A case study of the error growth evolution in a meiyu front heavy precipitation forecast and an analysis of the predictability

The prediction of heavy rainfall along the meiyu front in China in July 2003 is studied. Based on the study results of the sensitivity of precipitation prediction in the AREM to the different initial data analyses and the different initial errors of variables, the evolution of error growth and its mechanism are decribed and discussed in details in this paper. The results indicate that an initial error with smaller amplitude will grow faster and afterwards the evolution of error growth following the occurrence and development of rainfall is characterized by a transition from local growth to global spreading to the overall domain with the remarkable error growth occurring in the rain belt, suggesting that the rain belt is a sensitive area for error growth. The initial errors in the rain belt contribute significantly to the error of precipitation prediction. The distribution of initial moisture has not only important effect on the propagation of error growth, but also drives meso- and small- scale errors to grow rapidly and subsequently causes the error growth in larger-scale motions. Since the computational results based on the error energy formula show that the energy for error growth is largely provided by latent heat, it is therefor concluded in terms of energy that the precipitation and error growths have the same “energy source”, which imposes intrinsic limitation on the predictability of heavy rainfall.