Intense Atmospheric Vortices

This paper discusses the likely physical processes which are responsible for distinguishing between tropical disturbances which develop into tropical cyclones and disturbances which do not.

The pre-typhoon cloud cluster disturbance is differentiated from the prominent non-developing cloud cluster disturbance primarily by its surrounding 2–8o radius tropospheric wind and vertical wind shear patterns. Other parameters such as mean vertical motion, moisture, convergence, vertical stability, surface wind speed, etc. are not significantly different. The favorable wind and vertical wind shear patterns are initially established by general circulation features.

A paradox to the tropical cyclone development problem is the dual role of the tropical disturbance’s transverse circulation. The in-up-out mean radial circulation needed in a mechanical sense to spin-up the tangential flow also acts in an opposite sense to export moist static energy from the disturbance. Outflow air has higher energy than inflow air. Weather system maintenance and development can occur only if the disturbance’s surface energy flux is large enough to overcome the energy losses from the mean radial circulation and radiation. This requires extra energy flux from the ocean beyond that given by the bulk formula from space and time averaged wind speed and moisture values. Such extra surface energy fluxes have been calculated from budget analysis of our rawinsonde data as a residual and are observed to be related to the product of the weather system’s surrounding vertical wind shear and upward vertical motion.

It is hypothesized that the pre-storm weather system’s larger surrounding region tropospheric vertical wind shear patterns lead to the development of more frequent and more intense rainband convection. The downdrafts developed by these rainbands tap the ocean for more energy by simultaneous increasing surface wind speed and decreasing moisture and temperature. This results in a nonlinear increase of surface energy flux along the rainbands. These extra rainband energy gains cause a significantly larger surface energy flux than that specified by the broadscale bulk formula using composited data. This surface energy increase is diffused throughout the vortex and allows the energy depleting mean radial circulation to be balanced. Smaller residual energy amounts are often available to intensify the vortex. Recent numerical modeling experiments by W. Fingerhut (1981) are verifying that such enhanced surface energy fluxes and vertical wind shear relationships can differentiate developing weather systems from those which do not develop.

The tropical cyclone is a mesoscale power plant with a cyclone-scale supportive system. While the cyclone can be handled basically as a quasi-balanced axisymmetric vortex, the energetic core of the vortex is in a fully developed nonlinear regime of three-dimensional convection. One of basic problems of tropical cyclone theory has been the handling of the two interacting scales as one system. The purpose of this paper is to examine the conceptual development in tropical cyclone modeling, which is now returning to explicit calculation of the convective scale in order to simulate the full spectrum of scale-dependent dynamics.

Research aircraft have been probing the inner core of tropical cyclones on a regular basis for more than two decades. Most of what we know today about the detailed internal structure and dynamics of these storms has resulted from studies based upon data collected from these airborne instrumented platforms. Noteworthy among earlier studies are those by Riehl and Malkus (1961) and LaSeur and Hawkins (1963). Riehl and Malkus used data collected in Hurricane Daisy (1958) to try to deduce the thermal and dynamical characteristics of a developing and mature storm. LaSeur and Hawkins used multiple level data collected in Hurricane Cleo (1958) to study the three-dimensional structure of a mature hurricane. Later, Hawkins and Rubsam (1968) conducted a detailed diagnostic study of Hurricane Hilda (1964), computing many of the same quantities as those computed by Riehl and Malkus in their 1961 study. Although, in general, findings were similar, significant differences between the two studies appeared, such as the depth of the inflow layer and the role of local generation as compared to advection for increases in the kinetic energy in the inner core of the hurricane. Similarly, several other case studies (Colon, 1961, 1964; Sheets, 1968, 1973; Hawkins and Imbembo, 1976) have shown gross features similar to those of the early Hurricane Cleo study, but have also shown wide variations from storm to storm and even from day to day within a given storm.

Angular momentum budget equations for a moving tropical cyclone are presented and the physical role and relative magnitude of each component is described.

China is one of the areas upon which the activities of large-scale intense atmospheric vortices have their influence most frequently and seriously. Take typhoons for example, more than one third of the tropical storms are generated in the North-Western Pacific. In the recent 30 years, from 1949 to 1978, the total number of typhoons generated over the North-Western Pacific was 858, among which 204 typhoons had landed on Chinese coasts. And more than one third of these landed typhoons had moved farther into our interior provinces and regions with a recorded maximum wind speed of 60 m/sec.

Numerical experiments have been performed in search of the favorable conditions of environmental wind for the genesis of a tropical storm. The low level basic flow has an impact on the latent energy supply which is essential for genesis. The upper level basic flow has to be coupled with the phase speed of the low level incipient disturbance. The low level cyclonic shear of the basic flow is conducive to storm genesis.

The role of large-scale eddy fluxes of momentum in the intensification of Atlantic hurricanes from prehurricane cloud clusters, depressions and cyclones is investigated. This is accomplished by numerically integrating SUNDQVIST’s (1970) symmetric model for hurricane development with parameterized large-scale eddy fluxes of momentum. The initial wind and moisture distributions and the prescribed eddy fluxes are taken from atmospheric observations of Atlantic developing hurricanes as composited by MCBRIDE (1981a, b) and MCBRIDE and ZEHR (1981). With the use of the same initial wind, temperature and moisture distributions, numerical integrations were performed with and without the observed eddy fluxes of momentum.

Without eddy flux forcing, the prehurricane cloud cluster and the prehurricane depression do not develop into hurricanes, and the intensifying cyclone takes approximately 13 days to reach hurricane strength. With the observed eddy flux forcing, all three initial disturbances develop rapidly into hurricanes, the intensifying cyclone requiring less than 2 days to reach hurricane strength. The rate of hurricane development, and the size, intensity and eye structure at maturity are different, depending upon the details of the eddy momentum flux distribution. It is concluded that inward eddy fluxes of momentum associated with synoptic-scale wave asymmetries may be crucial to the development of Atlantic hurricanes from prehurricane cloud clusters, depressions and cyclones.

The weather of China is greatly affected by typhoon. Typhoons, which move over China or close Chinese coasts,usually cause gales and rainstorms in the adjacent sea areas. Furthermore, some of typhoons may generate heavy rainfall in North China when they are interacting with the cold air of westerlies. Such kind of rainfall process is considered as “indirect” typhoon rainstorm, which constituted the majority of the heavy rainfall (≥100 mm.day^-1) and flood. Since the “Aug.1975” heavy rain in Henan Province (1060 mm in 24-hour,685mm in 6-hour and 189.5mm in one hour), Chinese meteorologists have paid great attention to these rainfall processes (Beijing Group,1979; Xie An et al,1979). The case selected in this study is a landed typhoon. Besides a short synoptic analysis of this case,some simple numerical experiments of real data forecast are made by using a five-level fine-mesh model. A better understanding to the physical mechanism of the rainfall may be expected.

This paper describes an analytic solution relatingthe radial distribution of the radial component, K1 , of eddy viscosity to the observed radial distribution of tangential velocity in a mature hurricane. It is an extension of a previous paper by EVANS and DAVTES (1979), which was concerned with the vertical distribution of the eddy viscosity components. Over most of the region between the eye boundary and the position of maximum tangential velocity the radial distribution of Kl is found to be very close to a linear function of the radial distribution of tangential velocity, but increasing sharply in the peak velocity region. The corresponding radial distribution of vertical velocity is then found to be very similar to typical observed profiles.

Moist static energy and momentum budgets for an intensifying tropical cyclone indicate that the radial motion determines the angular momentum source and at the same time the moist static energy sink. On the other hand, surface transfers determine the angular momentum sink and the moist static energy source. It was pointed out by Pearce [1981] that whether or not intensification can occur depends crucially on the vertical distributions of the moist static energy and angular momentum of the lower and middle tropospheric inflow air.

A simple axisymmetric model of intensification, in which the air flowing through the system in a contracting convective annulus is treated separately from the air in the interior, is briefly described. The interior may be regarded as responding, through gravity wave propagation, to dynamical forcing by the air flowing through the clouds.

The maintenance of the radial flow (or secondary circulation) is discussed, with the buoyancy of the inflow air highlighted as the main controlling factor. The dynamics of cumulus convection in axisymmetric flow and the factors determining contraction (intensification) are described.

It is shown that the mass adjustment to the primary (tangential) flow, accomplished by gravity waves, is rapid but that gradient balance cannot be achieved in an axisymmetric system. Instead the central pressure fall is less than that required for balance, i.e. the tangential winds are super-gradient. This is the result of the existence of the secondary flow with its large non-linearities.

These concepts are briefly contrasted with ideas based on CISK.

The recent availability of doppler radar for observations of convective storms has produced convincing documentation of the importance of rotating storms and their close association with severe weather. Although not all severe thunderstorms rotate, an operational evaluation study conducted at NOAA’s National Severe Storms Laboratory showed that most of the storms containing meso-cyclones identifiable by doppler radar produced either tornadoes or large hail within an hour.

Intense cumulus rotation has been found to precede severe tornadic storms. Substantial advances in documenting and understanding precursor conditions for destructive tornadic vortices have been made using Doppler radar observations. A typical tornado-associated “mesocyclone” has a diameter of 5–10 km, tangential wind velocities 15–25 ms^-1 and vorticities in the neighs borhood of 10^-2 s^-1. Only about 60 percent of these strong meso-cyclones produce funnels, however, suggesting that a special juxtaposition of a number of conditions is required.

The tornado is the most intense atmospheric vortex, and the most mysterious. Many questions still remain about its formation. The field of tornado dynamics has been reviewed recently by DAVIES-JONES (1981). In theoretical work it is generally assumed that the tornado is a vertical vortex, but it is often inclined appreciably. The direction and degree of tilt offers insights, hitherto unexplored into the character of the velocity field in which the tornado is embedded.

Recent multiple Doppler radar observations of supercell thunderstorms have documented that a particular dynamic structure consistently develops when a storm enters its tornadic phase. This structure is characterized by a downdraft outflow which wraps around the center of storm circulation at low levels and an associated distortion from a single cell to a two-cell up-draft structure (Brandes, 1978; Lemon and Doswell, 1979). In this study we investigate the evolution of a storm’s tornadic phase through high resolution numerical simulations initiated within the central portion of a 3D cloud model mature thunderstorm.

Preliminary results from an axisymmetric numerical simulation of tornado growth in a mesocyclone updraught are described. Using an observed tornadic storm proximity sounding, the calculations show that the distribution and degree of buoyancy in a mesocyclone updraught can account for the generation and maintenance of an intense tornado when the initial level of storm rotation is within the observed range of values. The structure of the mature tornado is highlighted by a comparison of the vertical force balance in a rotating and non-rotating updraught simulation. The present paper extends our recent studies of tornadogenesis to include the effects of moisture diffusion and of negative buoyancy due to the water loading of the updraught.

It often occurs that tornadoes contain smaller subsidiary vortices which revolve about the tornado, which in turn, rotate in the same sense as the tornado (see FUJITA, 1970; FORBES, 1978). This phenomenon goes by various names (“suction vortices”, “satellite vortices”, and “secondary vortices”), however, for this work we follow CHURCH et al. (1979) and refer to the multiple vortex phenomenon (MVP). So, when we speak of the MVP we refer to that type of “suction vortex” which FUJITA (1976) has termed the “orbiting vortex”. That the MVP may be more than a minor detail of the tornadic flow is suggested by damage surveys which indicate the most intense destruction of life and property is associated with cycloidal paths which FUJITA has termed “suction swaths”. As with the tornado, very little is known about the internal circulation of the MVP. Photogrametric and ground survey data are inconclusive on such important questions as to i) what maximum wind speed is achieved by the tornado and ii) the relation between this and the MVP. MVs are clearly visible in photographs but it is extremely difficult to infer actual flow patterns.

Recent experiments on concentrated vortices produced by swirling flow into a constricted, cylindrical tube have resulted, we believe, in a new understanding of vortex flows in a variety of similar devices, the main purposes of which have been to simulate atmospheric vortices. As a major part of this review we present a synopsis of these new results and show their connection to various laboratory models of atmospheric vortices. The utility of such models in the study of large scale atmospheric flows is also discussed.

The addition of swirl to an axial flow leads to flow phenomena which are not yet well understood even for simple geometric configurations. An essential feature of many such flows is that they are characterised by an inner vortical core of small diameter surrounded by nearly irrotational flow. Since the core can support waves, it may be either supercritical or subcritical in nature, and a relatively sudden transition from the former to the latter can occur in the form of vortex breakdown. The present paper represents a summary of the outcome to date of a continuing long-term study of vortex flows confined within cylindrical tubes in which vortex breakdown plays a significant role. The paper is organised around a number of specific aspects of such flows: the governing parameters, visualization of the core and the outer zone, the criterion for breakdown, measured velocity and circulation profiles.

Wall static pressure fields beneath tornado-like vortices have been investigated using a vortex generator designed to model tornado-cyclone airflow. The data presented include (a) a series of radial profiles of the mean surface pressure beneath the vortex as a function of swirl, (b) measurements of the magnitude of the greatest pressure deficit to be found under the vortex as a functionof swirl, and (c) a set of measurements of the pressure deficits associated with the individual subsidiary vortices in a multiple vortex system. These data are interpreted to show the development of the model.tornado cyclone from the no-swirl state, and the evolution of its intense vortical core from a one-celled into a two-celled flow. It is shown that the greatest pressure deficits and largest pressure gradients are associated with single-celled vortices. The pressure deficits within the subsidiary vortices are found to be two to three times as great as that at the center of the parent flow. This work provides the background for better physical interpretation of barographic records obtained in actual tornado cyclones, and further supports a model of the evolution of these phenomena that is consistent with doppler radar observations of natural tornado cyclones.

Laboratory experiments have revealed substantial reorganization of the mean motion in enclosed uniformly rotating fluid when subjected to intense stirring. The potential importance of these effects is reviewed in relation to the earliest stages of tropical cyclone genesis. A laboratory experiment in which localized mixing simulates the kinematics of a tropical disturbance cluster indicates that turbulently induced recirculations can bring about significant and rapid vortical organization.

Intense vortices are generated by turbulent forcing on a small scale of a fluid body initially in solid body rotation. The vortices extend over the entire depth and their core vorticity can be by a factor of about 50, and possibly more, larger than the background vorticity, 2Ω . The size of the vortices and their average spacing . depend on the ratio of intensity of forcing to background vorticity. The concentrated vortex cores support waves consisting of helicoidal distortions and these waves induce strong axial flow which can cause local convergence and vortex intensification. Wave interaction frequently results in vortex breakdown. Complete distruction occurs on a time scale of about 10–20 rotation periods. The formation time of vortices, when the forcing is suddenly switched on,is of the order of 6–7 rotation periods.

A number of phenomena related to labqratory-produced tornado-like vortices are observed and discussed. These are the formation of a columnar vortex above evaporating fluid; appearance of multiple vortices in a single convergence system; a suppression of turbulence in fluid in rigid-body rotation; and the formation of a tornado-like vortex in rotating fluid heated from below. Qualitative explanations of these observed phenomena are given.

The regular and irregular non-axisymmetric flow regimes of thermal convection in a rotating fluid annulus subject to differential heating in the horizontal are characterized by the presence of upper level jet-streams, where intense concentrations of vorticity and high concomitant horizontal temperature gradients are found. The main features of the upper-level flow pattern can be interpreted by straightforward arguments based on general thermodynamic considerations and the requirement that the flow should be quasi—geostrophic nearly everywhere. Thus, when the distribution of applied heating and cooling is such that the corresponding gradient of the impressed radial temperature field has the same sign at all radii, the most conspicuous feature of the upper-level flow pattern in the regular non-axisymraetric regime is a single jet-stream meandering in a wavy pattern between the bounding cylinders. When, however, the impressed radial temperature gradient changes sign near mid-radius (as in the case when heat is introduced throughout the body of the fluid and withdrawn at both side-walls), the corresponding upper-level flow consists of several closed eddies, each circulating “anticyclonically” with the horizontal flow largely confined to a narrow jet-stream at the periphery of each eddy. In some respects these stable closed eddies are dynamically similar to long-lived anticyclonic eddies (including the Great Red Spot) seen in Jupiter’s atmosphere in the southern hemisphere. Previous work on stable baroclinic eddies is now being extended in various directions and supporting numerical work is also being carried out.