Tropical Cyclones in Rotating Radiative-Convective Equilibrium with Coupled SST

Radiative-convective equilibrium (RCE) is a useful idealized framework for studying the tropical atmosphere. In the simplest version of RCE one typically ignores spherical geometry and places the flow in a doubly-periodic domain in which the forcing and boundary conditions are all horizontally homogeneous. This allows the study of the interactions between radiation and moist convection, in a simple geometry. In many cases, the flow is assumed to be non-rotating, but it is also very interesting to consider rotating radiative convective equilibrium (RRCE) by using an f-plane geometry. If the domain is large enough it fills up with long-lived tropical cyclones (TC). Both the size and intensity of these tropical cyclones are intrinsically determined and their parameter dependence can be studied. RRCE can be investigated both with cloud-resolving models in which deep convection is partly resolved (Khairoutdinov and Emanuel, 2013), and also with lower resolution and the column physics of a global comprehensive model (Held and Zhao 2008; Zhou et al., 2014). The latter case provides an idealized setting for evaluating the TC simulations in global comprehensive models (GCMs). GCMs are moving towards higher resolution such that many aspects of their TC simulations are becoming more realistic. They are also one of the tools used to predict the impact of climate change on TC statistics. Being able to study these GCM-generated TCs in this idealized geometry, and eventually comparing to analogous simulations at cloud-resolving resolutions, should provide valuable information about the limitations and strengths of TC simulations in GCMs. Previous RRCE studies all prescribed a horizontally homogeneous sea surface temperature over the domain, and thus neglected the response of sea surface temperature to anomalous surface flux perturbations produced by tropical cyclones. In this study, we achieve RRCE with the column physics of the high resolution atmospheric model at GFDL following the line of Zhao et al. (2014), at 25km horizontal resolution, but with sea surface temperature coupled to surface fluxes through a simple slab ocean layer. The impact of this slab ocean layer on tropical cyclones is studied by varying the ocean layer depth H in a wide range from 0.02 m to 10 m. With H = 10 m, SST in the eyewall region is about 1 k cooler compared to the environment, due to both the negative cloud forcing and the anomalous surface fluxes there. As H decreases, the negative cloud forcing is almost invariant such that its absolute cooling effect on SST increases. However, such cooling effect is largely compensated by the surface fluxes at the eyewall region which are significantly suppressed as the SST cools. As a result, the relative cooling of SST at the eyewall region is limited at about 7 K with H = 0.02 m and tropical cyclones are still sustained but with a reduced intensity compared to those with H = 10 m (Fig. 1). The original CAPE algorithm for Potential Intensity (PI, Bister and Emanuel, 2002) assumes that SST, surface air temperature and relative humidity at the outer eyewall are the same as those of the environment. It is thus not surprising that it fails to predict the decreasing TC intensity as there are indeed significant differences between the outer eyewall and the environment when H is small. We introduce a modified PI algorithm, which is able to capture the deceasing tropical cyclone intensity as the ocean layer depth decreases. We also show that the value of surface relative humidity in the eyewall region has a significant influence on the relationship between the PI minimum pressure and the PI maximum wind velocity in this modified algorithm.