Abstract

Rainfall rates determined from airborne radar and infrared satellite images are combined to construct a space- and time-dependent heating function for Hurricane Anita (1977). The heating is assimilated into a three-dimensional primitive equation prediction during a 12 h pre-forecast integration, after which the heating rate is computed internally by the model. The specified heating forces initial wind and mass fields toward their observed values, and produces improved 12 and 24 h forecasts of both track and intensity compared to a control integration, for which the heating is computed internally for the entire period.

Calculations indicate that model adjustment during the period of heating can be viewed as a slow response of the vorticity field to continuous forcing of the divergence by the heating. The location and pattern of the heating relative to the center appear to be of greater importance than the magnitude of the heating. This may be of significance because remotely-sensed rainfall estimates are more likely to be accurate in the positioning of heavy rainfall than in its intensity. The initialization procedure appears capable of producing useful improvement in short-term hurricane prediction, particularly prior to landfall, when data coverage is best and accuracy is of greatest concern.

A number of authors have noted the importance of upper-level inward eddy momentum fluxes for hurricane intensification. Calculations from the simulated storm indicate that such eddy fluxes are present in Hurricane Anita and are associated in part with an anticyclonic outflow eddy over an intense local rainfall area 300 km east of the center.

Abstract

A formulation of Kuo's cumulus parameterization is described which satisfies arbitrary vertical profiles of apparent heat source (Q ₁) and apparent moisture sink (Q ₂). The approach requires little calculation, and for a given Q ₁ and Q ₂, contains only one parameter, the partitioning of available moisture between storage and precipitation. The proposed method is tested in the prediction of a mesoscale convective complex and its advantages and limitations are discussed.

Abstract

A closure is proposed for the b parameter of Kuo (1974), using the framework developed by Krishnamurti et al. (1976). Emphasis is placed on the time-dependent behavior of the solutions. The proposed closure is found to be the only one of several tested to produce an approach to moist neutrality in both temperature and moisture under strong external forcing. The sensitivity of the grid-scale evolution to the partitioning of moisture defined by the b parameter suggests that such partitioning must be carefully dealt with in any method for computing the effects of cumulus convection, whether or not b is explicitly present.

By including entrainment in the cloud lapse rate, the observed large-scale behavior of the vertical profile of moist static energy under disturbed conditions is simulated. The approach is shown to be easily invertible when precipitation rate is specified, thus insuring internal consistency in a model when such a procedure is used as part of a dynamic initialization.

Because it is relatively simple and general, and reproduces observed large-scale θ _e variations under strong forcing, the approach may be particularly suitable for large-scale models. An economical way to extend the procedure to mesoscale models is proposed.

Abstract

A simple framework is presented for adjusting the normal wind components in a polygon of data points which produces a vanishing vertical integral of horizontal divergence, allows correct calculation of flux and advective terms, and permits virtually any choice of vertical profile of divergence adjustment. The procedure was used to estimate precipitation as a residual from vertically integrated heat and moisture budgets for SESAME data, in order to evaluate the uncertainty introduced by commonly used approximations in diagnostic studies. Although the method cannot be applied on a grid in its current form, the results remain valid for gridded calculations.

Line integrals around the polygon were carried out analytically, allowing an exact calculation of eddy fluxes within the assumption of linearity along the edges. Finite difference approximations for nonlinear terms were shown to introduce significant errors, even under ordinary circumstances.

It is common practice to neglect the horizontal advecting velocity adjustment brought about by the adjusted divergence. Such an assumption produced negligible median errors in the integrated heat and moisture budgets. The median differences in calculated precipitation caused by differing choices of the divergence adjustment profile reached 1.34 andO.35 cm day⁻¹ in the heat and moisture budgets, respectively. Because the true divergence adjustment profile is unknown, these values represent median lower bounds on the errors in budget estimates of precipitation in middle latitude convection.

Abstract

The ability of several explicit formulations of convective heating to predict the precipitation associated with a mesoscale convective complex was compared to that of a cumulus parameterization on a ½ deg latitude-longitude mesh. In the explicit approaches, prediction equations were present for both water vapor and cloud water, or vapor alone. The simplest explicit approach, for which any condensed water was assumed to fall immediately as rain, produced localized excessive rainfall. This explicit heating instability arose as a result of the requirements of saturation prior to rainfall, which delayed condensation and allowed excessive convective instability to build, and neglect of fluxes, which prevented the instability from being released in a realistic manner. These results, combined with those of previous investigators, indicate that the simplest form of explicit heating is prone to instability and unsuitable for mesoscale models.

Instability problems were significantly reduced by the inclusion of the inhibiting effects of rainwater evaporation and a cloud phase with hydrostatic water loading, Nevertheless, bemuse significant nor occurred in nature in the absence of area-averaged saturation, rainfall was unrealistically delayed when a 100 percent saturation criterion was used. Reducing the saturation criterion improved the phase error of the rainfall prediction, but sometimes reintroduced local instability.

Although only simple explicit formulations were used, inclusion of more sophisticated microphysical parameterizations from cloud models may be unrepresentative of processes in nature for meso-α scale models, for which the grid spacing exceeds 50 km. It is proposed for such models that implicit approaches offer the greatest potential for improvement. For meso-β scale models the optimum choice remains uncertain.

Abstract

Cumulus and mesoscale downdrafts are incorporated into the cumulus parameterization of Kuo. Convection is driven by grid-scale moisture supply, and distributed vertically by temperature and specific humidity differences between the environment and an idealized cloud. The moisture supply is defined to minimize the problem of lag between instantaneous moisture accession and rainfall. Downdrafts are added to the idealized cloud profile by determining a weighted mean of the equivalent potential temperatures (θ _e ) for cumulus updrafts, saturated cumulus downdrafts, and unsaturated mesoscale downdrafts, and by extracting the cloud temperature and specific humidity iteratively from the mean θ _e . The θ _e values are weighted by the mean vertical eddy flux convergence of moist static energy by each component.

The addition of downdrafts sharply increases the rate of stabilization of the grid scale by the Kuo approach. Stabilization characteristics are also shown to depend upon precipitation efficiency, strength of grid-scale forcing, downdraft relative humidity, downdraft weighting, and intensity of surface fluxes.

The approach was tested in a real-data, three-dimensional primitive equation prediction of a mesoscale convective complex (MCC) on a 1° latitude/longitude mesh. Prediction of total rain volume was most accurate when downdrafts were included. Without downdrafts, a feedback instability occurred at the MCC center and rainfall was greatly overestimated. When convective heating was omitted, so that rainfall could be produced only after grid-scale saturation, predicted rainfall was less than 10% of that observed and the MCC decayed. Difference vectors between the full and no convection integrations showed strong outflow developing in the upper troposphere, evolving to a large anticyclonic eddy following the MCC by hour 12 of the forecast. Corresponding inflow and a weak cyclonic eddy developed at low levels. Influence of the MCC spread rapidly over several hundred kilometers through this divergent flow. The results indicate, not surprisingly, that maintenance of the MCC depends critically on the presence of cumulus convection. The failure of the explicit (nonparameterized) approach suggests that cumulus parameterization is necessary for realistic prediction of convective systems in meso-α scale models.

In the integration with downdrafts incorporated, a life cycle behavior occurred in the heavy rainfall region. The level of maximum upward motion shifted from middle to upper levels over several hours, and downward motion developed at the lowest levels. The, apparent heat source was initially positive at all levels, then became negative in the lower troposphere and more strongly positive aloft. Stratiform precipitation fell from saturated upper levels for a brief period after convection ceased. This life cycle behavior, which contains several aspects of that observed, took place only when cumulus and mesoscale downdrafts were incorporated.

Abstract

A 10-yr climatology (1986–95) was performed using ECMWF gridded analyses on isentropic surfaces to identify regions where the lower-tropospheric meridional potential vorticity (PV) gradient changes sign across Africa and Australia during their respective summer seasons. While an African sign reversal has been documented, no similar study has been performed for the Australian region, which also has desert on the poleward side of open ocean. In each hemisphere, a northward decrease of PV is sufficient to produce a sign reversal.

It was found that PV decreases northward in the lower troposphere across northern Australia, with the maximum reversal on the 315-K surface. It had comparable magnitude but smaller zonal extent (∼3000 km) than that on the 320-K surface in Africa (∼5000 km). In each region the sign reversal was associated with cyclonic PV anomalies on the equatorward side and anticyclonic anomalies on the poleward side.

OLR was used as a proxy for deep convective heating in order to evaluate the total convective forcing of PV. The vertical distribution of heating was specified. In both regions the maximum total convective forcing of PV was largest on the equatorward edge of the sign reversal region. The effects of dry convection were not included in the PV budget. Dry convection, located poleward of the maximum deep convection, acts as a lower-tropospheric PV sink and produces anticyclonic PV anomalies. In both regions these anticyclonic anomalies were larger in magnitude and areal coverage than the cyclonic anomalies associated with deep convection.

The potential instability implied by the sign reversal regions has traditionally been associated with the growth of easterly waves. In support of this argument, bandpass-filtered (2–6 day) meridional wind variance on the 320-K surface nearly triples from east to west along the African sign reversal. In Australia, little evidence was found of such waves in the 2–10-day meridional wind variance. Possible explanations for the lack of growing disturbances over Australia are discussed.

Abstract

The structure and evolution of lowpass-filtered background flow and synoptic-scale easterly waves were examined during the 1991 eastern Pacific hurricane season. Active and inactive cyclogenesis periods conformed well to the sign of the near-equatorial, lowpass-filtered, 850-mb zonal wind anomaly, consistent with the recent results of Maloney and Hartmann. This behavior emphasizes the importance of westerly wind bursts associated with the Madden–Julian oscillation (MJO) in creating an environment favorable for eastern Pacific tropical cyclogenesis.

Synoptic-scale easterly waves reached the western Caribbean and eastern Pacific regularly from upstream, usually from Africa. The amplitude of waves leaving Africa had little correlation with the likelihood of a wave producing an eastern Pacific storm. Rather, easterly waves intensified, and tropical depressions formed, during the convectively active phase of the MJO in the western Caribbean and eastern Pacific. Wave growth, measured by strengthening of convection within the waves, occurred in the regions of sign reversal of the meridional potential vorticity gradient found previously. For the 1991 season cyclogenesis occurs when westward-moving synoptic-scale waves amplify within the superclusters that represent the favorable MJO envelope. Analogously, waves existed but failed to grow during the unfavorable part of the MJO.

During each active period of the MJO, the region of active convection moved eastward and northward with time in the eastern Pacific, with strongest convection reaching as far as the southwestern Gulf of Mexico by the end of such periods. The locations of tropical depression formation followed a similar path, shifting eastward with time following the MJO, and northward following the eastern Pacific intertropical convergence zone. The latter was defined by the locations of low-pass-filtered background vorticity maxima at 1000 mb.

It is argued based on previous work in the literature that the western Pacific might behave similarly, with upstream easterly waves growing and producing depressions within the convectively active envelope of the MJO.

Abstract

Helicity was calculated in Hurricane Bonnie (1998) using tropospheric-deep dropsonde soundings from the NASA Convection and Moisture Experiment. Large helicity existed downshear of the storm center with respect to the ambient vertical wind shear. It was associated with veering, semicircular hodographs created by strong, vortex-scale, radial-vertical flow induced by the shear. The most extreme values of helicity, among the largest ever reported in the literature, occurred in the vicinity of deep convective cells in the downshear-left quadrant. These cells reached as high as 17.5 km and displayed the temporal and spatial scales of supercells.

Convective available potential energy (CAPE) averaged 861 J kg⁻¹ downshear, but only about one-third as large upshear. The soundings nearest the deep cells were evaluated using two empirical supercell parameters that make use of CAPE, helicity, and/or shear. These parameters supported the possible existence of supercells as a consequence of the exceptional helicity combined with moderate but sufficient CAPE. Ambient vertical wind shear exceeded 12 m s⁻¹ for 30 h, yet the hurricane maintained 50 m s⁻¹ maximum winds. It is hypothesized that the long-lived convective cells enabled the storm to resist the negative impact of the shear.

Supercells in large-helicity, curved-hodograph environments appear to provide a useful conceptual model for intense convection in the hurricane core. Helicity calculations might also give some insight into the behavior of vortical hot towers, which share some characteristics with supercells.

Abstract

The surface wind field in a developing tropical cyclone (Agnes, 1972) was analyzed over a 1660 km radius for four days using conventional surface data, as the storm evolved from a disorganized depression to a hurricane. The transition to hurricane intensity was characterized by a wavelike disturbance propagating inward at 15 m s⁻¹ from the outermost radii to the storm core over a 36-hour period. This propagating disturbance was clearly visible in the radial and vertical motion fields as a surge of inflow and upward motion. Rapid intensification of the storm began within hours after the leading edge of the surge reached the storm center. The analysis of consecutive 12-hour periods without compositing of data from nonsynoptic times was essential for identification of this feature.

The surge had the same asymmetry as the upper-level outflow channel, indicating the possible involvement of the outflow layer in its initiation. No clear evidence of an external forcing mechanism for the surge, such as the passage of an easterly wave across the circulation, could be found. No instability theory could account for propagation of this feature across regions with such strongly varying dynamical properties. As a result, it remains uncertain whether the inflow surge represented an environmental trigger to hurricane formation or a manifestation of an internal instability.

The boundary layer momentum budget was dominated by Coriolis torque and frictional dissipation. The sum of these two terms acted as a momentum source primarily during the passage of the inflow surge across each radial region. Inward lateral flux of momentum contributed significantly only within 440 km of the center.

A distinct diurnal oscillation in pressure tendency occurred until hurricane strength was reached, with maximum deepening at 1200 local time, and minimum deepening at 0000 local time. Diurnal oscillations in other variables were more subtle and often at variance with those described in other tropical cyclones.

Because the inflow surge developed at outer radii 36 hours prior to rapid deepening and had a clear signature in the time change of radial mass flux, it provides a potential tool for forecasting tropical cyclogenesis 24 hours or more in advance which requires only the use of conventional data. More study is needed to determine whether such an early warning signal frequently occurs in intensifying tropical cyclones.