Tropical cyclones, also known as hurricanes or typhoons depending on their strength and location, are rotating atmospheric phenomena characterized by strong near-surface winds and torrential rain. Although tropical cyclones help to moderate climate by transferring energy from warm equatorial regions to cooler higher latitudes, the combined effects of their extreme wind, precipitation, and storm surge threaten the lives of millions of people who live near the coast.
In the face of that threat, meteorologists strive to predict the intensity and location of tropical cyclones days in advance. However, accurate prediction becomes challenging when the physical processes governing tropical cyclone development are not well understood. Atmospheric scientists have devoted decades of research trying to reveal the mysteries of tropical cyclone dynamics. It is now well-recognized that tropical cyclones extract heat from the ocean through sensible and latent heat fluxes. That’s why relatively warm sea-surface temperatures are a necessary, albeit insufficient, condition for tropical cyclone development.
Figure 1. Diagram depicting the main features of tropical cyclones as seen in a vertical cross section of height versus horizontal distance. CREDIT: COMET Program
The heat extracted from the ocean is transported radially inward toward the tropical cyclone center, where it ascends within the eyewall region. The eyewall is the location of the heaviest precipitation and strongest winds, as shown in figure 1. Large amounts of latent heat are released within the eyewall once the water vapor in the ascending air condenses. The net effect is a thermal expansion of the atmospheric column, leading to a pressure drop at the surface with an associated strengthened horizontal pressure gradient between the storm and the environment. Stronger near-surface winds result.
If no significant retarding effects, such as colder sea surface temperatures or friction, emerge to halt intensification, a continuous feedback arises between surface fluxes, latent heat release, and further pressure drops. A tropical cyclone will then continue to strengthen toward its maximum potential intensity.
Multiple factors can limit development and intensification of a tropical cyclone. One such factor is the presence of dry air in the vicinity of the storm. At the same given temperature, dry air is less buoyant than moist air and thus limits ascending motions. Dry air may also prevent ascending air parcels from reaching saturation, which decreases both net condensation and latent heat release.
Previous studies have demonstrated that, all other factors being equal, near saturation from the surface to a height of 5–7 km is a necessary condition for tropical cyclone development. Therefore, it is imperative to understand the sensitivity of developing tropical cyclones to different moisture profiles. A quantitative assessment of this issue, which is what we have carried out and describe here, could prove beneficial to gain new insights into tropical cyclone dynamics and prediction.
Cloud Model 1
Our study employed idealized numerical simulations to investigate the role of water vapor on tropical cyclone development. The principal ingredient of the simulations was the two-dimensional, height-versus-radius framework of George Bryan’s Cloud Model 1 (hereafter referred to as CM1). A hurricane-like vortex was prescribed in the numerical setup with ambient conditions characteristic of favorable tropical environments, such as relatively warm sea surface temperatures, nonzero planetary rotation, and the absence of environmental winds. The initial ambient conditions were specified via a vertical profile of temperature and relative humidity.
We conducted three experiments with different moisture profiles, but only two of them will be discussed here: a nearly dry (20% relative humidity) profile and a nearly saturated (80% relative humidity) profile. The initial vortex was the same across the experiments, and the only variation was the initial amount of environmental water vapor above 1-km height.
Once initialized, CM1 was integrated forward in time to produce seven-day simulations. The intensity evolution of the simulated tropical cyclones can be summarized with a time series of the maximum wind speed at 10-m height for each experiment, as in figure 2. A 12-hour running average was applied to remove noise and show the smoothed evolution of the intensity.
Figure 2. Time series of maximum 10-m wind speed for the experiments with 80% (red) and 20% (blue) relative humidity.
The results in figure 2 demonstrate that, within the modeling framework of CM1, tropical cyclone development takes longer for drier environments. That finding is consistent with previous modeling studies and observations. Interestingly, all experiments presented here have a period of nearly constant or slowly increasing maximum wind speeds before the period of greatest intensification is reached. We call those two stages of intensification the preconditioning stage and the rapid intensification stage.
As evidenced in figure 2, the length of the preconditioning stage is longer for drier initial relative humidity profiles. Because the only difference between the experiments was the initial moisture profile, it is possible that the duration of the preconditioning stage corresponds to the time needed to develop axisymmetric deep moist convection. The associated latent heat release, surface pressure drops, and stronger surface wind speeds likely set the stage for the rapid intensification later.
We used a water vapor budget approach to quantitatively investigate the duration differences of the preconditioning stage in the different experiments. The budget analysis is based on a simplified equation for the local time rate of change of the water vapor in the atmospheric column:
∂(ρqv)/∂t = −∇h ⋅ (ρqvvh) − ∂(ρqvw)/∂z + E − C,
where ρ is the dry air density, qv is the water vapor mixing ratio, vh is the horizontal wind velocity, w is the vertical wind speed, and E and C embody, respectively, evaporation and condensation rates within a given location.
The first term on the right hand side of the equation represents the horizontal flux convergence of water vapor. In the two-dimensional framework considered here, only the radial flux contributes to this term. The second term represents the vertical flux convergence of water vapor and quantifies the change in water vapor due to the vertical wind. The third and fourth terms are a source and a sink of water vapor, respectively.
We included only the first two right-hand terms by taking an area-average of thee terms within a radius of 102 km from the storm center. We excluded the other terms because our goal was to compare the exchange of water vapor between the tropical cyclone and its environment via horizontal and vertical fluxes.
Figure 3. Height-versus-time cross section of the area-averaged horizontal water vapor flux convergence (units: g/m3 per day) evaluated at a radius of 102 km for the experiments with initial (a) 80% and (b) 20% relative humidity. Positive values represent an increase in water vapor, whereas negative values represent a decrease in water vapor. Also shown is the maximum wind speed, with the corresponding scale on the right ordinate.
The results of the water vapor budgets are presented in height-versus-time plots that show the evolution of the area-averaged horizontal and vertical water vapor flux convergences (figures 3 and 4). Horizontal water vapor flux convergence at low levels begins early in both experiments (figure 3). The result is expected due to frictional convergence that induces radial inflow in the lower troposphere, which in turn transports moisture radially inward. The corresponding area-averaged vertical term shows water vapor flux divergence (figure 4).
The early onset of the horizontal water vapor convergence is also expected from a mass continuity perspective, since horizontal convergence in the lower troposphere must be accompanied by upward motion over the same layer. Moreover, the magnitudes of the horizontal convergence and vertical divergence are larger for the moist case that for the dry case over the first 72 hours of the simulation. Since the two cases were initialized with the same vortex, this result suggests that a moisture-rich environment is more effective at transporting water vapor radially inward and vertically upward, which is more conducive for convection and latent heat release.
Figure 4. Height-versus-time cross section of the area-averaged vertical water vapor flux convergence (units: g/m3 per day) evaluated at a radius of 102 km for the experiments with initial (a) 80% and (b) 20% relative humidity. Positive values represent an increase in water vapor, whereas negative values represent a decrease in water vapor. Also shown is the maximum wind speed, with the corresponding scale on the right ordinate.
Further analysis unveiled interesting features in the water vapor budget terms. Initially, all experiments showed that the horizontal convergence and vertical divergence were both confined to the lowest 1-km layer. There is a sudden increase in horizontal water vapor flux convergence above 1 km around 12 hours in the simulation with initial 80% relative humidity, and at later times in the other, drier simulation. This temporary feature is quickly replaced by horizontal convergence and vertical divergence in the lower troposphere.
We interpret this changing pattern as arising from pulsing updrafts and downdrafts. Updrafts are defined as vertically ascending air motions. Since water vapor is generally more abundant in the lower troposphere and the vertical motion is assumed to be zero at the surface, vertical water vapor flux divergence extending upward from the surface (as seen in figure 4 at 29 hours) signifies the upward transport of water vapor. Similarly, downdrafts are defined as the descending vertical air motions. Downdrafts are indicated by vertical water vapor flux convergence near the surface, which also leads to horizontal water vapor flux divergence over the same layer.
Although both the dry and moist cases show pulsing updrafts and downdrafts, the pulses appear several hours later and are confined to shallower layers in the initially drier environments (that is, 20% relative humidity). This result is evidence that pulsing convection is initiated once the boundary layer (that is, the lowermost 1-km layer) is moistened enough by horizontal water vapor flux convergence. At the same time, the pulsing convection transports moisture vertically upward, providing a favorable environment for more vigorous and sustained convection and latent heat release. The pulses cease at the end of the preconditioning stage, confirming that the duration of this stage is related to the development of persistent convection near the tropical cyclone center. In future work, this finding will be investigated further to elucidate more information about tropical cyclones under different moisture environments.
Although the idealized numerical model employed here does not completely simulate all of the physical processes in the real atmosphere, the study’s results provide evidence that tropical cyclones in drier environments are more likely to develop later than those in more moist environments. The delay in drier experiments is found to be a consequence of longer-lasting pulsing convection that moistens the layer above the boundary layer during the preconditioning stage.
Further investigation is needed—specifically, into the role of moisture in a fully three-dimensional model—but our preliminary results could still help forecasters determine the likelihood of tropical cyclone development based on moisture observations in the lower and middle troposphere. The information could be used to provide more accurate forecasts that are crucial to give coastal communities enough time to prepare for the harmful effects related to tropical cyclones.
This article is based upon a group project for a course on tropical cyclones given by Kristen Corbosiero at the University at Albany, State University of New York. The work is currently being expanded with the collaboration of Corbosiero and her University at Albany colleague, Brian Tang.
Rosimar Rios-Berrios is a PhD student at the University at Albany, State University of New York (SUNY-Albany), where she explores the dynamics of tropical cyclones using various modeling frameworks. Josh Alland is a PhD student at SUNY-Albany, where he explores the role of downdrafts on tropical cyclone development and intensity change. Jeremy Berman is an atmospheric science PhD student at SUNY-Albany, where he is currently evaluating the sensitivity of severe convective forecasts to model initial condition and upstream forecast errors.
The Week in Physics" is likely a reference to the regular updates or summaries of new physics research, such as those found in publications like Physics Today from AIP Publishing or on news aggregators like Phys.org.
