ON THE ROLE OF UPPER-TROPOSPHERIC POTENTIAL
VORTICITY ADVECTION IN TROPICAL CYCLONE FORMATION:
CASE STUDIES FROM 1991
by
DANIEL HUNT REILLY
B.S. Physics, University of Virginia (1988)
Submitted to the Department of Earth, Atmospheric, and Planetary
Sciences in Partial Fulfillment of the Requirements for the
Degree of
MASTER OF SCIENCE IN METEOROLOGY
at the
Massachusetts Institute of Technology
September 1992
Massachusetts Institute of Technology 1992
Signature of Author
Department of Earth, Atmospheric, and
Planetary Science May 26, 1992
Certified by
Professor Kerry A. Emanuel
Thesis Supervisor
Accepted by
Thomas H. Jordan, Head
Dept. of Earth, Atmospheric,
and Planetary Sciences
1
MA
Ri
UBRAIES
ON THE ROLE OF UPPER-TROPOSPHERIC POTENTIAL VORTICITY
ADVECTION IN TROPICAL CYCLONE FORMATION:
CASE STUDIES FROM 1991
by
DANIEL HUNT REILLY
Submitted to the Department of Earth, Atmospheric and Planetary
Sciences on May 26th, 1992 in partial fulfillment of the requirements for
the Degree of Master of Science in Meteorology
ABSTRACT
Several cases of western North Pacific tropical cyclogenesis from the
1991 season are studied, with special emphasis on the antecedent
conditions in the upper troposphere. Specifically, I test the hypothesis that
tropical cyclogenesis takes place through an interaction between a
pre-existing lower-tropospheric disturbance (ITCZ or easterly wave
disturbance) and an independent upper-level trough.
The upper-level
trough provides a source of upper-tropospheric lifting via differential
potential vorticity advection which may aid in the genesis process.
In this study, locations of pre-cyclone disturbances are obtained from
JTWC best-track information. A measure of potential vorticity advection
by the upper-tropospheric shear or the "forcing" of ascent is then
calculated from NMC gridded analyses. It is found that the majority of
tropical cyclones studied are subjected to periods of positive forcing during
their pre-genesis stages. In addition, positive forcing is implicated in the
generation of the pre-cyclone disturbances for the majority of cases. The
findings support the notion of tropical cyclogenesis as an externally
triggered phenomenon.
Thesis Supervisor: Kerry A. Emanuel
Title: Professor of Meteorology
4
ACKNOWLEDGMENTS
I would like to thank Kerry Emanuel for suggesting this topic to me,
and for his help along the way.
The conversations with Michael
Montgomery were stimulating and enjoyable. I also value the friendships
made with the other graduate students in the department. Most of all, I
thank my parents and sister for their love and support.
NMC gridded analyses and access software were obtained from the
Scientific Computing Division of NCAR. The efforts of Harry Edmon,
who provided analyses to fill the gaps for a handfull of missing dates, are
much appreciated. Best track information was provided on diskette by
Joint Typhoon Warning Center (JTWC) at Guam. Frank Wells of JTWC
was prompt and courteous in his responses to my inquiries.
This research was supported under NSF grant number ATM-9103025.
CONTENTS
Abstract
3
Acknowledgments
5
1. Introduction
7
2. Statement of the Problem of Genesis
2.1 Necessary Conditions for Genesis
8
2.2 Population of Tropical Disturbances
9
3. Review of Previous Work
3.1 Observational Studies of Genesis
11
3.2 Conceptual Models of Externally Forced Genesis
16
3.3 Conceptual Models of Genesis Through Internal
18
Processes, and Their Relationship to the External
Triggering Hypothesis
4. Relevant Properties of Ertel's Potential Vorticity
19
5. Data and Methods of Analysis
5.1 The Database for and the Preparation of the NMC
22
Analyses
5.2 JTWC Best Tracks and Estimated Intensities
23
5.3 Definition of the Forcing Parameter
25
6. Case Studies from the 1991 Season
27
7. Summary and Suggestions for Future Work
44
References
46
Figures
51
1.Introduction
Tropical cyclones are awesome, and potentially devastating storms.
They can form quite suddenly, with disasterous consequences for mariners
and coastal or island communities caught unaware. Thus it is not only of
scientific interest, but also of great practical importance to better
understand the birth or genesis of these storms, so that their formation
can be predicted. Current practice is more of a careful monitoring of cloud
clusters for signs of development into tropical cyclones, rather than an
objective forecasting of such development.
The scale of the tropical cyclone and of pre-cyclone disturbances is on
the order of hundreds of kilometers, which is not resolved by the standard
observing network over the tropical oceans. However, it is possible to
view characteristics of the larger scale environment within which these
disturbances are embedded. One hypothesis which can be tested is that
tropical cyclogenesis occurs when a pre-existing lower-tropospheric
disturbance comes under the influence of positive potential vorticity
advection from an independent upper-tropospheric anomaly brought into
proximity by the relative mean flow.
In this study, NMC global analyses, received on 2.5 x 2.5 degree
hemispheric grids, are used to examine all of the cases of genesis found to
occur in the western North Pacific region between July 1st and September
6th of the 1991 season.
Attention is given to the evolution of the
large-scale environment surrounding the pre-cyclone disturbances, with
particular emphasis on the possible role of upper-tropospheric potential
vorticity advection in the genesis of these storms. In this paper, a few basic
questions related to the problem of tropical cyclongensis are first stated.
Section 3 contains a review of selected works relevant to this topic. A
sketch of useful and relevant properties of Ertel's potential vorticity is
then made in section 4. A description of the data sets, and of the methods
of analysis follows in section 5. The results for several western North
Pacific cases are presented in section 6, with some discussion for each of
the 13 cases occurring during the nine-week period. A discussion of the
overall findings of this study, and suggestions for future work conclude
the paper.
2. Statement of the Problem of Genesis
2.1 Necessary Conditions for Genesis
Tropical cyclones are observed to form only when certain necessary,
but not sufficient, conditions are satisfied. If genesis is to occur, the
environment must be characterized by (Gray, 1979; McBride and Zehr,
1981):
(i) a sufficiently warm oceanic surface layer; empirically, an SST
lower bound is found to be around 26.5 degrees Celsius (Palmen,
1948);
(ii) a Coriolis parameter greater than some finite value (no geneses
occur within a few degrees of the equator);
(iii) sufficiently weak vertical wind shear directly over the center of
the potential developer; significant shear will inhibit the
establishment of a deep warm core, a characteristic structual feature of
a mature tropical cyclone;
(iv) relatively large values of low-level relative vorticity on the
synoptic and larger scales.
On a climatological or seasonal basis, middle-tropospheric relative
humidity and lower-to-middle tropospheric moist entropy differences are
also believed to be important.
However, comparisons of composite
developing versus nondeveloping cloud-dusters showed little difference
in the potential instability to moist convection, (McBride and Zehr, 1981).
In addition, tropical cyclones are observed to form out of pre-existing
initial disturbances of apparently independent origin (Riehl, 1954). One
objective of genesis-related research is to understand the processes
involved in the transformation of these initial disturbances into tropical
cyclones more completely, and perhaps to add to or refine the current set
of conditions which are known to favor this transformation.
2.2 Population of Tropical Disturbances
During tropical cyclone season, the tropical atmosphere contains large
numbers of candidates for genesis, which include waves in the easterly
trades, disturbances that form within or near zones of cyclonic shear such
as the inter-tropical convergence zone (ITCZ) or monsoon trough, and the
cloud clusters associated with them. Precise definitions of terms useful for
the discussion of genesis can be found in the appendix of the Joint
Typhoon Warning Center's (JTWC) Annual Tropical Cyclone Report
(JTWC, 1990):
(i)
Tropical Disturbance:
A discrete system of apparently organized
convection, generally 100 to 300 nm (185 to 555 km) in diameter,
originating in the tropics or subtropics, having a non-frontal,
migratory character and having maintained its identity for 12 to
24 hours.
It may or may not be associated with a detectable
perturbation of the wind field.
(ii) Tropical Cyclone: A non-frontal migratory low-pressure system,
usually of synoptic scale, originating over tropical or subtropical
waters, and having a definite organized circulation.
iii) Tropical Depression: A tropical cyclone with maximum sustained
surface wind speeds of 33 kt (17 m/s) or less;
(iv) Tropical Storm: A tropical cyclone with maximum sustained
surface wind speeds of between 34 and 63 kt (17 to 32 m/s);
(v) Typhoon (or Hurricane): A tropical cyclone with maximum
sustained surface wind speeds of greater than 64 kt (33 m/s).
Whereas, by the above definitions, tropical disturbances abound, tropical
cyclones are relatively rare. It has been estimated that only 1% or 2% of all
tropical disturbances (called hurricane seedlings in the reference) become
tropical cyclones (Simpson and Riehl, 1981, p.70). A 24-year average for
the Atlantic (1967-1990) indicates that out of an annual average of 58
African waves per season, 11.6 become depressions, and 5.9 become
tropical storms (Avila, 1991). Thus, it is a fundamental question of genesis
to discover the differences between pre-cyclone and non-developing
tropical disturbances, which may involve differences in the internal
structures of the disturbances themselves, and/or in the environments
within which they are embedded. One proposition of the present study is
that the developers are often selected by a chance interaction with an
independent upper-level trough. The suggested role of the upper-trough
is to provide lifting through potential vorticity advection by the shear in
the flow.
10
3. Review of Previous Work
3.1 Observational Studies of Genesis
Riehl in his early works (Riehl, 1948a, 1948b,1950, 1954) recognized
the existence of migratory cyclones and anticyclones in the tropical upper
troposphere which appeared to be moving relative to, and independent of
lower to middle-tropospheric disturbances such as easterly waves or ITCZ
disturbances. He also proposed the need for a "starting mechanism" for
the tropical cyclone "engine" which he argued must come from an
external energy source.
In the western North Pacific he found that the
initial deepening and formation of tropical cyclones often occurred with
the onset of upper-level height falls after the passage of an
upper-tropospheric anticyclone past a lower-tropospheric trough, with the
upper feature moving westward relative to the lower one (Riehl, 1948a).
The author draws an analogy between this scenario and that found in
middle-latitudes, the difference being in the direction of the vertical shear,
i.e. the upper-tropospheric features tend to move eastward relative to the
lower-level troughs.
Over the years, with the addition of commercial
aircraft and cloud-tracked wind observations to the available database,
Riehl and others were able to observe upper-level flow patterns in the
vicinities of genesis events more completely, and many cases were
documented.
In his more recent work (Simpson and Riehl, 1981; Riehl,
1979), it is emphasized that a commonly observed precursor to genesis or
intensification of tropical cyclones is the positioning of an upper
tropospheric cold-pool associated with an upper-level cyclone or trough
adjacent to a lower-level disturbance. The lower-level disturbance is then
observed to intensify as the upper-tropospheric cyclone diminishes in
intensity, or is destroyed completely. The disappearance of the upper-level
trough or cyclone ("collapsing cold-dome" in Riehl's terminology) was
interpreted as signaling the release of potential energy through the sinking
of cold air, thereby providing the energy necessary for triggering the
genesis of the storm.
Other
observational
evidence
for
the
importance
of
upper-tropospheric troughs in the geneses of tropical cyclones is found in
the work of
J. Sadler (Sadler 1975, 1976, 1978). He observed that genesis in
the western North Pacific often appeared related to southward and
westward extentions of what he termed the Tropical Upper Tropospheric
Trough, or TUTT. The TUTT is a WSW-ENE oriented tongue of cyclonic
vorticity, or anomalously large potential vorticity (PV), which is a
persistent feature found on the equatorward side of the subtropical ridge.
This channel of high-PV is observed to vary in its eastern and southern
extent, as it tends to be stretched out by the flow associated with the
anticyclone to the north. Occasionally, this zonally stretched trough will
fracture, and a cutoff anomaly of high-PV will emerge, manifesting itself
an upper tropospheric cyclone or "TUTT cell" in Sadler's terminology.
Examples of these events will be seen in the results presented later in this
paper. Based on aircraft and radiosonde information, as well as subjective
inferences from satellite imagery, a number of cases of tropical
cyclogenesis were observed to occur at the approach of such a TUTT cell
toward the site of some trade-wind of monsoon trough disturbance.
Often, the TUTT cell was found to position itself to the north and west of a
tropical disturbance, with genesis following. The author proposed that
such an upper-level flow configuration facilitates the creation of an
outflow channel to the north and east of the storm, in addition to
providing upper-tropospheric divergence above the incipient center. The
outflow jets were considered favorable and important for cyclone
12
development due to their ability to remove high entropy air from the
periphery of the cyclone core region, the subsidence of which would
destroy the thermal gradients near the core of the cyclone, and kill the
development (e.g. Sawyer, 1947). Sadler also implicated the TUTT cells in
the formation of the initial low-level disturbances themselves, and thus
questioned the independence of the upper and lower features in those
cases.
More recently, a study was made of the atmospheric conditions prior
to the genesis of Hurricane Diana which occurred off the southeastern
United States coast in 1984 (Bosart and Bartlo, 1991). The authors found
that an upper-level cyclone, associated with the fracture of an intense
trough in the westerlies, approached the genesis site in the hours prior to
the formation of the storm, providing forcing of ascent (estimated from
calculated advection of upper-tropospheric vorticity by the thermal wind)
in
the upper-troposphere
near
the region of genesis.
The
upper-tropospheric cyclone was observed to weaken as the surface cyclone
developed, and the authors noted the similarity to the collapsing
cold-dome scenarios described by Riehl.
Another observational study relevant to this problem was made by
Nitta and Takayabu, 1985, in which ECMWF level-mb gridded analyses for
the FGGE year were utilized. In that study, 850 mb relative vorticity fields
were filtered to pick out waves of periods between 5.7 and 10 days in the
western North Pacific. They found that about half of the vorticity maxima
in the filtered fields within the observed tropical cyclone generation
region eventually developed into typhoons, while the other half did not.
The most striking difference between the pre-typhoon and the
non-developing vortices was in the horizontal structure of the 200 mb
vorticity fields composited relative to the vortices. The upper-level flow
13
pattern surrounding the developing composite was characterized by the
Mid-Pacific Trough (MPT) (equivalent to Sadler's TUTT) being extended
westward to a position to the north of the developing vortices, while for
the nondevelopers this feature was positioned much further to the east
(see figure 1).
Another study of western North Pacific cyclogenesis was performed
by Zehr, 1991. He looked at all 50 cases of named-storm formation which
occurred during 1983 and 1984, as well as 25 persistent nondevelopers.
Data sources included digitized satellite radiance data in both IR and VIS
channels (with 10 km, and 3 hour resolution), as well as Air Force
reconnaissance,
and objectively analyzed conventional (AIREPS,
cloud-tracked winds, rawinsondes, ships, surface obs) data. From the IR
radiance data he calculated the percentage of pixels with brightness
temperatures below some specified value over an area within some
specified distance from the disturbance center, and plotted this parameter
versus time. He found that for 85% of pre-typhoon disturbances the time
series exhibited marked peaks during the pre-cyclone stage which were
termed "early convective maxima".
It was found that such an early
convective maximum preceded tropical storm status by 36-60 hours on
average, but occasionally by several days. This finding led Zehr to propose
a two-stage model of tropical cyclogenesis, in which the occurrence of such
an early convective maximum signifies stage 1 (see figure 2). After the
convective maximum, persistent curvature in the deep cumulonimbus
and an identifyable low-level circulation center first become apparent.
Aircraft reconnaissance also showed a reduction in the scale of the
circulation at this time, with little change in low-level wind speed, and
hence an increase in the low-level vorticity.
The percent-coverage of cold
cloud-tops (the time series parameter) then decreases temporarily after
14
stage 1, but increases again with the arrival of stage 2 which is
characterized by decreasing central pressure, and winds increasing to
tropical storm force,
i.e., by the birth and early intensification of the
tropical cyclone. It should be noted that the author found little difference
in the separation between the developing and non-developing clusters
and the axis of the TUTT, or TUTT cells, during stages 1 or 2, and
suggested on this basis that the TUTT might not be important in the
genesis process.
A number of sounding-composite studies of developing and
nondeveloping clusters have been performed, principally by the research
group at Colorado State University.
These composite studies involve
taking whatever soundings are available from each individual disturbance
or cyclone, and assigning them to a radius and sector relative to the
estimated center of that disturbance or cluster. It is then possible to study
quantitatively various properties of the composite disturbances. Care is
taken to stratify the cases according to different stages of development, or
to their ability to develop into tropical cyclones at all. One drawback of
this approach is that features not having a preferred radius or sector,
which may be important to genesis for each individual case, tend to be
smoothed out in the composite. The smoothing will also apply in time,
the degree of which is sensitive to the precision of the stratification of
cases by stage of development. In one such study (McBride and Zehr, 1981;
McBride, 1981a,b), developing and non-developing composite clusters
were studied and compared.
It was found that pre-cyclone and
non-developing clusters had similar vertical and horizontal sturctures in
fields of temperature and moisture, but that the pre-cyclone clusters were
embedded in environments characterized by much larger lower-to
middle-tropospheric relative vorticity and tangential wind fields out to
15
radii of 8 degrees latitude. The same study also showed zero 200 mb-850
mb vertical wind shear directly over, and large horizontal gradients in
vertical shear in an anticyclonic sense about the centers of the pre-cyclone
disturbances. In a similar study (Lee, 1986,1989) the perceived importance
of a pre-genesis build-up of lower and middle-tropospheric relative
vorticity on the large scales is emphasized by the author. This build-up
was attributed to the agent of lower tropospheric momentum surges with
their origins remote from the genesis site (Love, 1985).
3.2 Conceptual Models of Externally Forced Genesis
Attempts to quantify mechanisms by which upper-tropospheric
forcing can influence the development of tropical cyclones have been
made in the works of Challa and Pfeffer (Challa and Pfeffer, 1980, 1990;
Pfeffer and Challa, 1981, 1991).
They used the equations for an
axisymmetric, balanced vortex (Eliassen, 1951) to diagnose the effects of
angular momentum transport by asymmetries in the flow. These were
parameterized by means of an effective angular momentum source term,
in the vortex equations, which depends on the radial derivative of the
eddy angular momentum flux. The effect of this forcing on the secondary
circulation is given by the vertical derivative of this effective momentum
source. Vertical motions can arise from the adjustment of the temperature
field to the forced momentum field under the prescribed balance.
The authors found from the McBride composites that the Atlantic
developing composites differ from the nondeveloping composites in the
organization of the eddy fluxes of angular momentum in the
upper-troposphere. For the composite developers, the flux convergence
pattern in the upper-troposphere is such as to encourage ascent near the
center of the vortex, hence favoring an in-up-and-out circulation which
16
might aid in the genesis of the storm. A series of numerical experiments
were performed in which the composite fields served as initial conditions
for an axisymmetric hurricane model (Sundquist, 1970) and later a
3-dimensional mesoscale primitive equation model (Madala et. al., 1987).
For both models it was found that the initial conditions taken from the
nondeveloping duster composites did not produce hurricanes, while the
unaltered developing cluster composites did produce hurricanes in the
model integrations. When the asymmetric component and hence the flux
forcing was removed from the developing composite, the resultant initial
conditions failed to produce a cyclone, leading the authors to suggest that
proper organization of upper-tropospheric angular momentum fluxes
might be an additional necessary condition for the genesis of Atlantic
hurricanes. Such organization might be brought about by interactions
with external agents such as independent upper-level troughs.
Montgomery and Farrell (1991) draw an analogy between the forcing
of ascent via flux convergence in the Pfeffer-Challa model, and that by the
differential advection of potential vorticity in a 3-dimensional model of
the process. They employ moist, geostrophic momentum models to
investigate the response of a low-level seed field to the approach of an
upper-level trough (positive PV anomaly) brought into proximity by weak
upper-tropospheric shear. It was found that when the effects of moisture
and a near-saturated deep column were incorporated, reducing the
effective static stability, the encroaching upper-level PV anomaly acted to
induce a narrow tube of strong deep ascent, and resulted in a dramatic
spin-up of a surface cyclone. Also noted in this process was a collapse in
the scale of the low-level circulation as the upper anomaly approached.
This was contrasted with the corresponding dry experiment for which
ascent was confined to the upper-troposphere, and the upper and lower
17
anomalies remained essentially decoupled.
3.3 Conceptual Models of Genesis Through Internal Processes, and Their
Relationship to the External Triggering Hypothesis
Other models of tropical cyclogenesis have tended to focus more on
the dynamics of moist convection, and on air-sea heat exchange.
Convective Instability of the Second Kind (CISK) is one such model which
seeks to explain the growth of a depression-scale circulation in terms of a
cooperation between convective and depression scales. The concept is that
the two scales support one another, the circulation providing a favorable
environment for the convection, through boundary layer convergence
and Ekman pumping, and the convection, through latent heat release in a
conditionally unstable atmosphere, providing energy for intensifying the
vortex. In this way, a positive feedback is established, and the system
intensifies.
This concept led to numerous numerical experiments (for
review, see Anthes, 1982) in which initial vortices were placed in
conditionally unstable environments to see if hurricane-like systems
would develop. Often, successful simulations would require seemingly
unrealistic assumptions about the intensity and/or geometry of the initial
vortex, and the vertical structure of the heating. Vortex development has
also been simulated in a nonhydrostatic primitive equation model
characterized by an initially conditionally neutral environment (Rotunno
and Emanuel, 1987), where the suggested positive feedback is between the
surface wind speed and the induced surface fluxes from the ocean, and
doesn't implicitly require ambient conditional instability of the
atmosphere as does CISK.
This study also demonstrates the finite
amplitude nature of this instability, termed WISHE (Wind Induced
Surface Heat Exchange) instability, as only vortices of sufficient strength
18
and concentration would intensify. The finite amplitude nature of the
WISHE instability has been explored further in Emanuel, 1989. A finite
amplitude instability resembling CISK has been investigated in Handel,
1990. The question which then arises is what processes in the atmosphere
might lead to the formation of the initial finite amplitude vortex. An
external transient forcing might play a role in producing such an initial
vortex, which
might then intensify by fundamentally
internal
mechanisms such as CISK or WISHE.
It has also been suggested (Frank and Chen, 1991) that processes
involved in the formation of the initial surface vortex may be similar to
those involved in the formation of lower to middle tropospheric vortices
in mid-latitude MCCs (e.g. Menard and Fritcsh, 1989).
A source of
synoptic-scale forced ascent, such as that provided ahead of a shortwave
trough, and an associated active stratiform rain area have been suggested
to be necessary conditions for vortex formation in midlatitude MCCs
(Chen, 1990). Thus, there is a role for external forcing in the tropics in this
model as well.
4. Relevant Properties of Ertel's Potential Vorticity
Plots of Ertel's potential vorticity (PV), defined by expression (1)
below, are used extensively in the analyses of the genesis events presented
in this paper. Ertel's theorem states that the quantity
Q, defined
Q
4
is conserved under conditions for which (i) DX/Dt = 0, (ii) frictional
effects are unimportant, and (iii) X depends only on density and
pressure (i.e. is a state variable) (Pedlosky, 1987). For unsaturated flows in
the atmosphere, the logical choice for 2 is potential temperature,
19
represented by the symbol 0. For this choice, Ertel's theorem states that
(dry) potential vorticity is conserved following the motion for adiabatic
(DO/DT=O), frictionless conditions, which are reasonable approximations
for large scale, unsaturated flows outside of the boundary layer. Plots of
PV on surfaces of constant potential temperature (also a conserved
variable for frictionless, adiabatic flow) are useful for observing the
movement of parcels or patches of air. The atmosphere is observed to
contain various structures which have signatures in the PV fields; the
conservation of PV allows the movements and changes in shape of these
features to be followed easily. The effects of latent heating can also be
addressed within the PV framework.
Expressions (2) and (3) are
statements concerning the Lagrangian time rate of change of PV due to
heating and frictional effects, and the integral effects of the heating in the
interior of a material volume (Hoskins et al., 1985):
41
DQ/Dt = (1/p)
- grad(dO/dt) + (1/p) K - grad(8)
d/dt ( ijj pQ dt)
=
-
If ((dO/dt)4
+ OK)-n ds)
(2)
(3)
where K is the curl of the frictional force.
Concentrating on the heating terms, expression (2) states that generation
or destruction of PV following a parcel depends on the dot product of
absolute vorticity with the gradient in heating, divided by the density.
Expression (3) states that, in the absence of heating or friction on the
boundaries of a material volume, the mass-weighted PV integrated over
the volume does not change with time. In a tropical cyclone, one can
imagine constructing the boundary of a volume in a way such that all
20
latent heat release occurs within its interior. Then expression (2) implies
that parcels below the level of maximum heating will gain PV, while
parcels above will lose PV. Expression (3) offers an integral constraint on
these changes, that the mass-weighted losses of PV aloft are compensated
by generation in the lower layers.
Such destruction of PV near the
tropopause in the vicinity of intensifying cyclones will be seen in many of
the results presented later.
Another aspect of PV which is worth noting is that PV anomalies in
quasi-balanced flow are characterized by vorticity and static stability
anomalies in the same sense, i.e. positive PV anomalies are characterized
by cyclonic circulations and relatively high static stabilities, and vice-versa.
The circulations associated with the anomalies will penetrate some
distance below the anomaly with a penetration depth dependent on the
scale of the anomaly, on the effective static stability in the column below,
and on the vorticity itself. Vertical shear in the horizontal wind will tend
to induce upward motion on the downshear side of the anomaly.
To
understand this result physically, one might imagine travelling with a
lower-level air parcel which is moving toward an upper-tropospheric
positive PV anomaly located upshear. The parcel will be required to
acquire vorticity consistent with the circulation from the anomaly above
as it approaches the region beneath the anomaly. This implies vortex
stretching from the time it enters the region of influence of the upper
anomaly's circulation until its closest approach, and then contraction as it
moves away. Translating this picture to an Eulerian frame, it follows that
upper-tropospheric ascent is expected where there is positive potential
vorticity advection by the relative velocity between the anomaly and the
parcels below. This is the physical basis for the forcing parameter defined
in the section 5.3.
21
5. Data and Methods of Analysis
5.1 The Database for and the Preparation of the NMC Analyses
The upper-air database used in the preparation of NMC analyses
consists of radiosondes, pilot balloons, aircraft reports, cloud-tracked
winds, and remotely-sensed temperature soundings. Figure 3 shows the
distribution of these elements over the western North Pacific for a "typical
day" (Dey, 1989).
Aircraft reports generally have their best coverage in the
upper troposphere and lower stratosphere, near the flight levels of
commercial aircraft, which have their highest density near 250 mb.
Cloud-tracked winds are derived from the identification and tracking of
clouds in time loops of images obtained from geosynchronous-orbitting
satellites. Crudely speaking, levels are assigned to the observations by
matching the measured IR brightness temperature of the cloud to the
appropriate level using some estimate of the environmental vertical
temperature profile, e.g. that obtained through remotely-sensed
temperature soundings.
Generally, low-level cumulus tracers provide
velocity measurements representative of the flow at 850 mb, while the
tracking of upper-tropospheric cirrus targets provide estimates of winds
near
the
tropopause
(Elsberry,
1987).
The
circulations
of
upper-tropospheric cold-lows with associated cirrus rings about their
perimeter can occasionally be well resolved by this data source
(Shimamura, 1981).
The remotely-sensed temperature soundings are
retrieved from measurements made by passive radiometers mounted on
polar-orbitting satellites.
Upwelling radiation is measured in various
frequency bands, each of which has a characteristic emissivity vertical
profile in the atmosphere.
Given the dependence of the emissions of
radiation on temperature, and given the emissivity profiles and observed
22
radiances for various bands, temperatures are determined for various
levels, and a sounding is retrieved. In practice, the emissivity profiles or
weighting functions overlap, and the sorting out of information to obtain
a vertical sounding is done by sophisticated retrieval methods (Smith,
1991). Temperature gradients are relatively weak in the lower to middle
tropical troposphere, and the remotely-sensed temperature soundings are
thus of little use there. However, this technique might be useful near the
tropopause where temperature anomalies are more significant.
For
example, temperature anomalies of +8 and -5 degrees Celsius have been
observed above and below upper tropospheric cold lows near 20 degrees
north latitude (Shimamura, 1982).
The raw observations are used to determine corrections, via
optimum interpolation techniques, to background or first-guess fields
which are provided by a 6-hour forecast from the NMC global spectral
model (Kanamitsu, 1989; Kanamitsu et al., 1991) which had been
initialized by the previous analysis. Obviously, in the absence of data,
these first-guess fields will dominate the analyses. In the results which
follow, attention is restricted to the 850 mb level and the 360K isentropic
level (near the tropopause) where the density of observations is greatest.
Presumably, the NMC analyses will accurately portray features which are
well resolved by information in the database.
The numerical model
should then be capable of maintaining these features for some time after
adequate resolution by observations has been lost.
5.2 TTWC Best Tracks and Estimated Intensities
The positions and intensities of the cyclones in this study are taken
from Joint Typhoon Warning Center (JTWC) post-season analyses for the
1991 season. This information includes positions of the pre-depression
23
disturbances when they were able to be determined by the JTWC analysts.
Position and intensity (maximum sustained wind, 1 minute average)
estimates both rely heavily on interpretation of satellite imagery. The
Dvorak technique (Dvorak, 1984) is chiefly used in the intensity estimates
(JTWC, 1990). The rules which characterize this technique are based on
the observation that, apart from diurnal and other short term fluctuations,
the patterns of deep convective clouds often follow a systematic
progression during tropical cyclone intensification. The primary indicator
of intensity for an immature (no eye) cyclone is the degree of coiling of
curved deep convective bands around the system center (see figure 4). For
a more mature cyclone, IR measurements of "eye-temperature" and
cloud-top brightness temperature for the surrounding inner core region
provide more objective measures of intensity (Dvorak, 1984). According
to the rules of this technique, the initial development of a tropical cyclone
is defined by the first appearance of curvature in a band of deep convective
clouds together with an area of dense overcast of extent greater that 1.5
degrees latitude found less than 2 degrees latitude from the estimated
system center. If this is observed, the disturbance is tagged "T1", and
assigned an intensity of 25 kts (for table relating intensity to T-number, see
figure 4). From that point, the T-number is expected to increase on
average 1 per day, with the disturbance anticipated to reach tropical storm
strength in 36 hours. The intensity progression expected by this Dvorak
timetable is to be used as a guideline for forecasters and analysts, but is
modified if the cloud-patterns clearly indicate a different intensity.
In the case studies which follow later, the JTWC estimated intensities
are plotted versus time. From these time series, it is possible to identify
various transitions in the intensification rates of the storms. Figure 6
shows a typical example. Often the time series is initially flat, with
24
intensities of magnitude less than or equal to 25 kts, and increasing at a
rate of 5 knots per day or less. During this stage, one can infer in the
context of the Dvorak technique that the disturbance is lacking signs of an
organized circulation (i.e. band curvature).
At some point a clear
transition takes place, and the system undergoes steady intensification of
10 to 25 knots or about 1 T-number per day from depression to tropical
storm and perhaps to typhoon stage. Occasionally, periods of more rapid
intensification are found, perhaps exceeding 30 knots per day for a limited
time interval. Of particular interest in this study are the environmental
conditions, and upper-tropospheric forcing during the days leading up to
the transition from the very-slowly intensifying, poorly organized tropical
disturbance, to the steadily intensifying well-organized tropical cyclone.
Such a transition will be referred to as genesis in this paper. Also of
interest are conditions at the time and place of the generation of the
tropical disturbance itself, defined by the first entry for the storm in the
JTWC best track. The JTWC positions are also used in the calculation of
the forcing parameter to be defined below, the physical significance of
which was described in the previous section. On potential vorticity and
850 mb vorticity maps in the results presented later, positions of JTWC
analyzed pre-cyclone disturbances or cyclones are indicated by triangles,
with the track of the disturbance and cyclone indicated by small squares.
5.3 Definition of the Forcing Parameter
In this work, we are especially interested in determining the degree to
which pre-genesis disturbances are fostered in environments characterized
by ascent, externally forced by migrating upper-tropospheric troughs or
cyclones. To quantify this effect, the forcing parameter defined below is
used:
25
F =
-
1/N
T
Y (Vijk - V40j) - grad( Q
k
i,j
)
where N is the total number of terms in the summation
Vjk is the analysis velocity at isentropic level k, horizontal grid points ij,
-_4
V40j is the 400mb horizontal wind velocity for grid points ij,
and Q
is the value of Ertel's PV at grid point ij, level k.
The summation in k is made over 5 levels of potential temperature, from
350K to 370K at 5 degree intervals. All grid points (ij) within a specified
radius from the JTWC-analyzed disturbance center are included in the
average. This radius is varied between 2 and 7 degrees latitude. The
forcing parameter essentially measures an average potential vorticity
advection near the tropopause as viewed in the frame of a parcel at 400mb.
Positive values should correspond to forced ascent, negative values to
descent. Time series of this forcing parameter are readily compared with
those of intensity. In this work, a value of the forcing parameter prior to
generation of the initial pre-cyclone disturbance is estimated by
extrapolating the first entry in the JTWC best track backward 6 or 12 hours.
Thus the forcing time series will begin 6 to 12 hours before the
corresponding intensity plot.
A second forcing parameter was also
computed, with 850 mb replacing 400 mb as the lower level in the
calculation. The results tended to be similar during the early stages of
development, with a few exceptions which will be noted in the discussion
of the results. In the time series presented later, the forcing is shown in
units of 10-11 m 2 K kg-1 s-2, or 0.864 PV units (Hoskins et. al., 1985) per day.
26
6. Case Studies from the 1991 Tropical Cyclone Season
In the following, results from every case of tropical cyclogenesis
which occurred in the western North Pacific between the dates of July 1st
and September 6th of 1991, are discussed. This 9 week period saw the
formation of 8 typhoons, 3 tropical storms, and 2 unnamed depressions.
Emphasized in the discussion of the results is the time evolution of
upper-tropospheric flow patterns and PV-structures for times leading up
to the geneses of the tropical cyclones. The results for each case are
presented in the following format. First, the tracks of each cyclone are
shown, which are useful for determining intensity changes which may be
related to a change in the lower boundary, i.e. passage of the disturbance
over land or colder sea-surface temperatures. Next, the time series of
JTWC estimated intensity for the cyclone is shown, from which phases in
the development, characterized by markedly different average
intensification rates, are identified. An attempt is made to identify the
time of transition from tropical disturbance to intensifying cyclone, i.e. of
genesis, on the basis of these time series. The time series for the forcing
parameter for the event are then discussed, with episodes of "positive
forcing" and implied forced ascent noted, and their relationship (if any) to
the genesis of the tropical cyclone, and to the generation of the precursor
tropical disturbance. Next, potential vorticity fields are shown on the 360K
isentropic surface, which indicate the upper-level flow patterns near the
time of genesis, and the structures responsible for episodes of positive
forcing. For most cases, maps of 850 mb winds and relative vorticities are
presented to portray the evolution of the lower-tropospheric large-scale
flow environment over the days leading up to genesis.
The month of July saw the formation of 4 tropical cyclones, Zeke,
Amy, Brendan, and Caitlin, which each attained typhoon status. The
27
tracks of these storms, including their pre-cyclone disturbance stage, are
shown in figure 5. The geneses of these storms occurred within a 2 week
period, which was followed by 2 weeks of inactivity in which no tropical
cyclones formed. From figure 5 note the similarity in the tracks of the
storms, the exception being the dramatic turn of Caitlin toward the north
on the 24th of July. Such a clustering of events in space and time has been
noted by others (Gray, 1979). Each of these storms will next be discussed
individually, in order of their appearance.
Typhoon Zeke
The intensity time series for tropical cyclone Zeke is shown in figure
6. The following phases, characterized by nearly uniform intensification
rates, are identified:
Phase
Time Inverval
1
07/06/OOZ-07/09/18Z
2
07/09/18Z-07/10/18Z
3
07/10/18Z-07/12/18Z
Description
Ave. Intensification
Rate (kts/day)
pre-cyclone
disturbance
steadily intensifying
incipient cyclone
rapidly intensifying
cyclone
3
10
27
The forcing parameter time series for this case (figure 6) shows an
extended period of positive forcing, and induced upper-tropospheric
ascent occurring roughly between OOZ of the 6th and OOZ of the 9th, during
the pre-cyclone stage of development.
The actual transition between
pre-cyclone disturbance and incipient cyclone, which we are calling
genesis, occurs by 18Z of the 9th, a time when the forcing is actually
reduced to near zero. In figure 7, the 360K PV-fields are presented every 24
hours for the days leading up to the generation of the pre-cyclone
disturbance on OOZ of the 6th. Of particular note is the westward and
28
equatorward advance of the upper-level trough extending from 165E, 17N
southwestward to around 145E, 7N at 12Z of the 5th, as well as the region
of negative PV centered around 146E, 3N to the south of the advancing
trough. The pre-cyclone disturbance associated with Zeke can initially be
found just west of this feature, as shown in figure 8. Over the next 48
hours the disturbance is approached by another trough, or extension of the
TUTT, which is seen by following the advance of the 3 PVU contour in
figure 8, to position itself to the northeast (upshear) of the tropical
disturbance by 12Z of the 7th. Figure 9 shows the subsequent movement
of the
pre-cyclone
disturbance
away from
the
TUTT,
with
nonconservative processes, i.e. destruction of PV aloft, exhibited between
12Z of the 8th and 9th.
The evolution of the large-scale flow at 850 mb can be seen in the
relative vorticity and wind fields shown in figures 10 and 11. Note the
intensification of the equatorial trough or ITCZ over the entire 40 degree
longitude span of the domain in the days prior to the generation of the
pre-cyclone disturbance (figure 10), with an elliptical large-scale circulation
between 130E and 145E, 5N and 1ON emerging by 12Z of the 5th. By 12Z of
the 9th the pattern has evolved to a more circular shape, with genesis, or
the transition to a steadily intensifying vortex, occurring around this time.
In summary, the generation of the pre-cyclone disturbance related to
Zeke was preceded by an intensification of the monsoon trough at 850 mb,
and the approach of an upper tropospheric trough from the northeast.
Once formed, the tropical disturbance remained in an environment of
positive forcing up until, but not including the transition from phase 1 to
phase 2, i.e. the genesis for this case. Also, the disturbance's proximity to
the Phillipines during the latter part of phase 1 (figure 5) may have
inhibited the onset of intensification (phase 2) until the passage of the
29
islands.
Typhoon Amy
The intensity time series associated with Amy (figure 12) suggests the
following divisions:
Phase
1
Time Inverval
Description
Ave. Intensification
Rate (kts/day)
07/12/18Z-07/15/12Z
pre-cyclone
3
disturbance
23
intensifying vortex,
2
07/15/12Z-07/17/06Z
incipient cyclone
3
07/17/06Z-07/18/12Z
rapidly intensifying
48
cyclone
The forcing parameter time series (figure 12) again indicates a bias toward
positive forcing during the pre-cyclone stage (phase 1) and through phase 2
as well, with maximum forcing occurring between OOZ of the 15th and OOZ
of the 17th. The forcing reduces to near zero during and after the rapid
intensification stage of the cyclone. The corresponding PV maps are
shown in figures 13 through 15. Note the evolution of the TUTT over the
days preceding the generation of the pre-cyclone disturbance (figure 13).
The TUTT appears extended from 165E, 24N to 135E, 5N as of 12Z on the
9th, and over the subsequent days, is seen to fracture, with an intense
upper-level cyclone remaining in the eastern portion, centered at 156E,
19N on 12Z of the 12th. The initial disturbance associated with Amy is
first detected by JTWC by 18Z of the 12th at a postion southwest of this
upper-tropospheric cyclone, or TUTT cell. Over subsequent days the two
features migrate westward maintaining approximately their original
separation (figure 14), the upper-cyclone tending to weaken and deform.
The period of maximum forcing centered around 12Z of the 15th is
characterized by a strengthening northeast flow over the pre-cyclone
disturbance down the axis of the upper-trough postioned to the northeast
30
(figure 18). By 12Z on the 16th, the upper-level forcing is bolstered by the
arrival of another trough from the northwest.
The flow at 850 mb is documented in figures 15 and 22, which show
the development and genesis of Amy taking place within a
poleward-displaced portion of the equatorial trough. Note the track of
Amy along the monsoon trough, which is intensifying on the large scale
by 12Z of the 16th.
In summary, the generation of the pre-cyclone disturbance related to
Amy is preceded by the formation and positioning of an intense
upper-tropospheric cyclone to the northeast. The forcing at this stage is
most impressive in the time series of the forcing parameter calculated
using the 850 mb winds for the lower level (figure 16). Once formed, the
tropical disturbance remains in a generally positive forcing regime
throughout the genesis and early intensification stages of development.
Typhoon Brendan
The intensity versus time graph for this storm (figure 17) suggests the
division of the formation and intensification of this cyclone into the
following 3 phases:
Ave. Intensification
Phase
Time Inverval
Description
Rate (kts/day)
1
07/15/OOZ-07/18/06Z
2
07/18/06Z-07/20/18Z
3
07/20/18Z-07/22/OOZ
pre-cyclone
disturbance
slow intensification,
genesis
rapidly intensifying
cyclone
2
6
27
The time series of the forcing parameter shows positive and increasing
values during phase 1, with a substantial peak at 12Z of the 18th. A lull in
the forcing ensues, until the onset of a second episode by OOZ of the 21st
31
which is correlated with the beginning of the rapid intensification phase.
The corresponding PV maps are shown in figures 18 through 21. The
generation of the pre-cyclone disturbance occurs to the southeast of an
upper-tropospheric trough similar to the case of Zeke, although the
forcing is apparently negligible at this time. Figure 19 shows the first
encounter of the pre-cyclone disturbance with an upper-tropospheric
trough on 12Z of the 18th as a tongue of high-PV is drawn equatorward by
northerly and northeasterly flow between OOZ of the 17th and 12Z of the
18th. At the same time, another PV anomaly is situated well to the north
centered along 30N between 125 and 140E. Over the next 48 hours (see
figure 20) this feature is observed to be advected southward and westward
on a collision course with the lower-level disturbance. The upper-level
feature is seen to encounter the low-level disturbance by OOZ of the 21st
(fig. 21) thus beginning another extended period of positive forcing. The
rapid intensification phase of the development begins at this time. Note
that the abrupt halt to intensification at OOZ of the 22nd is likely due to the
landfall made on OOZ of the 22nd (see figure 6) at the northern extreme of
the Phillipine Islands. Figures 22, 26, and 27 show the large-scale flow at
850 mb for this period.
The pre-cyclone disturbance appears to be
associated with a migrating circulation identifyable in the NMC analyses
as a vorticity maximum at least 48 hours prior to the JTWC's
identification of the pre-cyclone disturbance (figure 22).
In summary, the upper-tropospheric flow pattern around the
pre-cyclone disturbance at its generation resembles that for the case of Zeke
with a trough and region of negative PV positioned to the northeast and
east respectively. A negligible forcing parameter at this time, however,
makes a link between positive forcing and generation of the disturbance
questionable. The pre-cyclone disturbance encounters upper-level troughs
32
on 12Z of the 18th, and beginning on OOZ of the 21st, which correspond to
transitions in the intensification rates of the disturbance/incipient cyclone.
The second episode, in particular, appears to be implicated in the genesis
and early intensification of cyclone Brendan.
Typhoon Caitlin
The intensity time series for typhoon Caitlin (figure 23) shows 2
distinct phases in the development:
Phase
Time Inverval
Description
1
07/18/12Z-07/23/18Z
2
07/23/18Z-07/27/12Z
Ave. Intensification
Rate (kts/day)
pre-cyclone
disturbance
intensifying
tropical cyclone
3
19
The forcing appears positive but relatively weak for much of the
pre-cyclone phase with peaks around 12Z of the 18th and OOZ of the 22nd.
The early peak may suggest a role of positive forcing in the generation of
the initial disturbance.
The episode of positive forcing around 12Z of the
25th occurs well into phase 2 by which time the cyclone has attained an
intensity of 65 knots, clearly after the genesis of the storm.
Potential vorticity maps (figures 24 and 25) do show an upper-level
trough positioned to the northwest of the low-level disturbance during
much of phase 1, with the lower-level disturbance apparently overtaking
the upper-level trough by OOZ 23rd, but this event is not reflected strongly
in the forcing time series. For this case, on the basis of the quantitative
measure, we must conclude that a link between upper-tropospheric
potential vorticity advection and tropical cyclogenesis is not readily
apparent. It is worthy of note that the transition from stage 1 to 2 is also
accompanied by the marked change in the track of the storm which is
33
observed to occur around 6Z of the 24th (figure 5). It is not clear how or if
these events are related. At 850 mb (figures 26 and 27) the pre-cyclone
disturbance is seen to form in the vicinity of a vorticity maximum
migrating along the equatorial trough, similar to the case of Brendan.
Subsequent panels (figure 27) show increasing vorticity about the
disturbance on the large scale, prior to the genesis of the cyclone arount
July 24th.
Two weeks of inactivity ended with the genesis of Doug around OOZ
of August 9th. This event marked the start of another active period. The
next 10 days saw the geneses of 5 tropical cyclones. Their tracks are shown
in figures 28 and 35. While the track of Fred was similar to that of the first
4 typhoons, the other cyclones tended to form about 10 degrees latitude
further north. The clustering of events in time and space is again noted.
The cyclones within this second group will now be discussed individually
in order of their appearance.
Tropical Storm Doug
Doug achieved maximum sustained winds of only 35 knots, minimal
intensity for a tropical storm. The storm tracked quickly northward (figure
28) toward higher latitude and colder ocean, failing to intensify further.
The intensity and forcing time series are shown in figure 29.
The
transition from tropical disturbance to weak tropical cyclone (genesis) is
seen to occur around OOZ of the 9th. Forcings at this time and before are
primarily positive (especially for 7 degree averaging radius) but weak, with
no clear relation between genesis and positive forcing for this event.
Potenital vorticity maps are shown in figures 30 and 31, with relevant 850
mb maps displayed in figure 32. The PV maps for 12Z of the 8th, just prior
to genesis, do indicate high PV to the north and east of the disturbance, but
34
the forcing is apparently weak. At 850 mb (figure 32), the pre-cyclone
disturbances related to Doug and Ellie are observed to form within a
pre-existing region of high relative vorticity apparently associated with a
frontal zone intrusion from mid-latitiudes days before.
In summary, there is no clear indication of upper-level forcing of
either the genesis or the generation of the initial disturbance for this case.
Typhoon Ellie
The time series pertaining to Typhoon Ellie are shown in figure 33.
The intensity time series is unusual in that there is no clear pre-cyclone
disturbance phase, with a fairly rapid elevation to tropical storm status
within the first 30 hours of JTWC detection as an entity. The following
phases are defined:
Phase
Time Inverval
Description
Ave. Intensification
Rate (kts/day)
1
08/09/00Z-08/10/06Z
genesis
16
2
08/10/06Z-08/11/18Z
slow intensification
10
3
08/11/18Z-08/12/18Z
steady
4
08/12/18Z-08/14/18Z
moderate
intensification
0
20
The time series of forcing shows near zero PV advection by the shear
during the early stages of development and genesis of this storm. Thus,
there is no apparent relationship between upper-tropospheric forcing and
the generation of the pre-cyclone disturbance, or genesis of the tropical
cyclone for this case. The event around 12Z of the 13th appears to be
related to the approach of the tropical cyclone to a trough in the westerlies.
Note the enhanced outflow to the north, and the PV destruction aloft
35
exhibited by 00Z of the 13th (see figure 34).
Typhoon Fred
The intensity time series for Fred (figure 36) suggests that the
development of the cyclone be divided into 3 stages:
Ave. Intensification
Phase
Time Inverval
Description
Rate (kts/day)
1
08/08/OOZ-08/12/18Z
pre-cyclone
2
disturbance
2
08/12/18Z-0/15/12Z
intensifying
18
cyclone
3
08/15/12Z-08/16/OOZ
brief period of
40
rapid intensification
The time series of the forcing parameter (figure 36) shows impulses of
positive forcing at the onsets of both phases 1 and 2 (around OOZ of the 8th
and 13th respectively), with another more extended period of forcing
during the pre-cyclone disturbance phase, peaking at 12Z of the 11th. The
first impulse is implicated in the generation of the initial disturbance,
while the second and third episodes are implicated in the transition from
phase 1 to phase 2, i.e. in the genesis of the tropical cyclone. After OOZ of
the 13th, the forcing parameter is seen to hover around zero. It should be
noted that the brief landfall of Fred at the Phillipines during the 12th of
August may have also played a role in the timing of the genesis (figure 35).
The 360K PV maps show the structures responsible for the episodes of
positive forcing, and implied upper-tropospheric ascent. Figure 37
portrays the events surrounding the generation of the pre-cyclone
disturbance which occurs by OOZ of the 8th. Note the upper-tropospheric
positive PV anomaly centered around 147.5E, 7N as of 12Z on the 7th,
moving westward with the strong easterlies of about 50 kts, which
converts to about 10 degrees longitude per 12 hours, consistent with the
36
observed displacement between 12Z of the 7th and OOZ of the 8th.
Generation of the pre-cyclone disturbance is seen to occur just downshear
of this anomaly on OOZ of the 8th. Figures 38 and 39 show the events
leading to the genesis of the cyclone.
As of 12Z on the 9th, an
upper-tropospheric anomaly is positioned near 23N, 147.5E, and is being
advected west-southwestward.
By 12Z on the 11th, the anomaly is
positioned to the northeast of the pre-cyclone disturbance.
This
configuration remains fixed to some extent for the next few days as the
two features drift toward the west, with the flow configuration
corresponding to the peak in the forcing at OOZ on the 13th shown in
figure 39. Also noteworthy is the approach of the intensifying cyclone
toward an intense but decaying upper-level trough, whose axis is along
110E as of OOZ on the 14th. The lower-left panel of figure 39 shows the
situation just prior to the onset of phase 3, a period of rapid intensification
just prior to landfall.
The configuration appears conducive to the
establishment of an outflow channel to the north (Sadler, 1978). Also, the
demise of the upper-trough is reminiscent of the collapsing-cold-dome
scenarios described by Riehl (Riehl, 1979).
At 850 mb (figures 40 and 41) an intensification of the equatorial
westerlies and the monsoon trough is observed to precede and accompany
the genesis of the storm.
In summary, the pre-cyclone disturbance associated with Fred was
generated downshear of a rapidly moving upper-tropospheric PV
anomaly. The tropical disturbance then experienced positive forcing, from
another anomaly situated to the northeast, with a peak in the forcing at
the onset of steady intensification, or the genesis of the tropical cyclone.
The brief landfall of the disturbance on 12th of August may have also
influenced the timing of genesis.
37
Tropical Depression 13W
This disturbance was first detected on 12Z of August 11th, reaching its
maximum intensity of 25 kts by 6Z of the 12th. The forcing time series for
this disturbance is shown in figure 42, with positive forcing apparent at
the generation of the disturbance, and zero or negative forcing during its
lifetime as a depression.
The relevant PV maps (figure 43) show the
generation of the pre-cyclone disturbance occurring downstream of an
upper-level trough situated just to the east. Tropical cyclone Ellie can be
seen situated to the northwest. The negative forcing appears to result
from the relative movement of the depression toward the low-PV air in
the upper-troposphere left in the wake of cyclone Ellie. At 850 mb, the
disturbance is seen to develop on the cyclonic shear side of a belt of 25 kt
southeast winds (figure 44).
Note also the motion of TD 13W around
cyclone Ellie.
In summary, positive forcing is clearly indicated in the generation of
the initial tropical disturbance for this case. Negative forcing prevailed
during the depression stage, with the system failing to intensify beyond 25
knots.
Typhoon Gladys
The early development of Gladys, a minimal strength typhoon at its
most intense level, is divided into 4 stages (figure 45):
Phase
Time Inverval
Description
Ave. Intensification
Rate (kts/day)
1
08/13/12Z-08/16/OOZ
2
08/16/OOZ-08/18/OOZ
3
08/18/OOZ-08/20/OOZ
pre-cyclone
disturbance
intensifying
cyclone
steady at 55 knots
4
08/20/OOZ-08/21/06Z
renewed intensification
38
6
15
0
8
The time series of the forcing parameter shows an episode of positive
forcing during the first 24 hours of the JTWC best track, followed by lesser
events (peaks in the time series) over the following few days. The
relevant PV maps (figures 46 and 47) show that during its generation and
early lifetime, the pre-cyclone disturbance resides in an environment of
upper-tropospheric northerly flow, with a broad region of high-PV
situated to the north and east, and with structures within this region
occasionally rotating down to positions north of the disturbance. The
corresponding 850 mb maps (figures 48 and 49) reveal a complicated flow
pattern, including a large-scale cyclonic circulation on the order of 30
degrees latitude in diameter centered around 150E, 20N on OOZ of the 13th.
Pre-cyclone Gladys forms on the southeast extreme of this circulation,
where winds are strong and southerly, implying relatively large vertical
shear between the lower and upper troposphere during the early stages.
Note also the increase in the voriticity about the intensifying disturbance
on the large scale. By 12Z of the 16th, a large-scale circulation surrounding
now tropical depression Gladys dominates the domain.
For this case, positive forcing is implicated in the generation of the
pre-cyclone disturbance and, to a lesser degree, in the genesis of the
tropical cyclone. An increase in the low-level vorticity on the large scale
occurred during the tropical disturbance stage, preceding the genesis of the
tropical cyclone.
Tropical Depression 15W
The tracks of the next two cyclones of the season, TD 15W and TS
Harry, are shown in figure 50. The intensity and forcing time series
pertaining to TD 15W are shown in figure 51. The generation of the
39
disturbance is accompanied by an impulse of positive forcing, while its
later intensification from 20 kts (pre-cyclone disturbance) to 30 kts
(depression) around 12Z of the 25th is also accompanied by the onset of
positive forcing.
The question yet to be answered for this case is why did
it not intensify further. The depression skirts the southern tip of Japan by
00Z of the 28th.
The associated PV maps are shown in figures 52 and 53. The forcing
agent at the generation of the disturbance appears to be an eastward
extension of the TUTT (figure 52). At 850 mb (figure 54), the disturbance
appears to form within a large-scale, zonally elongated trough in
subtropical latitudes.
Note the pervasiveness of lower-tropospheric
westerlies in the tropics at this time.
Tropical Storm Harry
The time series related to the genesis of this weak tropical storm are
shown in figure 55. An increase in estimated intensity from 25 to 40 kts is
seen to occur between 12Z of the 29th, and OOZ of the 30th, so genesis is
defined as occurring around then. The forcing parameter time series show
a peak at OOZ of the 27th for the 7 degrees and 5 degrees latitude averaging
radii, and at 12Z of the 27th when 3 degrees is used. The corresponding
potential vorticity maps (figures 56 and 57) show the approach of an
upper-tropospheric cyclone toward the pre-cyclone disturbance. On OOZ of
the 27th the upper cyclone is centered around 143E, 17N, with the
pre-cyclone disturbance found near 133E, 20N. Over the next 24 hours the
northern portion of the upper-level cyclone appears to overtake the
low-level disturbance, while the bulk of the cyclone appears to be
destroyed and replaced by low-PV air to the south. The collapse of an
adjacent cold pool is consistent with Riehl's formula for genesis. In this
40
case, upper-tropospheric potential vorticity advection is implicated, with
maximum forcing leading genesis by roughly 36 to 48 hours. At 850 mb,
Harry is observed to form and develop along the cyclonic shear side of a
southerly and westerly jet, which disappears by 12Z of the 30th as the
cyclone moves away toward the north.
In summary, the forcing at the generation of the pre-cyclone
disturbance is positive for the 7-degree radius average, but near zero or
slighltly negative for the 3-degree average, making the connection
between positive forcing and generation suggestive but quesionable. The
forcing is generally positive throughout the pre-cyclone disturbance stage
for the 5 and 7 degree averages, with a peak near 12Z of the 27th for the
3-degree average. Thus, the connection between forcing and genesis is
suggested, although perhaps not as clearly as in other cases.
Typhoon Ivy
Figure 60 shows the tracks of the remaining 2 cases of this study, Ivy
and Joel. The intensity time series (figure 61) for Typhoon Ivy is divided
as follows:
Phase
Time Inverval
1
08/31/12Z-09/01/12Z
2
09/01 /OOZ-09/04/06Z
3
09/04/06Z-09/07/06Z
Description
Ave. Intensification
Rate (kts/day)
pre-cyclone
disturbance
slow intensification/
genesis
moderate intensification,
tropical cyclone
0
11
23
The forcing parameter time series indeed shows several episodes of
positive forcing during stages 1 and 2.
The corresponding PV maps
(figures 62 and 63) show the upper-level flow patterns over the days prior
41
to generation of the pre-cyclone disturbance (figure 62), and during the
genesis and early intensification of the storm. Note the southwestward
approach of the TUTT between 12Z of the 30th and 12Z of the 31st toward
the site of the generation of the pre-cyclone disturbance, which first
appears to the southwest of the upper-trough.
By OOZ of the 3rd, the
intensifying disturbance appears free from the influence of the first
trough, but another anomaly is shown to advect down from the north to
provide another period of forcing. By OOZ of the 5th, Ivy has reached
maximum sustained winds of 65 knots. Potential vorticity advection by
upper-tropospheric shear is indeed implicated in this low-latitude case of
genesis, as well as in the generation of the precursor tropical disturbance.
The corresponding 850 mb charts (figures 64 and 65) show some
evidence of a migrating vorticity maximum within the equatorial trough
prior to the generation of the initial disturbance.
Note also the
development of a remarkable circulation on the large scale as the
disturbance/incipient cyclone intensifies.
Tropical Storm Joel
The intensification of Joel (figure 66) is divided into the following 4
stages:
Phase
Time Inverval
1
09/01/OOZ-09/02/06Z
2
09/02/06Z-09/03/18Z
3
09/03/18Z-09/04/18Z
4
09/04/18Z-09/06/12Z
Description
Ave. Intensification
Rate (kts/day)
pre-cyclone
disturbance
slow intensification/
genesis
steady at 30 knots
renewed intensification,
tropical cyclone
4
10
0
14
The forcing parameter time series shows an initial peak at the beginning
42
of the period, OOZ of the 1st, for the 3 degree averaging radius, and 12
hours earlier for the 7 degree average. The forcing is variable but generally
positive throughout the period. The associated PV maps (figures 67 and
68) show first the westward movement of a broad region of high-PV air,
with the generation of the pre-Joel disturbance occuring to its east by OOZ
of the 1st (figure 67). Over the next several days, the incipient cyclone
remains under the influence of northeasterly flow aloft, and generally
positive potential vorticity advection near the tropopause.
Forcing of
ascent from upper-level positive PV anomalies is implicated in both the
generation of the pre-cyclone disturbance, and in the genesis of the tropical
cyclone for this case.
The corresponding 850 mb maps (figures 69 and 70) show a fairly weak
vorticity pattern 24 hours prior to the generation of the initial disturbance.
Again, a large scale circulation begins to form during the pre-genesis stages
of development of the disturbance and incipient cyclone.
43
7. Summary/ Suggestions for Future Work
Observations appear to support the hypothesis that upper-level
structures often play a role tropical cyclogenesis.
Of the 8 typhoons
studied, only Ellie showed no indication of an upper-level trigger, while
for Caitlin the link was not obvious. The remaining 6 cases all had in
common episodes of positive potential vorticity advection by the
upper-tropospheric shear during the few days and hours prior to genesis.
The upper-tropospheric agents could be traced back in the analyses and
were clearly initially independent of the lower-level disturbances
involved in the process.
The upper-level agents often appeared as
elongated upper-tropospheric troughs, i.e. equatorward and westward
extensions of the TUTT, or else upper-level cut-off cyclones.
Of the
tropical storms, both Luke and Harry showed some indication of a
correlation between positive forcing and genesis, while for Doug such a
relationship was not apparent. For many cases, positive forcing was also
indicated at the very beginning of the JTWC best tracks, i.e. in the
generation of the initial disturbances themselves. This was certainly true
for each of the nonintensifying depressions, TD13W, and TD15W.
As for future direction in this work, it is suggested that the
incorporation of satellite radiance data might be useful. In particular, it
would be interesting to see if the episodes of positive forcing found in this
study could be related to any changes in the organization or in the amount
of deep convection. Also, the analyses performed in this study should also
be tried on nondeveloping clusters. An attempt was made by the author
to identify nondevelopers from the NMC 850 mb analyses, and a few
persistent nondeveloping vorticity maxima were catalogued. Forcing time
series were calculated for these as well.
Figure 71 shows an example.
Note that the nondeveloping vorticity maximum is not immune to
44
episodes of positive PV advection aloft. It would be more desirable to use
satellite imagery to identify nondevelopers as other authors have done
(e.g. McBride, 1981a). Also, it is suggested that other factors such as shear
over the disturbance center and large-scale low-level vorticity be
incorporated into the problem, which might help differentiate between
developers and nondevelopers.
It should be added that although the data scarcity over the tropical
oceans is well-known, standard observations are often found to resolve
the upper-level features described in this paper. It would be especially
useful to have access to the database which goes into any given NMC
analysis, and weigh data availability in the selection of cases from which to
draw conclusions.
Furthermore, analyses of PV fields surrounding
genesis cases in the eastern North Pacific and western North Atlantic
basins for 1991 found similar configurations for the many of those cases as
well, and it is not believed that the processes documented in this study are
unique to the western North Pacific.
45
References
Anthes, R. A. 1982: Tropical Cyclones-Their Evolution, Structure and
Effects;. Meteor. Monographs, Vol. 19, Amer. Meteor. Soc., 208pp.
Avila, L. 1991: Atlantic tropical systems of 1990. Mon. Wea. Rev. 119, pp.
2027-2033.
Bosart, L. and J. Bartlo, 1991: Tropical storm formation in a baroclinic
environment. Mon. Wea. Rev. 8, pp. 1979-2013.
Challa, M. and R. Pfeffer, 1980: Effects of eddy fluxes of angular momentum
on model hurricane development. J.Atmos. Sci, 37, pp. 1603-1618.
Challa, M. and R. L. Pfeffer, 1990: Formation of Atlantic hurricanes from
cloud cluster and depressions. J.Atmos. Sci., 47, pp. 909-927.
Charney, J.G., and A. Eliassen, 1964: On the growth of the hurricane
depression. J.Atmos. Sci., 36, pp. 2617-2630.
Chen, S., 1990: A numerical study of the genesis of extratropical convective
mesovortices. PhD thesis, Pennsylvania State University, University
Park, PA 16803, 178 pp.
Dey, C. H. 1989: The evolution of objective analysis methodology at the
National Meteorological Center. Wea. Forecasting, 4, pp. 297-312.
Dvorak, V. F., 1984: Tropical cyclone intensity analysis using satellite data.
NOAA Tech. Rep. NESDIS 11, NOAA, 47pp.
Eliassen, A., 1951: Slow thermally or frictionally controlled meridional
circulation in a circular vortex. Astrophys. Norv., 5, pp. 19-60.
Elsberry, R. L. 1987: Observation and analysis of tropical cyclones. A
Global View of Tropical Cyclones, R. Elsberry (ed.), pp. 1-12.
Emanuel, K. A. 1989: The finite-amplitude nature of tropical cyclogenesis.
Atmos. Sci, 46, pp. 3431-3456.
46
J.
Frank, W. M. 1987: Tropical cyclone formation. A Global View of Tropical
Cyclones, R. Elsberry (ed.), pp. 53-90.
Frank, W. M., and S. Chen 1991: Simulations of vortex formation in convective
weather systems. 19th Conference on Hurricanes and Tropical
Meteorology, Amer. Met. Soc., pp. 241-244.
Gray, W. M., 1975: Tropical cyclone genesis. Dept. of Atmos. Sci. paper no.
323, Colorado State University, Ft. Collins, CO, 121 pp.
Gray, W. M., 1979: Hurricanes: their formation, structure and likely role in
the tropical circulation. Meteorology Over the Tropical Oceans. Royal
Meteorology Society, pp. 155-218.
Handel, M., 1990: Tropical cyclone intensification from finite amplitude
disturbances. PhD thesis, Massachusetts Institute of Technology,
Cambridge, MA 02139, 261pp.
Holland, G. 1987: Mature structure and structure change. A Global View of
Tropical Cyclones, R. Elsberry (ed.), pp.13-52.
Hoskins, B. J., M.E. McIntyre, and A.W. Robertson, 1985: On the use and
significance of isentropic potential vorticity maps. Quart. J.R. Met.
Soc., 111, pp. 877-946.
Joint Typhoon Warning Center, Annual Tropical Cyclone Report 1990, 278pp.
Kanamatsu, M., 1989: Description of the NMC global data assimilation and
forecast system. Wea. Forecasting, 4, pp. 335-342.
Kanamatsu, M. et al., 1991: Recent changes implemented into the global
forecast system at NMC. Wea. Forecasting, 6, pp. 425-435.
Lee, C. S., 1986: An observational study of tropical cloud cluster evolution
and cyclogenesis in the western North Pacific, Colorado State
University report no. 403, 250pp.
Lee, C. S., 1989: Observational analysis of tropical cyclogenesis in the
western north Pacific. Part I: Structural evolution of cloud clusters.
Journal of Atmos. Sci., 46, pp. 2580-2598.
Love, G. 1985: Cross-equatorial interactions during tropical cyclogenesis.
Mon. Wea. Rev. 113, pp. 1499-1509.
47
Madala, R., S. Chang, et. al., 1987: Description of the Naval Research
Laboratory Limited Area Weather Prediction Model. NRL
Memorandum Report 5992.
McBride, J.L., 1981a: Observational analysis of tropical cyclone formation,
Part I: Basic description of data sets. J. Atmos. Sci., 38, pp. 1117-1131.
McBride, J.L., 1981b: Observational analysis of tropical cyclone formation,
Part III: Budget analysis. J.Atmos. Sci., 38, pp. 1152-1166.
McBride, J.L., and R. Zehr, 1981: Observational analysis of tropical cyclone
formation. Part II: Comparison of non-developing versus developing
systems. J. Atmos. Sci., 38, pp. 1132-1151.
Menard, R. and J.Fritsch, 1989: A mesoscale convective complex generated
inertially stable warm core vortex. Mon. Wea. Rev. 117, pp.
1239-1261.
Montgomery, M., and B. Farrell, 1991: Tropical cyclone formation, submitted
to Journal of Atmos. Sci.
Nitta, T. and Y. Takayabu 1985: Global analysis of the lower tropospheric
disturbances in the tropics during northern summer of the FGGE year
part II: regional characteristics of the disturbances. Pageoph 123, pp.
272-292.
Ooyama, K.V., 1982: Conceptual evolution of the theory and modelling of the
tropical cyclone. J. Meteor. Soc. Japan, 60, pp. 369-380.
Palmen, E. 1948: On the formation and structure of tropical hurricanes.
Geophysica 3, pp. 26-38.
Palmen, E. and C. Newton (1969): Atmospheric Circulation Systems: Their
Structure and Physical Interpretation. Academic Press, 603 pp.
Pfeffer, R. and M. Challa, 1981: A numerical study of the role of eddy fluxes
of momentum in the development of Atlantic hurricanes. J. Atmos. Sci.
35, pp. 2392-2398.
48
Pfeffer, R. and M. Challa, 1991: Formation of Atlantic Hurricanes from Cloud Clusters
and Deptessions, 19th Conference on Hurricanes and Tropical
Meteorology, Amer. Met. Soc., pp. 214-216.
Pedlosky, J., 1987: Geophysical Fluid Dynamics, 2nd ed., Springer-Verlag,
New York, NY, 710 pp.
Riehl, H. 1948a: On the formation of typhoons.
J.Meteor. 5, pp. 247-264.
Riehl, H. 1948b: On the formation of west Atlantic hurricanes, Misc.
Reports No. 24 of the Dept. of Meteorology of the University of Chicago,
University of Chicago Press, pp. 1-67.
Riehl, H. 1950: A model of hurricane formation.
J.Appl. Phys. 21, pp. 917-925.
Riehl, H., 1954: Tropical Meteorology, McGraw-Hill Book Co., Inc., New York,
NY., 392 pp.
Riehl, H., 1979: Climate and Weather in the Tropics, Academic Press, 611 pp.
Rotunno, R. and K. Emanuel, 1987: An air-sea interaction theory for tropical
cyclones, Part II. J.Atmos. Sci., 44, pp. 542-561.
Sadler, J.C., 1975: Tropical cyclone initiation by the Tropical Upper
Tropospheric Trough, NAVENVPREDRSCHFAC Tech. Paper No. 276.,
103 pp.
Sadler, J. C., 1976: A role of the Tropical Upper Tropospheric Trough in
early season typhoon development. Mon. Wea. Rev., 104,
pp. 1266-1278.
Sadler, J.C., 1978: Mid-season typhoon developments and intensity changes and the
tropical upper tropospheric trough
Sawyer, J.S. 1947: Notes on the theory of tropical cyclones. Quart. J.Roy.
Met. Soc., 73, pp. 101-126.
Shimamura, M. 1981: The upper-tropospheric cold lows in the northwestern
Pacific as revealed in the GMS satellite data. Geophys. Mag., 39, pp.
119-156.
49
Shimamura, M. 1982: An application of GMS satellite data in the analysis of upper
cold lows in the western North Pacific, Geophys. Mag., 40, pp. 113-152.
Simpson, R. H. and H. Riehl 1981: The Hurricaneand Its Impact; Louisiana
State University Press, Banon Rouge, Louisiane, 398pp.
Smith, W. L., 1991: Atmospheric soundings from satellites-false
expectation or the key to improved numerical weather prediction?
Quart. J.R. Met. Soc., 117, pp. 267-297.
Sundquist, H., 1970: Numerical simulation of the development of tropical
cyclones with a ten-level model, part I. Tellus 22, pp. 359-390.
Yanai, M. 1964: Formation of tropical cyclones; Rev. Geophys. 2, 367-414.
Yanai, M. 1968: Evolution of a tropical disturbance in the Carribean Sea
region. J.Met. Soc. Japan (Ser. I) 46, 86-109.
Zehr, R. M., 1991: Observational investigation of tropical cyclogenesis in
the western north Pacific. 19th Conference on Hurricanes and Tropical
Meteorology, Amer. Met. Soc., pp. 235-240.
50
200 mb
COHPOSITED
VT
case .
Figure 1. Horizontal distributions of 200 mb relative vorticity in units of
10-6 s-1 referred to the composite vortex centers for (top) pre-typhoon
vortices and (bottom) nondeveloping vortices (Nitta et. al., 1985).
51
DATB/TlI6
RAOIUS.*-4 39CON*A
%bo
DATB/T1
.
Ow.
C
WM
go
M-uumum Sea -Level Pressure
--
1010
*o*
r
-
I
axmm
Sufcwtd(m)
clouds
Oeep Conknective
STGE
Doy
-
STG
eoree
Tropico Storm
u
Figure 2.
Sample of time series from two case studies exhibiting early
convective maxima.
Also, schematic illustrating Zehr's two-stage model for
tropical cyclogenesis (Zehr, 1991).
52
0 0 &COE
0
00
00
dc
00
0
CI0
I
Ok
o 0
0o
0
01 0
0
V
I
0
0I
0
0 0
O
.
-
-~--
0
Iv
01
*
C
NJ
U
0
;0
-
IN
IfI
0
:1
Figure 3. Components of database for NMC analyses on a "typical day" (Dey,
1991):
(a) upper air soundings, (b) commercial aircraft reports, (c) cloud-tracked
Latitude
wind observations, (d) satellite temperature profile observations.
Figure 3a.
and longitude lines are shown for every 30 degrees.
Radiosondes (circles), and Pibals (asterisks).
53
0
U
4
I
I
I
I
___
I
I
CYD
0)
*1
0
P1
TROPICAL STORM HURRICANE PATTERN TYPES
PRE
DEVELOPMENTAL
PATTERN TYPES
STORM
T1.5 :.5
CURVED BAND
PRIMARY PATTERN TYPE
(Sirong)
(Minimal)
T2.5
F'TF
T3.5
(Minimal)
TA.5
(Strong)
TS.S
(Super)
T6.5 . Ta
)
CURVED BAND
EIR ONLY
C_________________
CDO PATTERN TYPPE
VIS ONLY
SHEAR PATTERN TYPE
MSLP
(NW Pacific)
MSLP
(Atlantic)
MWS
nots)
(KY
CI
Number
25 K
1.5
25 K I
2
2.5
3
30 K
35 K
45 K
3.5
4
4.5
5
|
1 1009 mb
1 1005 mb
1000 mb
1000 mb
997 mb
991 mb
-987 mb
979 mb
970 mb
976 mb
966 mb
954 rb
55 K
i 994 mb
65 K
77 K
90 K
1
1
5.5
102 K
960 mb
6
115 K
127 K
140K 1
948 mb
6.5
7
7.5--
155 K
8
170 K
984 Mb
941 -b
I
935 mb
I
921 mb
I 906 mb
I 890 mb
927 Fb
914 mb
898 mb
879 mb
858 mb
Figure 4.
Dvorak technique: Primary cloud patterns and associated
"T-numbers" (top).
Table relating Cl-numbers to intensity (bottom).
For
the developing stage of tropical cyclones, Cl-numbers and T-numbers are
equivalent (Dvorak, 1984).
57
---
----------------------
L----------------a-
.
....
a
a
a
.0
O
a
U
--
- - - - - - - - - - - T --
a
a
a
a
a
7
k
- -
- - - -
L
--------
a-----aI
a
a
a
a
I
- - a
a
------------
X
- - - - - - - -
d
,.
a
*N
a
a
a
aCa
a
aO
-----------------------------------------fa aa
a
a
a
X
a
a
a
a-'0*
z
L
~1
-
C
0
z
C
a
a ~..I
~.'-'
a If
a
a
a
%J
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
-
04
01,
~~~~0
Figure 5. Tracks of typhoons Zeke (Z), Amy (x), Brendan (+), and Caitlin
(*). Plotted also is the day the observation was made (in July 1991), with
OOZ, 06Z, 12Z, and 18Z fixes shown sequentially for any given full day of
observation.
58
INTENSITY OF ZEKE
80
75
70
5
60
55
-50
40
35
30
25
20
15
70700
70800
70900
71200
71100
71000
ORTE MOOHH
71300
71400
71!
FORCING. RADII 3(R)5(B),7(C) ZEKE
4.0
3.5
3.0
JV\V
2.5
2.0
1.5
1.0
.5
0
-. 5
-1.0
-1.5
-2.0
70600
70700 70800 70900 71000 71100 71200 71300 71400 71500
DATE MMOOH
Figure 6.
Intensity (top) and forcing parameter (bottom) versus time for
typhoon Zeke.
Times are of the format MDDHH where M is the month
number, DD is the day, and HH is the hour. Forcing parameter series is
shown for averaging radii of 3, 5, and 7 degrees latitude.
59
to
to0
ON~
to 1.-
to -
4.
0
fl40
0
V- 0
0(i
in
-
L0
.,V>in)
0
71,
-~OF
A
CD,
r
~00.F
I'
-, Li
i
C',-
*94
I
r
"
I f3 F
r.P a 41
in)
9P
in
1-N
ON
to
C,,
C)
in
Le-a
M,
Q
a2 itP
r0
§v;. 9?
f
Figure 7. Ertel's potential vorticity (Q) fields plotted on the 360K isentropic
surface for the times shown at the bottom of each panel. Contour intervals
are 0.5 PVU for Q less than 1.5 PVU, and 1.5 PVU for Q greater than or equal
to 1.5 PVU. Negative values are indicated by thin solid lines, 0.5 and 1.0 PVU
contours by dashed lines, and contours of 1.5 PVU and greater contours, by
Winds are plotted in knots, with half-barbs, full-barbs
heavy solid lines.
The track of
and flags each representing 5, 10, and 50 knots respectively.
the disturbance/cyclone is indicated by the squares, with positions every 6
hours shown, beginning at OOZ of the 6th.
T
60
ID
I-
a
ION
1~1~
aN
a
Sn..
q.~j
'-C.)
4t0U*
0
%
Sn
CSa
-
'-Q.
W
LOa
LU
kno
IIt
CLF
'O
Figure 8. Ertel's potential vorticity (Q) fields plotted on the 360K isentropic
surface.
Fields are shown at 12 hour intervals during the pre-cyclone
disturbance phase of Zeke. Large triangles denote current positions of the
pre-cyclone disturbance or cyclone, and squares represent its track.
Contouring conventions are as in figure 7.
61
*U
U,
0N
CD
U,
0
0..
I-
wn
0O
\I 1
.i
[I
gi
V-J
Ln
K~7
I-
Nh.
N
1CO
0n
U,
0q
r0
r'.:
0j
_
'I
113
rr~
Figure 9. Ertel's potential vorticity (Q) fields plotted on the 360K isentropic
surface.
Fields are shown at 12 hour intervals during the pre-cyclone
disturbance phase of Zeke.
Contouring conventions and meaning of
symbols are same as in figures 7 and 8.
62
0
0
U3
CV)0
V- u,
r
mr
0
N4
ON
Ln1
N
N
1
U
N
1
U,
In
-M-
-*M
an
.I.-
C
Figure 10. 850 mb winds and relative vorticity fields in units of 10-5 s-1.
Contour interval is 2 with positive values indicated by heavy solid lines,
and negative values dashed. The 1 and 3 contours are also plotted, indicated
Fields shown for times shown beneath each panel.
by thin solid lines.
Winds are plotted in knots, with half-barbs, full-barbs and flags each
The track of the precyclone
representing 5, 10, and 50 knots respectively.
disturbance and then cyclone Zeke is indicated by squares as in figure 7.
63
I
%
II
I
I \1
-I/
\
\k \A
In
r
I,
In In
VU,
in
C)I
40
F
'V.
0
CI
CD
R,
CD
rI
'~)
N1
~ :''Z".
Om
0
r0-
:3
~'~
C
0
N
850 mb winds and relative vorticity fields in units of 10-5 s-1.
Figure 11.
Triangles denote position
Contouring convention is same as in figure 10.
Track shown as in previous
of pre-cyclone disturbance related to Zeke.
figure.
64
INTENSITY OF RMY
SI
71300
I
i
71400
71500
I
I
71600
71700
ORTE MO0HH
71800
71900
i
I
72000
72100
FORCING. RADII 3(A).5(B).7(C) AMY
2.0
1.8
1.6
1.4
1.2
S1.0
-.
8
R 0 .6
CL
.4
GS
- .6
-2
-. 4
71300
71400
71500
71600
71700
MMOODH
71800
71900
72000
DATE
Intensity (top) and forcing parameter (bottom) versus time for
Figure 12.
typhoon Amy. Times are given in the format MDDHH where M is the month
number, DD is the day, and HH is the hour.
65
-,.-
-
%_h
1,0
W tc+- Y~~..o
%
.'
04.~
4.4
.00
_O
~
k,
VII >
If
0
I
In
Io
o11
In
In
L
It
-
If
'.--
gal
IL
0
-rL
71,
13. PV on 360K surfce.FedDr hw a
Figure
Ertel'
r-yloecutrreaeoAy
leading
u to the gnerationof h
poition
at. 1re' 8Z of theK
12h sotura conentios
66
aeshoin fiur
4hu
nevl
7.huritevl
('M
0-
AA
to
rC.,,
U,
wU
Cn
t
C31
4"
E3
C
LO
* r
0
Figure 14. Ertel's PV, 360K surface. Fields shown every 12 hours during the
pre-cyclone disturbance phase for Amy, whose location is indicated by the
large triangle. The track of Amy is given by the squares. Track of Brendan
also indicated in panel for 12Z of the 14th. Contouring conventions as in
figure 7.
67
/L
I
o~)1
V)
U1
r
U)
Prr0
r
In
in0nL
r
on18 n h 1t.
or r
/0
otorngcnvnis as in fiueD0
M68
)V
FORCING2, RROII 3(R)1 5(B),7(C) RMY
2.0
1.8
1.6
1.4
.,
1.2
1.0
z
.8
.6
I-
L.
CL
.4
.2
0
0
-. 2
-. 4
-1.0
-1.2'
-71300
71400
71500
71700
71600
DRTE MMOOHH
71800
71900
72000
Figure 16. Forcing parameter for Amy computed with 850 mb wind
velocity used to represent the lower-level in the calculation.
69
INTENSITY OF BRENDAN
5
s-
60
--
55
50-
30-
25S
25
-10
71600
71700
71800
71900 72000 72100
OATE MO0M
72072300
72400
72500
FORCING, RADII 3[R),5[B).7[C) BRENDAN
3.5
3.0
2.5
15
2.0
10
1.0
--
-20 -
-2.s
71500 7160
71700 71800 71900 72000 72100 72200
ORTE
72300 72400 72500
MMOOH
Intensity (top) and forcing parameter (bottom) versus time for
Figure 17.
typhoon Brendan. Times are given in the format MDWDHH where M is the
month number, DD is the day, and HH is the hour.
70
OQJ
(Q
~CD
0
0
co
hiL
WD
oco
C)
~<CD
00
135
135
155
160
150
145
140
ERTEL'S PV, 360K SURFACE: 071600
155
160
145
150
140
ERTEL'S PV, 360K SURFACE: 071500
165
165
170
170
1.5
0.5
0.5
1.5
0.5
41S
135
0
10-0
06-
165
1;?
140
145
150
155
160
ERTEL'S PV, 360K SURFACE: 071612
tj
r
10,
140
145
150
155
160
ERTEL'S PV, 360K SURFACE: 071512
6
Awl
135
170
S
a
,,
k-
o
-
DC7
Lo
o
InI
Fiur 19-re)kVo
Trck0f
m and Brna
6K ufc.
arMhwsi
Cnorcnenin
72O
si
iue7
peiu iue
ry
N
iLA
LULs
41
3
L
r;
,'
n
C4
'-
,
:; TC
4
o
*'u.
a,I-~
0
'10
0
%
).-OOUL-
0
0
(0
CI,
In
N
%
%%
n
0..
C,,
%%
z; r
in
, %.6
-.
,-
I--r.
E
%0
U,
0
Ui
It..t-
1W7 r I
i
* . r VI
%~
'i
It
\
LnL
I.,
00 0
~
'-0
N
0
a -
A
>
0
10
C.,
U,
N>
NA9
'-a-
.o-
SI,
I
'I
LU
U,
. ,,V,
#
-k
'k
'4
~,
:~
Figure 20. Ertel's PV, 360K surface. Fields shown every 12 hours. Asterisks
denote current positions of typhoon Amy. The track (squares) and position
of Brendan are indicated. Contouring conventions as in figure 7.
73
-C--A
t-*
0
-14
0-k
*.
lku
% V,
-
,~
0
U1%
In
Lj
In'n
04
04
%
Y,
rr AK~~0C
,)
dlA
I
.
Fiue2.
EtlsP,30
ufc.
seik
eoepstoso
h
prcuso dstrbnc
t Citin
rak ndpoiton o Benanasi
prvosfgr.
Cnornovnin
iue7
si
74
%o o
to
W
r
r
r r
r
o
covIon
asi
fgr
a
r
ro7
0
(r0
10.
75n
INTENSITY OF CRITLIN
100
90
70
40.
150
30
20
---
.
OATE MOOit
FORCING, RADII 3[R),5[B),7[C)
2.4
CAITLIN
tI-
I..
2.2 -2.0
1.8
--
1.4
z
1.2
1.0
.8-
0
1.0'
-
-
-. 2 -l
-. 4
71900 72000 72100 72200-9230 72400 725W 28
ORTE 1900m
Figure 23.
72700 728W 72900 7
Intensity and forcing parameter time series for typhoon Caitlin.
76
-
-
-
, I
.
,
I
I
I
Io
040
I,-
I*
4/%
M
%a
2
L
I..
<
4C%
%--
IF
7t-3
C
re'V
6Ksufc.Atrssdnoepsin fcdn
Breda.
(qures racad osiios tranges o Catln re ndcaed
CotuigcnetonVsi iue7
77
Fr 24
0~w
%1!J I0.
4I
%
'.
SL
1
j
----
a.
Figure 25. Ertel's PV, 360K surface. Asterisks denote positions of cyclone
Track (squares) and current positions (triangles) of Caitlin are
Brendan.
Contouring conventions as in figure 7.
featured.
78
-
o
0
0
~
160
160
155
150
145
140
135
130
910717/1200 850 MB VOROBS (*10*5)
155
150
140
145
135
130
910719/1200 850 MB VOROBS ('10"5)
3
165
-2
165
0
0
-2
850 P4flVOROBS (*10*5)
130
135
140
145
150
155
910720/1200
850 MB VOROBS (*10**5)
910718/1200
160
165
N
N
Figure 27.
850 mb winds and relative vorticity.
current positions (triangles) of Brendan and
Contouring conventions are as in figure 10.
80
Tracks (squares) and
Caitlin are shown.
7-
a
r"
xX
aO
xxX
x
x
-
-
x
x
Cc
x
+
C4
-X
M
...x
X
CD
--
+
4.
n
x
+0
+
--
+0
M
+
*
x
O2
Figure 28. Tracks of Doug (x), Ellie (a), TD13W (*), Gladys (+). Plotted also is
the day the observation was made (in August 1991), with 00Z, 06Z, 12Z, and
18Z fixes shown sequentially for any given full day of observation.
81
INTENSITY OF DUG
34
32
0
e
s
ze
-28
28
807M0
800
0850
80900
80950
ORTE MooHH
81000
81050
81100
81150
FORCING. RRDII 3(R.5(B),7(C) 0UG
.2
0
I
-. 2
C-
.4
-. 8
-. 8
80750
80800
80850
ORTE
80900
80950
81000
81050
MM00HH
Figure 29.
Intensity (top) and forcing parameter (bottom) time series for
the case of Tropical Storm Doug. Note changes in the scale of the axes.
82
m
00
o
.
O
140
140
0.'
--
V
r
145
ERE'
150
155
36KSUFCE
or
160
801
170
1.5:
16.5 170
150
155
160
165
145
ERTELS PV, 360K SURFACE: 080612
7-
21
1.
i
17
1
r
5
3
3
140
4§5 0 10
155
180
165
ERT L'S PW 360K SURFACE: 080800
170
175
cxn
0
o.
0n
0
o
Oi
OQ
'1
I3b'%
I
"
-'-:'
150
155
160
165
PV, 360K SURFACE: 080912
i~. -I 13t! 1.5
170
13 0
140
140
145
150
155
1§0 165
ERTEL'S PV, 360K SURFAGE: 081000
170
145
150
155
160
165
170
ERTEL'S PV, 360K SURFACE: 0809%
1
175
175
n>
5.
n
0r910806/
~
-7r
00
0
(n(D
oil
-14
'1
1A
200
O
M
17
170
15
160
155
150
850 MB VOROBS (*10**5)
00
'163
5 '16'0
' 150~
850 MB VOROBS (*10**5)
910808/1200
140
-
-
175
175
0
2
INTENSITY
OF
ELLIE
so
80
70
Sso
-40
30
20
10
81000 81100 81200
81300 81400
81500
81600 81700
81800 81900 82000
OATE MOOHH
FORCING. RAOII 3(A).5(B).7(C) ELLIE
3.0
2.5
2.0
-1.0
-'.5
-2.0
80900 81000 81100 81200 81300 81400 81500 81800 81700 81800 81900
ORTE MM00MH
Figure 33.
Intensity (top) and forcing parameter (bottom) time series for
the case of Typhoon Ellie.
86
o
o
C
D
(3<13
0~0.
5
3
10
140
-
0.5
ERTEL'S PV, 360k SURFACE: 081300
ERESP,36kSRAE:010
't
i
166
155
160
145
150
ERTEL'S PV, 360K WACE: 08110
N
z-
170
4.5
.
0.5
11
175
\
3
140 0 145
150
155
160
165
1-0
ERTEL'S PV, 360K SURFACE9-%812001-5
175
0
a
a
u
x
x
X
X
a
0
0
X
--0
X
x
X
x
x
a.
I
ax
;r
x
-.
In
X
X
I-
I-X
...
I
X
X
X
-.
N
Figure 35.
Track of typhoon Fred.
88
-
-
INTENSITY OF FRED
80900
81000
81100
81200 81300 81400
ORTE MOOHH
81500
81600
81700
81800
FORCING. RROII 3(R),5(B),7(C) FRED
-1.0
-1.5
-2.0
80800 80900 81000 81100 81200 81300 81400 81500 81800 81700 81800
DATE MMODHH
Figure 36.
Intensity (top) and forcing parameter (bottom) time series for
the case of Typhoon Fred.
89
0D
to
LM
U,
Uf)04
oo
0
M,
to
U,
0:j
Vw
U
Ln
04
LO
LM
LO
V_
U,
00
0
Go
0
0
An
'
'
.00
to
'T-0
*A-F
CI,
76
i
I~
Triangles denote positions of
Ertel's PV on 360 K surface.
pre-cyclone Fred, with its track indicated by squares.
Contouring
conventions as in figure 7.
Figure 37.
90
.
1~
10
4
IVI
x0
F
cm
-WC
r..
~
N
K.
~
4I
Ln
cvit
Figure 3 8. Ertel's PV on 360 K surface. Positions and track of Fred are
shown as in previous figure. Note gap between panels 1 and 2. Contouring
conventions as in figure 7.
91
0
00
M1
0
0.5
ofr
0
10 5
1
0.5>
f30
3
E
15
31
S
120
'125-
0
-~0.
EITTEL'S PV, 360K SURFACE: 081500
1
110
30
130
125
120
115
110
L'S PV, 360K SURFACI: 081300
'10
4-.-
0y
p
1
CO
105
E
15
31''100
3
15
1.5
25
30
100
~1
1
-i
15n
2
15
125'
4.51
1
1
125
120
RFACE: 081400
A
ERTEL'S PV, 36V SURFACE: 0811600
110
EdL'S PV, 360
105
o.
--
t
'0001))Lnr
r
Ak
Po'
In
r-
E
CV)
C41
-7
0
IPA
LI ,-0
/
0-
'0-
0
CU41
931
0
C
0
0
0
0
0
1.0
CA
S00
JQQ
0
0
13
125
850 MB
120
k
130
135)
VOROBS *10**5)
1/1'Q
135
130
125
120
115
110
850 MB VOROBS (*10*5)
910811/1200
115
910809/1200
110
\1
I
140
140
145
110
200
135
130
125
120
850 MB VOROBS (*10**5)
135
130
125
120
115
910810/1200
850 MB VOROBS (*10*65)
140
140
145
145
FORCING, RROII 3(R),5(B) ,7(C) T013W
3.0
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
a 1I 4
p I 1a
2.5
2.0
1.5
1.0
.5
z
.49
0
-D
-
--
-0
N-I
-. 5-
-.
0
R- -1.5
-2.5V
-3.0
-3.5
-4.0
' ' '
81150
81200
81250
81300
81350
DRTE MMDOHH
81400
Figure 42. Forcing time series of Tropical Depression 13W.
95
81450
81500
F
CY
AV
j
k
I
I
ORO,
4 ,
N
I
L
CY
-A
cz.
CO
co
r V,
Le)
LO
In
90
cc
tr) C',
LnW)
>
%
%r
f7
Cn
Cn:j
LO
lu
r
Ln
r
Ln
Ij
DCO
y-
k,
Op
Nlooo
F
Owl%
Ilu
CD
rd-,
%
CO
Ln
%
.%
13 Ir4
%
T-C3
to
cn
LO
0
$3
r
r
%
:j
Uj
elolf7
,
%
r %
inLn LU
%
r
f
03
LO
C
Ln
O"S
M
M
OF
Q
Ln
Asterisks denote positions of
Figure 43.
Ertel's PV on 360 K surface.
tropical cyclone Ellie. Positions and track of Tropical Depression 13W are
Contouring conventions
indicated by triangles and squares respectively.
as in figure 7.
96
14
0
r)r
(Dr
(D~
ot~
INTENSITY OF GLADYS
85
80
ss
5
50
45
-40
3s
30
25
20
15
10
81400 81500 81500 81700 81800 81900 82000 82100 82200
DATEMM0tH
8230
82400 82500
FORCING. RADII 3(R).5(B).7(C) GLADYS
2.0
1.8
1.6
1.4
1.2
1.0
.8
.5
.2
0
-. 2
-. 8
-. 8
-1.0
Figure 45.
71400 71500 71600 71700 71800 71900 72000 72100 72200 72300
DATE MMOt
Intensity and forcing parameter time series for Typhoon Gladys.
98
(0
0o 0
0CD
CD
0
o -01
to 0
En
0
(IQ
uC
-J
-+po
0
ca
3 1W
.5
Z5r
140
'
,
D
-X
.
'I
1450. 150
155
(160
165
ERTEL'S PV, 360K SURFACE: 081300
w 0o
3I
170
4
1P5
-
14Q).5
135'
'160
-1
ERTEL'S PV, 360K SUFACE:
-145-'' 150
150
155
160
165
L'S PV, 360K SURFACE: 081312
1tq
.
175
0
0
~0
0
00Q
0
r),
0
cu1
130
4 0130
15
5
5
.5
10
14CI0
145 -
SOr
-O-N-.a
150
0
155
140 445
150
55
L'S PV, 360K SURFACE: 081512
ERTEL'S PV, 360K SURFACE06B1712
15
135
1
160
160
%
165
165
15
1I
1.
t30
0
-
130
--
-m-
3',
i
--%-I &
E- --
--
m us .
I
'145
I
150
f5
1 160
160
p
ERTEL'S PV, 360K SURFACE: 081812 0.5
5 -0. 40
z I
1350.5 140
145
150
155
ERTEL'S PV, 360K SURFACE 081612
r~t
3Q4
6T5
1.5
- '166
1
0. 0
ts
~CD
oLA~
CDi
.r
04
0'*
o'
o
SOO
-0
140
150
910813/0000
145
910811/0000
155
160
16
170
850 MB VOROBS (*10**5)
850 MB VOROBS R&0*5)
-
175
910814/0000
850 MB
VOROBS (*10**5)
c
0
0
0 co
0
0
0
o0
2
130
5
0
"-0
I
/
N
r
910816/140
850 MB
145
140
105
910814/1200
850 MB
ra
231
(|/.
Y.
AW
///\
|
.
82
VOROBS (*10**5)
/
160
155
150
VOROBS (*10**5)
t
.
3
-
-2
1 5
.
r
-2
-2
x
300
29+
21
2aaa
a
I
a
a
a
a
a7
a
29
30
3030
30-- - - - - - - --- ------------------- u 2
+ 27 2 X
a
4
a+
a
+
202
-- - - - - - - - - - - - a - - - - - - a
h2
a6
:a
x
ao
a+
2a
25 aa
+
2+
+
+
a222 t~
++
+++
27X
a
xa
a
X289
ax28
a
27aa
20----------------
------
13
a
x
----------------------------------------------------------
1Aa
TOL514 (+19
Figure 50.
(x).
a
a5
HAIRRY (XI
Tracks of Tropical Depression 15W ()and Tropical Storm Harry
103
FORCING AND INTENSITY VS TIME
TD15W
A
-7
-C
t
I
82200
I
I
82300
I
82400
-#
I
I I I
II
I
82500 82500 82700
ORTE MMOOHH
I
82800
I
82900
1
I
83000
FORCING. RADII 3(A).5(B).7(C) TD15W
2.0
1.8
1.6
1.4
1.2
1.0
.8
.8
.4
.2
0
-. 2
-. 4
-. 8
-1.0
-1.2
82200
82300
82400
82500 82500 82700
om MM00m
82800
82900
83000
Figure 51. Top panel shows intensity (curve B) versus time, as well as the
forcing (multiplied by 40) as computed with a 5 degree latitude averaging
radius (curve A).
Bottom panel contains usual forcing time series for
Tropical Depression 15W.
1 04
C)
0o
-l
1-
p~*0
0
..
M
*'
-- o
0
o
0 M
(- r.
Lt
155
145
150
135
140
ERTEL'S PV, 360K SURFACE: 082012
Z~iWA1
TI# . I .f
130
0
1.5
20
259-
3
2-
'
135
140
13(0
1
30
--
-F
150
155
.1:6-- ^*^
145
*
.5
1.5
.
160
4.
135
140
145
150
155
160
ERTEL'S PV, 360K SURFACE: 082200
-F
EiTEL S PV, 360K SURFACE: 082100
A; r41 .
130
65.5
3
.5
165
f165
on
S--
-
SE.
<0
01
(-)
00
M
o
-
30
.53
,o
1AI
- -
,P~~:.
A
,
a-
*
05~--2A&S
155
135
140
145 0 150
125 0 130
ERTEL'S PV, 360K SURFACE: 082300
..
1_1
I L- -
2K,
0-
1.
----
i
160
I .7
e
16
4U
1.
12
25
30
.1
35'
35
125
1P5
4P
150
140
145
135
130
ERTEL'S PV, 360K SURFACE: 082600
130
1
140
145
150
155
ERTEL'S PV, 360K SURFACE: 082400
/~ 'r
-0
160
-
)
O
o Ni
0U
0D'
o
t
O~
Omla.
0
1.-
20
125
125
-
r
I I
w-
130
135
910822/1200
135
130
910820/1200
aAll
IN
140
850 MB
140
850 MB
\
-a"
lila
S.>
)..**
"It NI N
145
150
155
VOROBS (*10**5)
145
155
1g0
VOROBS (*10**5)
-,I i
a;
160
160
--
-2
125
130
135
910823/1200
140
850 MB
145
150
155
VOROBS (*10**5)
160
INTENSITY
OF
HARRY
*
127s50
62900
1250
32900
2950 8300 0ss
ORTEMOOM
p.
63003100
. . * I..
s3150
FORCING. RRDII 3(A).5(B).7(C) HARRY
DATE HWIOHH
Intensity and forcing parameter time series for Tropical Storm
Figure 55.
Harry. In time labels, HH=50 corresponds to 12Z, i.e. 83150 is August 31st,
12Z.
108
Q
IS
V>
40.
(n3
~C i
IP
Ir
t
T'
Ix
x
I
11
T
~1
Cs
I4
T_
In
A
r\
'f
/ L' 4
I
I
MIS
a
U)
11
Nf
N0
r--n
iT
1I-
IV
V'
cp.
of
IQ' O
3.0
I'-
0
-l
in
Figure 56.
Ertel's PV, 360K surface.
Triangles denote positions of
pre-cyclone Harry, which first appears on OOZ of the 27th near 133E, 19N.
Track of Harry is also shown (squares). Contouring conventions are as in
figure 7.
109
0
0
C0
(
-
1.
300
15
2U
%.
JTn
30
35
40
I
A2
,f r
3(
160
:,Ile %'oft
1510
135
140
145
150
EE'
130
5
ERTEL'S PV, 360K SURFACE: 082912
I
I,
4
( k5"
0 125
130
135
140
145
150
155
ERTEL'S PV, 360K SURFACE: 082812
1.5
q
.
-5
--
.5
15
201
25
30
35 -
12%
I
A
ikk
3
130
135
140
145
150
155
160
IRTEL'S PV, 360K SURFACE: 082900
1.51
I
40Q
ICO
90
-Q
$O M
o ul
0o
0.
S2Z1
155 2 160
150
145
140
135
130
125
850 MB VOROBS (*10**5)
0 910826/1200
135
140
145
150
8A MB VOROBS (10**5)
h*5~
155
160
o
o
o
0
od
0
m
0
*1
9-'
0dO
0
(n
JA
o (D
0 0"
120
910830/1200
850 MB
125
130
135
850 MB
910828/1200
VOROBS (*10**5)
140
145
150
VOROBS(*10*5)
155
120
910831/1200
850 MB
125
130
135
910829/1200
850 MB
VOdbBS (*10**5)
140
145
150
VOROBS (10**5)
155
a
a
*
+ +
aa
-
I
a
+
M
a
+1
a~
+
Saa
+
+~
(1
X
+
X
+X
i+
------------------------- ---------------------
---
----------
----
i+
+
+
+
+
+
X
a
Figure
60.
Tracks
of
Typhoon
Ivy
a
and
113
Tropical
Storm
Joel
(x).
INTENSITY OF IVY
90100
90200 90300 90400
90500 90600 90700 90800
ORTE MOOH
FORCING. RRDII 3(R),5(B),7(C)
90900 91000
91100
IVY
1.5
1.4
1.2
1.0
.8
.6
.4
.2
0
-. 2
-.6
-.0
-1.0
90100
Figure 61.
90200 90300
90400 90500 90600 90700
ORTE MMDDHH
90800 90900
91000
Intensity and forcing parameter time series for Typhoon Ivy.
114
0
CO
'LN-
T_
D
00
0
EME/it
-M-
-
F,,
rL2
o~
-f
I
\tNO
EMMM
(ave
MDO
M-U
0
U,::
LC
In
WN
0
0
V
to
LnCv
U,
0
U,
'-U
Ul
Figure 62. Ertel's PV, 360 K surface. Positions of pre-cyclone Ivy denoted
by triangles.
Track of Ivy is also indicated.
Contour conventions as in
figure 7.
115
a.
In
Lna
CDaD
in
p.
nQ
IT
a W
An
%
0
~
~R
U11
o
o
CA
P1o
ro
cn3
0
C) P-
CDi
.
oo
0-.
160
165
170
145
150
155
850 MB VOROBS (*10*5)
910831/1200
175
180
0
o
(D
00
-- j
(D1
'-1
0-
o,OP..
o
o
5. os
<<
:!
o C
160
155
150
145
140
135
850 MB VOROBS (*10*5)
910904/0000
165
170
135
140
145
150
155
160
910905/0000
850 MB VOROBS (*10*5)
165
170
INTENSITY OF JOEL
55
so
45
a
z35
6o
2S
20
15
10
DATE MOOMIH
FORCING, RADII 3(A),5(B),7(C) J0EL
1.4
i~i'I'I'I'I'
I
11
1.2
1.0 |-
.6
LU
0|
C-
aLU
.2
0
-. 2
90100
Figure 66.
Intensity
Tropical Storm Joel.
90200
90300
and forcing
90500
90400
ORTE MOOmI
90600
parameter time
119
90700
90800
series pertaining to
%
-"cr
In
In M
,-0
0D
C,,
r
Lnl
m) CI
CD)CV
0.
A~
I
LMLL
In
In
In
LO
Ur,
V)
n U,
iC,,
00
'C/)
0-
W)1
wVV
0\
M
-18U*l
N4
._1C~
04
'U)." '0
-
1-
In
Figure 67.
Ertel's PV, 360 K surface.
Track of Joel
denoted by triangles.
conventions as in figure 7.
120
Positions of pre-cyclone Joel are
indicated by squares.
Contour
0
t
Im.
CD
g
e
CD
00
CD
.0
D02
cn
CD
o c71
O\
110
5
20
115
-Oo
1
125
130
420
UEffTS PV, 360K SURFACe.9 0400
SN
0 145 0
N)
00
0
000
Q3r
0
-
-
-~
I
L .j
!e
jR
11
-
IN
I
r
4
SIN
04
04
Ln
M >
of
Harry. Contouring conventions as infigure 10.
122
gn
0
inl~
U))
0
o
::rO
0
m~
(Ap
0(3
~oo
0.
110
9?0904/0000
115
120
910902/0000
850 MB
VOROBS (*10"5)
1
130
135
140
85 MB VOROBS (*10"5)
6
145,
4
110
910905/0000
115
120
910903/0000
850 MB
VOROBS (*10**5) ~
125
130135
140
850 MB VOROBS ('10**5)
145
10
FORCING, RRDII 3(A).5(B).7[C] ND4
2.4
2.2
2.0
1.8
1.6
1.4
1.2
1.0
.8
.6
.2
0
-. 2
-.A
-. 6
-. 8
-1.0
-1.2
72400 72500 72600 72700 72800 72900 73000 73100 73200 73300 73400
ORTE MMO0HH
Figure 71. Forcing time series for nondeveloper number 4.
124