| description abstract | his paper explores the response of the tropical hydrologic cycle to surface warming through the lens of large-domain cloud-system-resolving model experiments run in a radiative?convective equilibrium framework. Simulations are run for 55 days and are driven with fixed insolation and constant sea surface temparatures (SSTs) of 298 K, 300 K, and 302 K. In each experiment, convection organizes into coherent regions of large-scale ascent separated by areas with relatively clear air and troposphere-deep descent. Aspects of the simulations correspond to observed features of the tropical climate system, including the transition to large precipitation rates above a critical value of total column water vapor, and an increase in convective intensity with SST amidst weakening of the large-scale overturning circulation. However, the authors also find notable changes to the interaction between convection and the environment as the surface warms. In particular, organized convection in simulations with SSTs of 298 and 300 K is inhibited by the presence of a strong midtropospheric stable layer and dry upper troposphere. As a result, there is a decrease in the vigor of deep convection and an increase in stratiform precipitation fraction with an increase in SST from 298 to 300 K. With an increase in SST to 302 K, moistening of the middletroposphere and increase in lower-tropospheric buoyancy serve to overcome these limitations, leading to an overall increase in convective intensity and larger increase in upper-tropospheric relative humidity. The authors conclude that, while convective intensity increases with SST, the aggregate nature of deep convection is strongly affected by the details of the thermodynamic environment in which it develops. In particular, the positive feedback between increasing SST and a moistening upper troposphere found in the simulations, operates as a nonmonotonic function of SST and is modulated by a complex interaction between deep convection and the environmental relative humidity and static stability profile. The results suggest that projected changes in convection that assume a monotonic dependence on SST may constitute an oversimplification. | |
