Numerical simulation of tropical cyclone genesis

Abstract

Past hurricane or typhoon simulation studies have concentrated on the development of the mature storm. Most of these simulations began with initial states that had closed, relatively strong cyclonic circulations and that ultimately produced hurricane disturbances during the integrations. Numerical modelers have generally not attempted to simulate the development of these initial closed tropical cyclones from which hurricanes grow. This research simulates some aspects of such tropical cyclone development and the factors that influence it.The numerical model uses simple but adequate physics for ease in interpretation and computational economy. Model characteristics include the following: (1) primitive equations, (2) three layers, (3) incompressible fluid, (4) boundary layer latent and sensible heat exchange with the sea surface, and (5) CISK parameterization as a function of boundary layer convergence and a vertical stability parameter. Because the model has incompressible dynamics, the cumulus convection parameterization simulates the basic effects of cumulus heating as a boundary-layer-forced mass sink in the lower level and source in the upper level. Model integration proceeds with semi-implicit time differencing on a space-time staggered grid.Most experiments used an initial state based upon a GATE composited easterly wave. Simulated development was highly sensitive to thermodynamic conditions. With a sea surface temperature of 28.6(DEGREES)C and an initial boundary-layer equivalent-potential temperature of 350 K, the composite easterly wave intensified little; but with analogous conditions of 30.3(DEGREES)C and 360 K, the wave rapidly developed.The experiments suggest that low level vorticity is a much more important intensification criterion than upper level factors. Results do not support the hypothesis that vertical shear of the horizontal wind has a large negative effect on tropical cyclone genesis or intensification. Experiments with a crude cumulus-momentum-transfer parameterization revealed a large damping effect on intensification.