Abstract: During the period of 20-23 July 1998 ("98.7"), an extraordinary heavy rainfall event occurred in the eastern and southwest parts of Hubei Province, which was in close relationship with the low vortex with shear line on 700 hPa as well as with the successively generated and intensively developing MαCS and MβCS along the shear line. The numerical simulation for this event is conducted successfully using nonhydrostatic version MM5 (V2.12). The mesoscale heat and moisture budgets are diag nosed using the output data of MM5. The finding sare as follows: strong apparent heat sources Q₁ and apparent moisture sink Q₂ would appear while the strong convection occurs with heavy rainfall; Q₁ and Q₂ are basically correspondent with the regions of heavy rains; Q₁ increases with height while the maximum heating level is located at 486.1 hPa (σ=0.54); the deep and thick heating layer in the middle troposphere is the main thermody namic mechanism for the both cumulus convective activities and the successive generation and development of severe rainstorms; the relatively cool layer in the upper troposphere provides a favorable thermody namic condition for intensive cumulus convection over the heavy rain areas; the condensation latent heating of cumulus convection in the lower and middle troposphere is not only heating the middle troposphere atmosphere, but also heating the upper environment atmosphere through transporting heat upward; the double-peak structure of Q₂ at the early stage of heavy rainfall is relevant to the drying associated with convective condensation of both stratocumulus at the lower levels and cumulus at the middle levels; the peak value of Q₂ in the middle level is basically corresponding with that of Q₁; the deep and thick drying layer is consistent with the deep and thick condensation-heating layer of Q₁. The diagnostic results show that it is reasonable to diagnose Q₁ and Q₂ using successfully stimulated output data with high resolution. The reliable physical proof for improving heating and moisturizing profiles in the cumulus convective parameterization can be obtained through diagnosing Q₁ and Q₂ of the rainstorm event.