active storms, such as cold pool or gravity wave. In addition, primary initiation is apt to take place in warm and cold advections intersection regions and positive humidity advection regions, which are not needed to be satisfied for secondary initiation.The accurate forecast of timing, location and strength of strong convective initiation is still a difficult and hot issue. Although there have been many observation and simulation studies, it is rare to use data assimilation method to study convective initiation. In this study, surface observations including wind and temperature fields are assimilated into WRF3DVAR data assimilation system. The results indicate that it is almost impossible to accurately reproduce the convective initiation and subsequent storms evolution during a long period only by WRF model, and data assimilation of surface observations can greatly improve the forecast of the storms. In addition, the mechanisms of convective initiation and evolution of strong storms of this study are believed to be helpful in further studying strong convective storms in the middle and east of China.