A Diagnostic and Numerical Study on Surface Rainfall Process of Tropical Cyclone Soudelor (2015) over the Ocean: Sensitivity Experiments on Precipitation Response to Sea Surface Temperature Change

To investigate possible impact of sea surface temperature (SST) change on surface rainfall process of tropical cyclone（TC） Soudelor over the ocean, sensitivity experiments are conducted using the Weather Research Forecasting (WRF) model and a 3D WRF-based precipitation equation based on the previous study of Wang et al. (2018). SST in the control experiment (CTL experiment, SST changes with time) is much different to SST in the sensitivity experiment (SNC experiment, SST remains constant). The results show that the difference in the simulated TC track over the ocean is small in the two experiments, but the simulated TC intensity in the SNC experiment is stronger than that in the CTL experiment. For the TC circulation, differences in the vertical velocity between the two experiments are basically positive in the troposphere (the SNC experiment yields stronger updrafts than the CTL). As the difference in SST increases, the vertical motion difference also becomes larger. Precipitation rate (P_S )in the SNC experiment is larger than that in the CTL experiment, but the difference in P_S between the two experiments increases non-linearly with the SST difference, which reflects the complexity of P_S variation. Q_WVA (the three-dimensional moisture flux convergence or divergence rate) in the SNC experiment is basically larger than that in the CTL experiment (especially in the later period of the TC life cycle). This is because SST difference can affect vertical motion and cause Q_WVA difference, which in turn affects P_S. During the study period of the two experiments, the atmosphere inside the TC continuously becomes drier (positive Q_WVL), and evaporation from the sea surface (positive Q_WVE) is significant._. There is little difference in Q_WVL between the two experiments. However, Q_WVE in the SNC experiment is generally stronger than that in the CTL experiment (especially in the middle and later stages of the TC life cycle). Temporal variations of the change rate of hydrometeor-related processes in the two experiments are complicated, but the magnitude of the maximum difference is comparable to that of Q_WVE. SST affects the growth of cloud hydrometeors content and deep convection, and thus differences in the variation of hydrometeors content gradually increased with SST difference between the two experiments. Among liquid-phase hydrometeors, large difference is found in raindrops, while differences in ice-phase hydrometeors, particularly large ice particles (snow and graupel), are even greater. In the SNC experiment, more graupel particles and raindrops are concentrated below the melting layer, which, concomitant with stronger updrafts, contribute to the occurrence and enhancement of P_racw(accretion of cloud water by rain) and P_gmlt（melting of graupels） and eventually enhance precipitation. Comparative analysis of regionally and temporally averaged macroscopic and microphysical processes related to the TC rainfall between the CTL and SNC experiments shows that the rainfall processes in the two experiments are qualitatively similar but quantitatively different. Compared with the CTL experiment, P_S in the SNC experiment increased by 8.8%, which is mainly resulted from the difference in macroscopic and microscopic physical processes of precipitation between the two experiments. The different in Q_WVE is_{the most obvious with different SST.}