The Mesoscale Predictability Of One Heavy Precipitation Event

Using mesoscale model MM5,the influences of realistic initial perturbations from model on the mesoscale predictability of the"0185"Shanghai heavy precipitation are investigated. The initial differences between NCEP and T106 are divided into two parts as MM5 initial perturbations by two-dimension (2D) spectral decomposition, one is the part with different initial scales and amplitudes, another is the part with different initial scales and rescaled amplitudes. The influences of random and idealized initial perturbations and moisture processes on the mesoscale predictability of the"0185"Shanghai heavy precipitation are also studied. The major results and conclusions of this study are summarized as follows:1,The results show that the evolution of the synoptic environment and the intensity and location of the heavy precipitation in Shanghai in a 24-h lead time can be reasonably simulated by MM5. Moreover, the time evolution of hourly rainfall in Shanghai simulated in CNTL also compares favorably to the relevant average observations during the same period. The evolution and power spectrum of the DTE between CNTL and T106ic (T106bc) shows that the initial uncertainties lead to larger forecast divergence than those with different lateral boundary uncertainties. It's also found that forecasting error growth is independent on the scales and amplitudes of the initial perturbation.2,The evolution of DTE indicates that it grows quickly in small scales, and the DTE error growth approaches a level of saturation about 12h, then the atmosphere will involve into the state with random process. The time of 12h is also named predictable time of this heavy precipitation event. The spectrum distribution of the accumulated precipitation differences between CNTL and T106ic shows that the accumulated precipitation differences beyond that of CNTL at the scales below 300km at the same time, it can be almost said that it is a complete loss of predictability below these scales for the heavy rainfall.3,The predictable forecasting time and space can be expanded under the function of initial perturbations with small scale and amplitude, but the time and space are mainly determined by the perturbed amplitude.4,Moisture processes are found to be the key to the predictability of the heavy rainfall. The intensity and location of such heavy precipitation are especially sensitive to the initial perturbation of moist field. The spectrum distribution of the accumulated precipitation differences between CNTL and the sensitive experiments which added moist perturbed fields shows a complete loss of predictability of precipitation in the larger scales, but the predictable forecasting time and space can be also expanded under the function of initial moist perturbations with small scale and amplitude...