This project is funded by ONR DRI “Unified Parameterizaiton for Extended Range Prediction” Program.
Recent observational, numerical, and theoretical studies have conclusively demonstrated that roll vortices (rolls) are prevalent in tropical cyclone boundary layers (TCBL). Rolls consist of overturning flows in the plane roughly perpendicular to the mean flow direction. They are spatially periodic with wavelengths that range from hundreds of meters to several kilometers. When rolls are present, the direct transfer of momentum, heat, and water vapor by these structures across the TCBL represents a potentially important contribution to the overall TCBL transport of momentum and enthalpy that is not currently included in hurricane forecast models. The lack of roll-induced fluxes in existing TCBL parameterizations has potentially major implications for simulating or forecasting hurricane intensity. The roll-induced fluxes are inherently nonlocal and non-gradient and hence cannot be captured by standard turbulence parameterizations. Thus, if the effects of TCBL rolls are to be included in hurricane models, new parameterizations of their effects are needed.
The objectives of this project are:
Our approach to parameterization of rolls in a TC model resembles the “super-parameterization” approach used to simulate cloud physics processes in general circulation models (Grabowski 2001). This methodology includes embedding a rolls resolving high-resolution 2D-LES model into the 3D equation system representing a TC model. The decomposition of a 3D equation system into two coupled equation systems for the mean flow and convective scale motions (rolls) is described in detail by Ginis et al. (2004) for an idealized 2D mean flow. In this project, we extended this procedure for a general case of a 3D TC model. The two models are coupled and explicitly solve the two-way interaction between the large-scale flow and rolls.
We coupled the 2D-LES model with an axisymmetric TCBL model to simulate the generation and evolution of rolls, as well as the interactions between rolls and the large-scale wind in TCBL.
We found that the TCBL flow is inherently unstable because of the existence of the inflection point in the radial wind profiles. Rolls generated by the inflection point instability are characterized with tilted streamlines in the vicinity of the inflection point. Kinetic energy budget consideration reveals the tilted streamlines are critical for the rolls to extract kinetic energy from the mean flow. The inflection point instability can easily operate if the stratification in the TCBL is weak. If the mixed layer is sufficiently high, rolls can couple with internal waves in the inversion layer.
After rolls reach finite amplitude, the roll-induced fluxes start to act on the mean flow. We found rolls can trigger persistent inertial oscillations in the mean hurricane flow, and they can also greatly enhance the momentum mixing in the boundary layer. As a result, the radial (inflow) wind speed is reduced and the inflow layer height is increased. In addition, the supergradient tangential wind in the TCBL is reduced as well.