Skip to main content

Full text of "DTIC ADA153085: Characteristics of North Indian Ocean Tropical Cyclone Activity"

See other formats


_ UNCLASSIFIED _ 

SECURITY CL ASS1 FICAT10N-0F THIS PAGE fWhen Date Entered) 


REPORT DOCUMENTATION PAGE 

READ INSTRUCTIONS 

BEFORE COMPLETING FORM 

report number NAVENVPREDRSCHFAC 

Contractor Report CR 84-11 

2. GOVT ACCESSION NO. 

3- RECIPIENT'S CATALOG NUMBER 

4. TITLE (and Subtitle) 


5. TYPE OF REPORT 4 PERIOD COVERED 

Characteristics of North Indian Ocean 
TroDical Cvclone Activity 

Final 



6 PERFORMING ORG. REPORT NUMBER 

7. author^*; 


8. CONTRACT OR GRANT NUMBERfaJ 

Cheng-Shang Lee and William M. Gray 

N00228-83-C-3122 

9 PERFORMING ORGANIZATION NAME AND ADDRESS 

Department of Atmospheric Science 

Colorado State University 

Fort Collins, CO 80523 

10. PROGRAM ELEMENT, PROJECT, TASK 

AREA 4 WORK UNIT NUMBERS 

PE 62759N NEPRF WU 6.2-22 

H CON TROLLING 0 F FICE N AME ANO ADDRESS 

Naval Air Systems Command 

Department of the Navy 

Washington, DC 20361 

12. report oate 

December 1984 

13. NUMBER OF PAGES 

no 

U MONITORING AGENCY NAME & AOORESSf// different from Controlling Office) 

15. SECURITY CLASS, (of this report) 

Naval Environmental Prediction"Research Facility 
Monterey, CA 93943-5106 

UNCLASSIFIED 



15*. CECLASSI FICATI ON/DOWNGRADING 

SCHEDULE 

16. DISTRIBUTION STATEMENT (of this Report) 



Approved for public release; distribution is unlimited. 

17. DISTRIBUTION STATEMENT (of the abetrmct entered in Mleek 20, It different frem Report) 

18. SUPPLEMENTARY NOTE5 

19 KEY WORDS (Continue on reverse eide if neceeeery m^d identify by block number) 

North Indian Ocean 

Tropical cyclone 

Tropical cyclone climatology 

Tropical cyclone structure 

- 

20. ABSTRACT (Continue on reveree eide it neceeeery end identify by block number) 

Characteristics of North Indian Ocean tropical cyclones are discussed 
from a combined climatological, composite, and individual-case perspective. 
Presented are both a general discussion of the monthly climatology of these 
spring/autumn tropical cyclones and a comparison of their structure, genesis, 
intensity, and movement with these qualities of northwest Pacific Ocean 
tropical cyclones. 

((continued on reverse)) 

1 JAN 73 1473 EDITION OF 1 NOV 65 IS OBSOLETE 

S/N 0 10 2-0 14- 660 1 | 

UNCLASSIFIED 


SICUFITY CLASSIFICATION OF THIS PAGE (When Dete Entered) 




































SECURITY CLASSIFICATION Of THIS PAgerWTitn gjt* Erutrtd) 


Block 20, Abstract, continued. 

Detailed individual case analyses are made of each North Indian Ocean 
tropical cyclone that occurred during the First Global GARP Experiment (FGGE) 
period. Each tropical cyclone's characteristics from genesis to decay are 
discussed. All data sources of the FGGE year period were consulted, including 
the analysis of the European Centre for Medium Range Weather Forecasts (ECMWF). 


UNCLASSIFIED 


SECURITY CLASSIFICATION OF THIS PAG EfW7i#n Dm (m Entmrmd) 





hN 

i x ) 

AD—A153 085 

f3 

42) 

040200 

C X 

U3) 

<U) 

CA 

5) 

COLORADO STATE UNI 
SCiENCE 

T X 

t6) 

C ia r a c t e r i s 1 . 1 c s o t 
Activity. 

TC 

18 ) 

(U) 

DN 

t 9) 

F i n a i c o n t r a c t. o r r 

AU 

t 10) 

Lee,C.. S. 

AU 

« 1 0 ) 

Gray, W. n. 

RD 

( 1 1 > 

Dec 1984 

PG 

i 12) 

114p 

N00228 - S 3 - C -• 3122 

CT 

K IS) 

Rn 

(18) 

NEPRF—CR-S4-1I 

RC 

( 20 ) 

Unclassitied repor 

DE 

(23 ) 

*TROP1CAL CYCLONES 
CLIMATE 

DC 

(24) 

(0) 

ID 

(25) 

F06E1First G1obai 

1C 

(26) 

(U) 

A B 

< 27) 

Characteristies o+ 


? r U K I C 0 L_ IN S D £ P1 ’ Q F A T i v i G S F* hi fc£ RIC 
t'M o r t h I n d i a n G c e a n T r o p i c a 1 C y c 1 o n e 


reDt. 


HU 

l)t„ 

SE 

cc 


( 28 ) 
4 33 ) 
4 34 ) 
( 35 ) 


1ND1AN OCEAN, NORTH<DXRECTION) 


G A R p £ x p e r i m e n t ) , P E6 2 759N, W U6222 

N o r t h I n cl i a n 0 c e a n tropic a I c y c 1 o n e s 

a r e d :i. s c u s s e d f r o m a c o rn d x n e d c 1 1 m a t o I o q i c a i , 
c: om p os 1 1 €? , a n d i n d i v i d u a 1 — c as e p er sp ec t: I ve.. Pr ese n t ed 
a r e D o t h a q e n e r a i d :i s c u s s i o n o + t h e m o n t hi 1 y 
c 1 11 V i a t o I o q y “ o f t h e s e s p r i r» q / a u t u m n t r o p i c a X c y c X o n e s 
and a campar i son o*f tnei r s tr ucat r e „ qenesi s» 

i n t. e n s 1t y „ a n d m o v e m e n t w :i. t h t. h e s e q u a X 1 t 1 e s o f 

northwest Pac11 1 c Ocean trap x cai cvcX ones* Detaieo 
x n d i v i d u aX c a s e a n a .i. v s e s a r“ e m a d e o f e a c h« N o r t n X n d i a n 
0 c e a n t r o q i c a x c y c x o n e t h a v; o c c: u r r e d d u. r i n q t i n e F i r s t 
G X o D a X G A R P £ x p e r i rn e r 1 t (F G Ghi > a b r i o d» E a c h t r o p i c a X 
c y c x o n e ‘ s c h a r a c t e r x s t x c s t r o m q e n e s x s c o cl e c a y a r e 
d .L s c u s u e d « A X X d a l: a s o u r c e s* o f t n e r G G £ y e a r p e r x o c; 
were consul tedincluding the analysis o-f the European 
C e n c r e t a r ri e d x u rrt R a n q e 2 e a t r i e r r o r e c a u t s £ ...2 fl U h ) * 

(u > 

0 1 

F 

0hi83 1 yj 



NAVENVPREDRSCHFAC CR 84-11 


NAVENVPREDRSCHFAC 
CONTRACTOR REPORT 
CR 84-11 



CHARACTERISTICS OF NORTH INDIAN OCEAN 
TROPICAL CYCLONE ACTIVITY 


Prepared By: 

Cheng-Shang Lee and William M. Gray 

Colorado State University 
Fort Collins, CO 80523 


Contract No. N00228-83-C-3122 


DECEMBER 1984 



APPROVED FOR PUBLIC RELEASE; DISTRIBUTION IS UNLIMITED 


Prepared For: 

NAVAL ENVIRONMENTAL PREDICTION RESEARCH FACILITY/ 
MONTEREY, CALIFORNIA 93943-5106 






TABLE CF CONTENTS 


Page 

1. INTRODUCTION . 1 

2. TROPICAL CYCLONE CLIMATOLOGY. 5 

2.1 Yearly Seasonal and Monthly Frequency . 5 

2.2 Favorable Environmental Setting for Cyclone Formation . 14 

3. SYNOPTIC ENVIRONMENT IN WHICH NORTH INDIAN OCEAN TROPICAL 

CYCLONES FORM. 24 

4. DATA SOURCES FOR CASE STUDY ANALYSIS AND CYCLONE 

STRUCTURAL DEFINITIONS. 3 0 

5. SUMMARY OF N.I.O. TROPICAL CYCLONES DURING THE FGGE YEAR 

AND THEIR BASIC LARGE-SCALE CIRCULATION PATTERNS. 3 5 

6. ANALYSIS OF INDIVIDUAL CASES OF NORTH INDIAN OCEAN TROPICAL 

CYCLONES DURING FGGE. 41 

6.1 TC 17-79. 41 

6.2 TC 18-79. 53 

6.3 TC 23-79. 65 

6.4 TC 22-79. 76 

6.5 TC 24-79. 80 

6.6 TC 25-79. 83 

6.7 TC 26-79. 87 

7. DISCUSSION. 90 

7.1 Characteristics of the Large-Scale Circulation 

Patterns During Tropical Cyclone Formation and 

Development. 90 

7.2 Comparison of the Structural Characteristics of N.I.O. 

Tropical Cyclones with Western North Pacific 

Tropical Cyclones. 96 

7.3 Characteristics of the Motion Characteristics of N.I.O. 

Tropical Cyclones. 100 

8. ACKNOWLEDGEMENTS. 10 3 

9. REFERENCES. 10 4 

DISTRIBUTION . 107 


i 
























LIST CF ACRONYMS AND SYMBOLS 


ATR 

ECMWF 

FGGE 

ITCZ 

JTWC 

NCAR 

NEPRF 

NHC 

N.I.O. 

NOAA 

NMC 

SMS 

V t 

V 

r 

V 

max 

TO 

*t 

V 

r 


Annual Typhoon Report 

European Center for Medium-range Weather Forecasts 
First Global GARP .Experiment 
Inter Tropical Convergence Zone 
Joint Typhoon Warning Center, Guam 

National Center for Atmospheric Research, Boulder, CO 
Naval Environmental Prediction Research Facility 
National Hurricane Center, Miami, FL 
North Indian Ocean 

National Oceanic and Atmospheric Administration 
National Meteorological Center, Washington, D.C. 
Synchronous Meteorological Satellite 
tangential wind 
radial wind 

maximum low-level sustained wind in a cyclone 
tropical cyclone 

mean tangential wind at 6° radius about a tropical 
disturbance of storms between 1000-700 mb levels 

mean radius outflow at 6° radius about a tropical 
disturbance or storm between 150 and 250 mb 


ii 


1. INTRODUCTION 


Growing international tensions over the Arabian Sea shipping routes 
and the recent acceleration of US-Indian research activity on the Indian 
sub-continent monsoon circulations have heightened the interest in the 
characteristics of North Indian Ocean (N.I.O.) tropical cyclones. How 
do tropical cyclones of the N.I.O. compare with the tropical cyclones of 
other global basins as to their genesis, intensification, structure, 
movement, and decay characteristics? What influence do N.I.O. tropical 
cyclones have on the onset and retreat of the monsoon? These are 
difficult questions to answer because the tropical cyclones of this 
region have not been as well observed as the tropical cyclones of the 
northwest Pacific or those of the western Atlantic for the following 
reasons: 


1 . Although routinely done in the Atlantic and the northwest 
Pacific for the last 35 years, there have been no aircraft 
reconnaissance flights into N.I.O. tropical cyclones. To obtain 
intensity or maximum wind estimates of N.I.O. cyclones one must 
rely on the Dvorak (1975) satellite technique for estimation. 
Before the satellite era one estimated cyclone strength as best he 
could from the surrounding wind and surface pressure information. 
When the Dvorak scheme is used by inexperienced meteorologists and 
when cyclones have a-typical structural features that do not well 
fit the Dvorak intensity curve, this very excellent technique can 
lead to errors in estimates of maximum wind (as occurred when the 


2 


technique was tested in other ocean basins). In addition, as will 
be discussed later, a maxinun wind determination in a cyclone nay 
not be a very good estimate of the net damage or flooding potential 
of the cyclone. Damage and flooding can, in many instances, be 
more related to the cyclone's outer wind strength than to its 
maximum wind or minimum central pressure. Weatherford and Gray 
(1984) have recently been making special studies of the 
relationships between tropical cyclone intensity (maximum wind) and 
cyclone strength (mean tangential wind over 1-3° radius about the 
cyclone) from analysis of western Pacific US Air Force weather 
reconnaissance flights. As will be discussed, we found that 
cyclone intensity and strength are often times not well related to 
each other. 

Despite these difficulties, we have no other choice but to use 
the Dvorak satellite technique for an estimate of N.I.O. cyclone 
intensity. In a general sense, this technique, as it has been 
applied over the last 7-8 years, has proven itself to be quite 
reliable, especially in the qualitative sense concerning change in 
cyclone intensity. 

2. Synchronous meteorological satellite (SMS) data have not been 
available over the Indian Ocean. In the Atlantic we have had SMS 
data since 1967 and in the western Pacific since 1978. SMS 
satellite data have been a very valuable source of tropical cyclone 
information. We are fortunate that SMS data were available for a 
portion of the FGGE year in which we analyzed individual N.I.O. 
tropical cyclones. This is one of the reasons we chose the FGGE 
year for our individual case tropical cyclone analysis. 


3 


3. Island and land stations to the equatorial side of tropical 
cyclones in the N.I.O are mostly absent as compared with the 
western Pacific and the western Atlantic. This hinders the proper 
analysis of Southern Hemisphere influences, which later in this 
study will be shown to be very important for the formation and/or 
intensity change of many N.I.O. tropical cyclones. 

Although a number of climatological studies and other individual 
case studies of tropical cyclones in the N.I.O have been made, there 
have not been, to our knowledge, any comprehensive studies of the 
combined climatological, .synoptic setting and structure of individual 
tropical cyclones in the region, or a comparison of N.I.O tropical 
cyclones with the better documented tropical cyclones of other ocean 
basins. This report attempts to make this more comprehensive study of 
N.I.O. tropical cyclones. We have made a detailed analysis of all the 
tropical cyclones that formed in the N.I.O. during the FGGE year. More 
reliable information exists on N.I.O cyclones during the 1979 FGGE year 
than at any other time because of the special observing systems and 
analysis procedures which were established for that year alone. This 
included the very unique and well archived European Center for Medium- 
Range Forecasts (ECMWF) FGGE year analysis of this region. We are also 
making much use of the ECMWF analysis to study tropical cyclones in the 
other global storm basins (see Lee, 1984; Chen and Gray, 1984; and Askue 
and Lee, 1984). We have found the ECMWF objective analysis to be very 
reliable and most useful for the understanding of tropical cyclones in 
the northwest Pacific and Atlantic. Comparisons using independent 
rawinsonde analysis (to be shown later) have shown the ECMWF analysis to 


be quite reliable — especially the large-scale rotational or tangential 
component of the wind fields and the general wind analyses at lower and 
upper tropospheric levels where SMS satellite wind vectors and/or 
aircraft data are available in large numbers. 

This study does not treat the more numerous tropical depressions 
that form in the Bay of Bengal during the summer season (Saha eit al. . 
1981). 

We hope that this report will give the reader a better overall 
understanding of N.I.O. tropical cyclones that form almost exclusively 
in the late spring and autumn seasons. 

We will first discuss the long-term climatology of N.I.O. tropical 
cyclones, then discuss the typical environmental flow features in which 
these cyclone systems typically form. A detailed analysis of each 
cyclone system by itself will then be given. We will also compare 
N.I.O. cyclones with northwest Pacific tropical cyclones and discuss the 
distinctive features of N.I.O. tropical cyclones. 


5 


2. TROPICAL CYCLONE CLIMATOLOGY 

2.1 Yearly, Seasonal and Monthly Frequency 

Unlike tropical cyclone formation in the other ocean basins 
tropical cyclones of the N.I.O. form in the late spring and autumn 
seasons. The summer is mostly devoid of tropical cyclones although many 
monsoon depressions form in the northern part of the Bay of Bengal 
during the summer period of June through September. But very few of 
these monsoon depressions, ever intensify to cyclones of tropical storm 
strength. This report deals mainly with tropical cyclones that develop 
sustained low-level wind velocities and were reported as significant 
tropical cyclones by JTWC or Indian Meteorological Agency. 

Table 1 gives a yearly listing of the number of N.I.O. tropical 
cyclones and hurricanes that have been reported since 1958. Generally, 
the yearly variation of cyclone number is not large. N.I.O. tropical 
cyclones represent about 7-87o of the 80 or so tropical cyclones and 
hurricanes that form per year about the globe. Figure 1 gives the 
tracks of N.I.O. tropical cyclones during the 10-year period of 1968- 
1977. 

Earlier Climatology . A study by Bansal and Datta (1972) has shown 
that from 1891 to 1970 there were 362 cyclonic storms (2 17 m/s) that 
originated in the Bay of Bengal, or about 4.5 storms per year. Another 
study by Mooley (1980) indicates 453 cyclonic storm systems in the 101- 
year period of 1877-1977. These figures are higher than that obtained 



6 


TABLE 1 


Yearly 

number of reported tropical 

cyclones 

in the North Indian 

from Indian Meteorological Agency 

and JTWC. 



Number of reported 




tropical cyclones 


Number of reported 


from Indian Meteor¬ 


tropical cyclones 

Year 

ological Aeencv 

Year 

from JTWC ATR 

1958 

5 

1975 

6 

1959 

6 

1976 

5 

1960 

4 

1977 

5 

1961 

6 

1978 

4 

1962 

5 

1979 

7 

1963 

6 

1980 

2 

1964 

7 

1981 

3 

1965 

6 

1982 

5 

1966 

9 

1983 

3 

1967 

6 



1968 

7 

Average 

4.4 


1969 

6 

1970 

7 

1971 

6 

1972 

6 

1973 

6 

1974 

7 


Average 6.2 


Ocean 


from the JTWC ATR (Annual Typhoon Report) 1975-1983 average of 3.0 
significant tropical cyclones per year in the Bay of Bengal. The 
differences between these figures appear to be due to the use of 
different data sources and the uncertainty in determining the cyclone 
intensity. The previous two studies utilized the data collected by the 
Indian Meteorological Agency, which uses different definitions for the 
intensity of tropical cyclones (see Table 2). On the other hand, JTWC 
uses different criteria for issuing tropical cyclone warning. For 
example, Typhoon Hope (1979) moved westward across the south China Sea 
and SE Asian Peninsula and into northeast Bay of Bengal in August, 1979. 








7 



if) 


] 

[ : 



; ! 

■ : : 

;rll inr : : -j 

y 



t 5 J* - * 

r, 

_ 

i \ 


Fig. 1. Typical tropical cyclone tracks in the North Indian Ocean for 
the 10-year period 1968-1977. 


TABLE 2 


Indian Meteorological Department classification of tropical disturbances 
in the Bay of Bengal. 


Classification 


Range of Wind Speeds (m/s) 


Low 

Depression 

Deep depression 

Cyclonic Storm 

Severe Cyclonic Storm 

Severe Cyclonic Storm (Hurricane) 


< 8.5 


8.5 - 

13.5 

14.0 - 

16.5 

17.0 - 

23.5 

24.0 - 

31.5 

32 




















































8 


The JTWC ATR did not report it as a tropical cyclone in the N.I.O., but 
the Indian Meteorological Agency (as discussed by De Angelis, 1980c) 
did. 

We believe that the Indian Meteorological Office sometimes listed 
Bay of Bengal monsoon disturbances as tropical cyclones when (by JTWC 
criteria) they were not. In other words, JTWC rarely warned on systems 
that originated in the northern Bay of Bengal and could possibly be 
interpreted as monsoon depressions. A comparison of the monthly 
frequencies of tropical cyclones reported by JTWC and the Indian 
Meteorological Agency (Tables 3 and 5) reveal that the largest 
difference occurs during .summer (June-August) or the most active period 
for monsoon depression. 

Bansal and Datta's (1972) tabulations indicated that there were 98 
cyclonic storms in the Arabian Sea region during the 80-year period of 
1891-1970. The average number was 1.2 storms per year. This is not 
much different from the recent year value of 1.4 cyclones per year which 
have occurred over the last nine years (as monitored by JTWC). The 
difference between the number of tropical cyclones in the Bay of Eengal 
and Arabian Sea by Bansal and Datta's estimate was 3.75 to 1. Recent 
year information places the figure at about 2 to 1. These differences 
are due to the new case count of tropical cyclones in the Bay of Bengal 
being decreased. 

Among the 362 cyclonic storms (or systems with V 2 17 m/s) in 

ID 8.X 

the Bay of Bengal as listed by Bansal and Datta, 132 or 36% were 

reported to be of severe cyclonic storm strength (V 2 24 m/s). In 

max 

the Arabian Sea- 53 of 98 or 54% were so classified. These numbers agree 
qualitatively with those reported by JTWC, in terms of the numbers, 


9 


Arabian Sea tropical cyclones (although fewer in number) are, in 
general, somewhat stronger than Bay of Bengal tropical cyclones. 

Table 3 gives a summary by month of the distribution of tropical 

cyclones (V max 2 17 m/s, and V max > 24 m/s), in the N.I.O. during this 
80-year period. November and October are the most active and second 
most active tropical cyclone months in both the Bay of Bengal and 
Arabian Sea. Hay is the second most active month for the more intense 
Arabian Sea cyclones, however. The large numbers of storm systems in 
the Bay of Bengal in July and August are believed by us to be mostly 
monsoon depressions that occurred in the northern Bay of Bengal. 


TABLE 3 

Number of cyclonic storms in the N.I.O. during 1891-1970 with intensity 
(V max > 17 m/s) and the number (in parentheses) that were estimated to 
be severe storms (V max > 24 m/s). Data from Bansal and Datta, 1972. 



JAN 

FEB 

MAR 

APR 

MAY 

JUN 

JUL 

AUG 

SEP 

OCT 

NOV 

DEC 

TOTAL 

Bay of 

Bengal 














All Storms 

5 

1 

4 

19 

39 

35 

38 

26 

32 

62 

68 

34 

3 62 

(Severe 

Storms) 

(2) 

(1) 

(2) 

(8) 

(26) 

(4) 

(7) 

(1) 

(10) 

(26) 

(32) 

(13) 

(132) 

(36%)* 

Yearly 
Average of 
All Storms 

0.1 

0.0 

0.1 

0.2 

0.5 

0.4 

0.5 

0.3 

0.4 

0.8 

0.9 

0.4 

4.5 

Arabian 

Sea 














All Storms 

2 

0 

0 

5 

16 

15 

3 

2 

5 

19 

26 

5 

98 

(Severe 

Storms) 

(0) 

(0) 

(0) 

(4) 

(13) 

(9) 

(0) 

(0) 

(1) 

(7) 

(18) 

(1) 

(53) 
(54%)* 

Yearly 
Average of 
All Storms 

0.0 

0.0 

0.0 

0.1 

0.2 

0.2 

0.0 

0.0 

0.1 

0.2 

0.3 

0.1 

1.2 

♦Percentage of 

the 

number 

of all storms 

that 

are 

severe. 


















10 


In the Arabian Sea regions, only 5 systems formed in the July- 
August period and none of them attained severe storm intensity. In the 
spring active period (April-June), only 28% of the storm systems 
reported by Bansal and Datta (36 out of 129) formed in Arabian Sea 
region. But of this 28% of systems which formed in the Arabian Sea, 72% 
(or 26 out of 36) attained severe cyclonic storm intensity. A summary 
of storm numbers during these two active seasonal periods in both basins 
is shown in Table 4. It is also interesting to note that depressions 
can form in the Bay of Bengal in almost every month but only those 
depressions that form in the active monsoon transitional periods 
intensify into severe storms and hurricanes. 

TABLE 4 

Number of tropical storms (1891-1970) during each of the two active 
season periods and the percent that are severe cyclonic storms (i.e., 
systems of tropical cyclone or greater intensity). 


Apr-Jun Sep-Dec 

Bay of Bengal 

All Cyclones (V > 17 m/s) 93 196 

incix 

All Cyclones (V > 24 m/s) 38 (41%) 111 (57%) 

Ula.X 

Arabian Sea 

All Cyclones (V > 17 m/s) 36 55 

Ula.X 

All Cyclones (V > 24 m/s) 26 (72%) 27 (49%) 

IflciX 


Figure 2 shows, by three month intervals, 20 years of origin points 
of significant tropical cyclones. Note the concentration of formation 
in the spring (April-May-June) and autumn (October-November-December) 
months. In comparison with the western North Pacific region, the N.I.O. 
has as high or higher number of tropical cyclone formations per unit 








11 



60‘Nr 


2G*E 40* 60* 9C- 0C* >2C* 6C # 


JAN - FEB - MAR 
CYCLONE ORIGIN LOCATIONS F 


i F 


20 YRS. 



40* 


eo* 


20*E 40* 

60*N 


leo- 


Fig. 2. Origin points by three month intervals for a 20-year period of 
significant tropical cyclones. Some of the original locations 
in the northern Bay of Bengal during July- August-Sept ember 
can be classified as monsoon depressions. 























12 



60*N 


40* 


2CTE 


oc- 


i20* 




> 60 * 


180 * 


JUL - AUG - SEP 
CYCLONE ORIGIN LOCATIONS F 


20 YRS. 



GOTH 


ar t 4p* ecr ecr txr go* wr __ >ecr ecr 

OCT - NOV - DEG_, O 
CYCLONE ORIGIN LOCATIONS Fjdtk 20 YRS.®°* 


<o* 


Fig. 2. Continued. 





















13 


area as the Pacific. Differences in total numbers of storms are more a 
result of the larger Pacific formation area. 

M.I.G. Tropical Cyclones Monitored by JTWC in Decent Years . Before 
1974 the Guam Joint Typhoon Warning Center (JTWC) area of responsibility 
did not include the Arabian Sea region. Valid statistics using 
satellite information and the Dvorak intensity classification (1975) are 
available only since 1975. By JTWC standards, using the Dvorak 
classification scheme, the number of tropical cyclones in the N.I.O. is 
about 4.4 per year. This is about 1.3 fewer storms per year than 
indicated by previous Indian estimates. This is, as has been discussed, 
a consequence of the Indian Meteorological Agency classifying quite a 
few monsoon depressions as cyclone systems during the summer period. 

Table 5 shows the number of tropical cyclones that occurred in each 
month during this nine year period (1975-1983). The second line lists 
the yearly averaged number of tropical cyclones in each month. As 
expected the autumn transition season was the most active period for 
tropical cyclone formation. 


TABLE 5 

Comparison of the number of significant tropical cyclones during 1979 
with the number of tropical cyclones during the 9-year period 1975-1983 
as determined by the JTWC. 

JAM FEB MAR APR MAY JUN JUL AUG SEP OCT MOV DEC TOTAL 

Morth 























14 


Among these 40 systems of tropical cyclones, 27 or about two- 
thirds formed in the Bay of Bengal region and 13 (or one-third) in the 
Arabian Sea. Two systems formed in the Bay of Bengal but traveled into 
the Arabian Sea and maintained their tropical storm intensity (maximum 
sustained winds 2 35 kts) in both basins. Among the 27 tropical 
cyclones that formed in the Bay of Bengal, 6 (or 22%) were estimated to 
have attained typhoon intensity (2 65 kts) . The average maximum 
intensity of these 27 Bay of Bengal tropical cyclones was 56 kts. In 
the Arabian sea region, 4 of the 13 tropical cyclones (or 31%) attained 
typhoon intensity. Their average maximum wind speed was 60 kts. There 
was about a two to one difference in the number of tropical cyclones 
that formed in the Bay of Bengal in comparison with the Arabian Sea. 
This is in general agreement with previous estimates. A higher 
percentage of Arabian Sea tropical cyclones appear to form in the 
spring period as compared with the Bay of Bengal. And a slightly 
higher percentage of the cyclones that form in the Arabian Sea reach 
hurricane intensity. 

2.2 Favorable Environmental Setting for Cyclone Formation 

We will now discuss the climatology of the general environmental 
setting in which tropical cyclones can form in the N.I.O. 

Sea-Surface Temperature and Ocean Energy Contents . Warm ocean 
temperatures and the depth to which this warm ocean water penetrates 
into the N.I.O. are quite substantial in all seasons. These warm ocean 
temperatures are a necessary ingredient to tropical cyclone formation. 
Provided that wind and other dynamic factors are satisfactory, tropical 
cyclone formation (at least from an ocean energy point-of-view) can 
occur in any month in the N.I.O. 



15 


Figures 3 and 4 show seasonal values of these parameters. Note how 
the 28°C sea-surface temperature and ocean temperature above 26°C 
extends down to below 60 meters depth in all seasons in the southern 
half of both the Bay of Bengal and in the Arabian Sea. Tropical cyclone 
formation requires that ocean temperatures be above 26°C and that warm 
temperatures extend to a depth of 50-100 meters or so. 

The lack of tropical cyclones in the N.I.O. during the summer 
season is due to dynamical processes associated with the shift of the 
monsoon trough to northern India, not to any reduction in ocean energy 
content. 

Seasonal Cyclone Frequency Related to Seasonal Climatology . As 
previously discussed by the second author (Gray, 1975, 1979) long-term 
seasonal tropical cyclone genesis frequency (in all the global basins) 
is closely related to a combination of six seasonal meteorological 
features which will henceforth be referred to as primary climatological 
genesis parameters. These seasonal parameters are: 

(1) low-level relative vorticity (X ), 

(2) Coriolis parameter (f), r 

(3) the inverse of the tropospheric vertical wind shear (S ) 

of the horizontal wind between the lower and upper Z 

troposphere or (1/S ), 

(4) 'ocean thermal energy' - integrated sea temperature excess 

above 26°C to a depth of 60 m (E), 

(5) vertical gradient of Q between the surface and 500 mb 

oe e /0 P ), 

(6) middle troposphere relative humidity (RH). 

A tropical cyclone Seasonal Genesis Index (SGP) can be defined as: 

SGP — [ (vorticity) x (Coriolis) x (vertical shear) parameters ] x 
t (ocean energy) x (moist stability) x (humidity) parameters ] 



16 




60*N 


40* 


20*E 


40- 


00 - 


20 - 


W0* 


APR - MAY - JUN 
SIA SURFACE TEMP (°C) 


Fig. 3. Seasonal sea surface temperature over the Indian Ocean. 
























17 


2 o*e 40* 60* ecr oc* ec* *cr «o* 




Fig. 3. Continued. 



























60 *N 


«r 


-ZTL 40* 60* 80" 00* 




APR - MAY -JUN 
DEPTH OF 26 0 C ISOTHERM (m 


Depth (in meters) of the 26° isotherm over the Indian Ocean 
by season. 


Fig. 4. 










































19 





Fig 


4. Continued. 























20 


This seasonal tropical cyclone Genesis Index may also be thought of 
as in the form: 

SGP = (dynamical potential) x (thermal potential). 

where 

dynamic potential = [(vorticity)x(Coriolis)x(vertical shear) parameters] 
thermal potential = [(ocean energy)x(moist stability)x(humidity) parameters 

The product of dynamic potential and thermal potential is defined 
as the Seasonal Genesis Index. When expressed in this way with the 
proper units (see original papers) this SGP gives a very good estimate 
of the seasonal tropical cyclone genesis frequency in number per 5° 
latitude-longitude square per 20 years in all seasons and at all global 
locations. 

The very close correspondence between predicted and observed 
seasonal cyclone frequency lends support to this argument concerning the 
relevant seasonal climatological factors which are necessary for cyclone 
genesis. 

Figures 5 and 6 show these dynamic and seasonal potentials and 
their product (the Seasonal Genesis Parameter) for the spring and autumn 
periods of major cyclone formation. The bottom diagram in both of these 
figures shows the observed number of tropical cyclone formations per 5° 
latitude-longitude square per 20 years. This is to be compared with the 
third figure from the top. 

It is seen how closely the seasonal climatology of these six 
parameters dictates the number of cyclones which form in the N.I.O. even 
though the days of cyclone genesis (1-15 per 5° square per 20 years) 


21 




no. per 5° l AT .-LONG, per 2 


60-N 


40* 


2D-E 


40 * 


60 * 


cc- 


GC* 


M3C* 


DYNAMIC POTENTIAL 

/vorticity^x f CaidisV ( V |^A 

iPoometefl iPcrarr.etey Iparcmeier/ 

(inlOsec 2 per m/s per750m 


60*N 


ao*E 40- 


CO* GO- 


' wo* 


t60- 


i ecr 


APR - MAY - JUN 


«o* 


THERMAL POTENTIAL - [E] + JjtfH p 

(units lO’Cal/cm 2 °K pegs pQpjbJ 


or a met 


60*N 


sc* bf ar ao* kg- 




20"E 


APR - MAY - JUN 

SEASONAL \ / \/ 

DAOAiurTCD W DVNAM1C VTHERMAL 
PARAMETER j^OTENTlAM^OTENT^ 


60*N 


tc- 


20-E ^ 


-:-r 


tv." 








APR - MAY - JUN /— w /( 
OBSERVED CYCLONE OR IG I N^R EQ ti E N CY 
NUMBER PER 5° LAT. - LONG.^R 20 YEAf^S 


«0- 8C* 


PC" 


Fig. 5. North Indian Ocean April to June Dynamic and Thermal potentials. 
Seasonal Genesis Parameter and observed number of tropical 
cyclone formations per 5 latitude-longitude square per 
20 years. 



















































22 


Fig. 6 



ecru 


40* * 




ecr 




oc* 








acr 



OCT - NOV - DEC 

SEASONAL \ / \ / 

DAO a HP! Id W DYNAM| C VthERMAL 
PARAMETER nPOTENTiAlJPOTENTIAL 

no. per 5 0 LAT.-LONG. per 2( 
(V.nils k5X10’ 8 cal °K sec' 



60*N 


2D-E 


4 CT 


«r 






OCT * NOV - DEC 
OBSERVED CYCLONE ORIGIN 
NUMBER PER 5° LAT S - LONG, f 


J 3 C - 


CY 

R3 


. Same as Fig. 5 but for October to December. 




















































23 


make up but a very small percentage of the season. It was not expected 
that the agreement between predicted and observed genesis frequencies 
would be this close. Not only is the correlation of the six seasonal 
parameters with genesis frequency noteworthy, but the physical rationale 
concerning the effects of these parameters which also is important. 

There is thus some rational quantitative explanation as to why the 
long-term frequency of tropical cyclones in the N.I.O. is as observed. 

The N.I.O. thermal potential changes little during the spring and 
autumn seasons. It is mainly the day-to-day variations in the dynamic 
potential that must be watched for in predicting the day-to-day 
potential for cyclone formation. This is mostly determined by the 
strength and location of the monsoon trough. The presence of the 
monsoon trough between 5-15°N latitude assures that upper tropospheric 
winds will be such that easterly zonal winds will prevail equatorwards 
of the monsoon trough and westerly zonal winds on the poleward side of 
the trough. 


3. 


SYNOPTIC ENVIRONMENT IN WHICH NORTH INDIAN OCEAN TROPICAL 
CYCLONES FORM 


Tropical storms form in the monsoon transitional seasons of spring 

and autumn in the N.I.O. because it is only during these periods that 

the monsoon trough is located sufficiently south (~ 5-15°N) that a broad 

oceanic area is available for cyclone development. During the summer 

periods of late June through early September, the monsoon trough is 

typically located over the Ganges Valley and at the head of the Bay of 

Bengal. Vertical wind shear over the Bay of Bengal and Arabian Sea are 

too strong to permit tropical cyclones in these seasons. As discussed 

by the second author (Gray, 1968, 1975), tropical storms form only when 

the tropospheric vertical wind shear over the incipient disturbance is 

very weak. Such small wind shear conditions typically occur only during 

the monsoon transitional periods of late April through June, and late 

September through early December. It is at these times that the 

tropospheric vertical shear of zonal wind (U) or (If. .-U„,« ) 10° to 

200mb 850mb 

the poleward side of the monsoon trough is strongly positive while 10° 
to the equatorwards side this shear is strongly negative. The change of 
sign of this tropospheric zonal wind shear from south to north assures 
that a region of zero or minimum shear exist near the monsoon trough 
axis. 

Figure 7 shows the type of tropospheric vertical wind shear 
conditions required for tropical cyclone formation in the N.I.O. 

^200mb~^850mb must change sign as one proceeds northward from the 
equatorward to the poleward side of the monsoon trough. The greater the 


25 




l -i- 

Eq. fO°M 

LAT1TUC€ 


20°N 



Fig. 7. North-south cross sections of the typical locations of various 
classes of tropical cloud clusters relative to the monsoon 
trough, and the usual zonal winds present with these systems 
(top diagram). The bottom diagram portrays the vertical 
distribution of typical zonal wind velocity, U, occurring with 
different, A to D, types of cloud clusters whose general location 
is specified in the top diagram. The usual zonal distribution 
or cluster velocity is designated U. and the difference in 
cloud cluster and environmental wind velocity at any level is 
given by U-U d< 


magnitudes of 200 to 850 mb zonal wind shear at points B and D (as 
indicated in both figures) the greater is the strength of the monsoon 
trough and the greater is the potential for cyclone formation. Tropical 
cyclones, in general, do not form in weak monsoon trough conditions or 
when wind shear conditions at points B and D are weaker than normal. 

























26 


As will be discussed in Chapter 6, all seven tropical systems that 
developed in the N.I.O during the 1979 FGGE year exhibited this type of 
vertical wind shear about the monsoon trough. In general, the potential 
for tropical cyclone formation increases the larger the tropospheric 
vertical wind shear is about 6° north and about 6° south of the 
incipient disturbance. See previous CSU research by McBride, 1981; 
McBride and Zehr, 1981 for more discussion of this genesis criterion. 
This criterion implies that the vertical gradient of vcrticity between 
850 mb and 200 mb around the incipient pre-cyclone disturbance at 6° 
radius be as negative as possible. The more negative this value is, the 
higher the potential for.cyclone development. The most important 
ingredient for tropical cyclone development is not just the amount and 
intensity of Cb convection associated with the tropical disturbance, but 
rather the existence of the tropical disturbance in a favorable strong 
monsoon trough environment. If the surrounding environmental wind 
conditions through the troposphere are not favorable then tropical 
cyclone formation will almost never take place irrespective of the 
amount and intensity of deep convection currently occurring. 

Figures 8 and 9 show a composite of the 850 mb and 200 mb zonal 
wind patterns surrounding the initial location positions of 54 N.I.O. 
tropical disturbances which later developed into tropical cyclones. 

Mote that the 850 mb westerly winds south of these disturbances are 
typically stronger than the trade winds to the north. Also note that 
200 mb easterly winds are blowing over the disturbance center. Figure 
10 shows the composite of 850 mb to 200 mb zonal wind shear patterns 
surrounding these pre-tropical storm cases. Note the westerly 
tropospheric wind shear to the north and the easterly tropospheric wind 


27 


shear to the south of these early stage developing systems. Also the 
near zero zonal wind shear near the center of these early stage 
developing systems. 

These are very typical of the lower and upper tropospheric wind 
patterns which are associated with tropical cyclones which develop 
within the monsoon trough in this and at other global locations (Gray, 
1968; Holland, 1984a,b.c). 


N 


V 
/ 

/ 


/ 

/ 




X 

\ 

3(79) \ 


\ 


4(56) 


\ 


64 

<3> 


/ 


2150 


/ 


\ 

2(52) \ 


16(19) ' 


/ 

% 

I / 'n 

\ / 

A--. 

\ ^ ; 

,'\ 6(44) — 10(4 

N. / \ - ) 'N 

\/ 1(34) \ HO / 

/ "f /\ < yy -7 

_ / / 3<2SN / ''1302)/ - 

! 0 — 1 583) 

W ! *> I ( 

> / rc \ 

/g(5) /_N^'2I(I3) 


12* 




- y \ 7io 

A X 14(23) "x 

' \ —* / -s 

\ / 

\x x N I NO AN OCEAN 
\ ZONAL WIND (U) 
850 mb 
Knott 


Fig. 8. Composite of average rawinsonde information in each area 
relative to the initial center positions of 54 tropical 
disturbances which later developed into tropical storms 
between 1955 and 1966. Length of arrows is proportional to wind 
speed in knots (at left). Values in parentheses are number of 
wind reports in each area average. Distance from the center is 
given by the lightly dashed circular lines at 4° latitude 
increments (from Gray, 1968). 


28 



200 mb 
Knof* 


Fig. 9. Same as Fig. 8 but for the 200 mb level. 




29 


N 


\ 

\ / '‘*™ " " 

A 

/ \ y \ 

/ «£» "v( v\ j_j _ . 

— \KX32/ 4.^512^ \ 

*., I ww) i 21SL '1 $*(^£'#'*1 I Thz«) i i3(-m) * _ 

W ~ ~ i “ i “ E 

yiTlrg^ /—i I 

r r - ^Z7 I ?g I 1 _^>2?(!5r 

\ -^ 0 --^> / •£!£! ?-20 

\ ^ X-y&ilL^ r7\ / 

- ^5?r w«). N <7TBir a^ y 

\ /— \ 


-40 


;< 


-b i 



N. INOIAN OCEAN 
VERTICAL SHEAR OF ZONAL WINO 

* U 200rri U B50(rJ' KnOt ’ 


Fig. 10. 


Same as Fig. 8 but for the 200 mb minus 850 mb shear of zonal 
wind. 








4. DATA SOURCES FOR CASE STUDY ANALYSIS AND CYCLONE STRUCTURAL 

DEFINITIONS 

The data sources used in this study's analysis of individual N.I.O. 
tropical cyclones are the FGGE Ill-b data of the European Center for 
Medium-Range Weather Forecasts (ECMWF) and the National Oceanic and 
Atmospheric Administration (NOAA) polar-orbiting satellite (TIROS) 
picture. These satellite imageries are available on mosaics twice a day 
in both IR and visible images for most of the days during the period 
that tropical cyclones were present. The Annual Typhoon Report (ATR) 
published by the Joint Typhoon Warning Center (JTWC), Guam was used to 
determine the position and the intensity of our tropical cyclone cases. 

A supplementary data source was the flow features and satellite pictures 
during the summer MONEX (Quick look ''Summer MONEX Atlas'', Part I Saudi 
Arabia phase) edited by Florida State University (Krishnamurti, et al.. 
1980). These data are used for TC 17-79 (the most intense of the North 
Indian Ocean systems) because of the missing TIROS data during this time 
period. 

The FGGE Ill-b data are available on plotted and analyzed charts in 
book form published by ECMWF in 1981 and on larger maps which were 
specially duplicated and purchased from the ECMWF facility. The 
original FGGE Ill-b data are also available on 1.875 degree 
latitude/longitude cartesian grids at 15 pressure levels. These data, 
including zonal and meridional wind (U,V), height, temperature, vertical 
motion, and specific humidity, are packed and contained in 82 Terabit 


31 


Memory volumes at the National Center for Atmospheric .Research (NCAR). 
Full utilization was made of these archived data sources. 

The FGGE level Ill-b data used in this study are the raw initial 
values of the horizontal zonal and meridional components. These data 
were extracted directly from the initial analysis of the pressure 
levels. They have not been subjected to any vertical interpolation. 

The analysis procedure used a primitive equation forecast for the first 
guess, but after the initial analysis step, no model initialization took 
place. For a complete description of the ECMWF analyzed FGGE Ill-b 
data, please refer to the FGGE data management manuals or the FGGE 
newsletter No. 1 (May, 1983). 

For tropical cyclone studies, a cylindrical coordinate system is 

preferable. The FGGE data, which were initially on cartesian grids, 

have been linearly interpolated onto cylindrical grids. There are 13 

radial and 16 azimuthal grid segments. The grid spacing is 1° latitude 

radius (starting from 2° out from the cyclone center) and 22.5° azimuth. 

The cylindrical grid point data of zonal and meridional wind components 

(U, V) are converted to radial and tangential wind components (V , v ) 

r t 

with respect to the cyclone center. The radial wind is then mass 
balanced vertically from 100 mb to 1000 mb at every radius. The 
pressure levels used in this study are 100, 150, 200, 250, 300, 400, 

500, 700, 850, and 1000 mb. However, due to their better quality, the 
low-level and upper-level data are more extensively used. (There are 
many more surface observations than there are rawinsonde data. Many 
satellite winds are used in the analysis). 

By comparing the FGGE analysis with our independent rawinsonde 
data, Lee (1984) has previously shown that the FGGE data can reasonably 


32 


well represent the tropical cyclone circulation, especially beyond the 
4-5° radius. 

Laver Considerations . We have made much use of the ECMWF analysis 
of the outflow layers of those cyclone systems. This analysis was based 
on the combined satellite, rawinsonde and aircraft jet wind vectors 
taken during the FGGE year. Due to the uncertainty of the height of the 
satellite cloud vector wind, we considered it to be more appropriate to 
average the 150 mb, 200 mb (weighted by a factor of 2) and 250 mb wind 
vectors together to represent the upper level flow. This is an 
important thing to do because of different heights of the maximum 
outflow. Equatorial outflow layers are typically 25-50 mb higher than 
poleward outflow layers. By averaging the 150, 200 and 250 mb layers 
together we obtain a measure of the outflow which can be compared at all 
azimuths. 

On the other hand, we considered the low-level circulation to be 
represented by the layer between 700 mb and 1000 mb. Low-level data are 
considerably better due to the larger number of surface observations. 

Definitions of Cyclone Structure . In discussing the 
characteristics of tropical cyclones and comparing one cyclone with 
another it is important to have methods of structural classifying 
besides those of just maximum wind and central pressure. We have 
recently established a new method of defining the structure of tropical 
cyclones. It is based on the strength of the tangential wind at 
different radii. We defined the cyclone structure by the low-level 
tangential wind characteristics in four radial bands: 




33 


(1) Intensity - highest sustained low-level wind speed between 

0-1° radius. (Intensity of the North Indian Ocean 
tropical cyclone listed in JTWC ATR is estimated by 
Dvorak scheme, 1975). 

(2) Strength - mean low-level tangential wind between 1-3° radius. 

(3) Outer Circulation - mean low-level tangential wind between 

3-7° radius as measured from ECMWF and other analysis. 

(4) Environment - mean low-level tangential wind between 7-15° 

radius as measured from ECMWF and other analysis. 

One of the major points made by our studies of tropical cyclones in 
the northwest Pacific and Atlantic is that changes in one cyclone 
structural component (such as strength) do not necessarily imply changes 
in another structural component such as intensity. Figure 11 shows this 
relationship schematically. An increase in intensity does not require 
that the cyclone strength (or outer circulation) increase 
proportionately. In fact, the strength (outer circulation) may 
increase, decrease, or stay the same for a given change in intensity. 

In terms of a tropical cyclone's net rainfall and its inland destruction 
potential, it is likely that cyclone strength and, to some extent, outer 
circulation are as important or more important than cyclone intensity. 






TANGENTIAL 
WIND SPEED 


34 



i i-1-1-1-1-1-1-1-r~ 

0 I 2345 6 789 

RADIUS (° Latitude) 

Fig. 11. Definition of tropical cyclone intensity, strength and outer 
circulation and their changes on a radial profile of 
azimuthally averaged tangential wind. 









35 


5. SUMMARY OF N.I.O. TROPICAL CYCLONES DURING THE FGGE YEAR AND THEIR 

BASIC LARGE-SCALE CIRCULATION PATTERNS 

During the 1979 First GARP Global Experiment (FGGE) year, the JTWC 
observed seven tropical cyclones forming in the North Indian Ocean. 

This is higher than the averaged number of four and one-half which JTWC 
observed during the period of 1975-1983. Only one 1979 cyclone system 
attained typhoon intensity. This is slightly lower than the yearly 
average of 1.3. Table 6 gives a summary of the 1979 N.I.O. tropical 
cyclone systems. Figure 12 shows a summary of the storm tracks. The 
monthly distribution of these cyclone systems has already been presented 
in Table 5. As in most other years the most active storm period was the 
autumn. 

Among the seven tropical cyclones in 1979, three, that is TC 17-79, 
TC 18-79, and TC 23—79 attained an intensity level whereby their maximum 
sustained winds (or ^ max ) was greater than 50 kts. TC 25-79 reached an 
intensity of 40 kts. It persisted as a tropical cyclone for two days. 

TC 24-79 attained an intensity of 35 kts and maintained this intensity 
for 12 hours shortly before it reached land. TC 22-79 and TC 26-79 did 
not quite reach an intensity level of 34 kts. 

Generally speaking, these weaker tropical cyclones are relatively 
short-lived compared to those in other ocean basins, especially in the 
northwest Pacific and northwest Atlantic. This is because tropical 
cyclone formation in' the N.I.O. typically occurs close to land. The 
short life of some systems makes it a little more difficult to ascertain 


TABLE 6 

NORTH INDIAN OCEAN 
1979 FGGE YEAR TROPICAL CYCLONES 


CYCLONE 

DESIGNATION 


PERIOD 

OF 

WARNING 

CALENDAR 
DAYS OF 
WARNING 

MAX 

SUS 

WIND 

kts 

EST 

MIN 

SLP 

mb 

NUMBER 

OF 

WARNINGS 

DISTANCE 
TRAVELLED 
(n mi) 

TC 17-79 

06 

MAY-12 

MAY 

7 

85 

967 

26 

1267 

TC 18-79 

18 

JUN-20 

JUN 

3 

50 

985 

12 

581 

TC 22-79 

21 

SEP-23 

SEP 

3 

25 

1000 

10 

694 

TC 23-79 

21 

SEP-25 

SEP 

5 

55 

980 

14 

1108 

TC 24-79 

29 

0CT-01 

NOV 

4 

35 

995 

13 

720 

TC 25-79 

16 

NOV-17 

NOV 

2 

40 

994 

8 

547 

TC 26-79 

23 

NOV-25 

NOV 

3 

30 

995 

10 

1071 


1979 TOTALS 

24* 



93 



•Overlapping days included only once in sum. 


the physical processes that are important to their development. The 
following case studies will emphasize the three stronger systems, two of 
which occurred over the Arabian Sea region where there is new interest 
in Arabian Sea tropical cyclones because of political tensions in the 
Persian Gulf. There is also new interest in the association between 
springtime Arabian Sea tropical cyclones and the onset of the summer 
monsoon (Krishnamurti, et al., 1981). 

Tropical cyclones in the N.I.O. generally originate from cloud 
cluster disturbances embedded in the monsoon trough. The low-level wind 
flow is from the west on the equatorward side and from the east on the 
poleward side of the developing system. A strengthening of the westerly 
winds on the equatorward side or of the easterly wind to the poleward 
side (or both) is a .common feature occurring prior to the formation of 








37 



CN 

bfl 

•iH 

u. 


The track of all 1979 North Indian Ocean tropical cyclones. 














































38 


tropical cyclones in this and other monsoon trough regions where cyclone 
formation occurs (Love, 1982). 

Figures 13 and 14 show the north-south and east-west vertical cross 
sections of zonal and meridional winds (U and V), averaged for all seven 
1979 systems at their initial center position at their first reported 
12Z time period. As expected, a surface to 400 mb deep layer north- 
south monsoon trough circulation exits at these initial periods. The 
strong upper level westerly flow to the north of the center is just the 
prevailing mid-latitude westerlies. Near the equator the winds are, as 
expected, strongly easterly at upper levels. 

The combination of diagrams 13a and 14b represents the net 
rotational or vorticity part of the wind around the initial cyclone 
center. It is interesting to find that the vorticity inside 5° radius 
and through a deep layer around the initial cyclone is as large as it 
is. The most distinguishable feature of the N.I.O. cloud cluster 
systems that did not form cyclones was, in general, a lack of such a 
strong surrounding deep-layer cyclonic vorticity field. As the tropical 
cyclone spins up and gains inner-core strength this outer cyclonic 
circulation field undergoes little change, however. 

The upper-level anticyclonic circulation is much shallower and much 
less defined. Figures 13b and 14a show the divergent part of the wind. 
To the south and west of the center this low-level convergence is very 
shallow and appears to have originated from the cross-equatorial 
monsoonal flow across the equator. Notice that the equatorial monsoon 
flow (and rotational part of the wind about the pre-cyclone disturbance) 
is present through a very deep layer but the convergence only through a 
shallow layer. In the upper level, divergence is found most in the 


39 


meridional direction with outflow channels towards the pole and the 
equator. The equatorial outflow branch corresponds to a cross- 
equatorial return flow. 



( 0 ) 

U 



(b) 

u 


Fig. 13. North-south cross section of zonal (diagram a) and meridional 
(diagram b) winds in m/s, averaged with respect to the center 
of all seven North Indian Ocean tropical cyclones in 1979 at 
their first 12Z time period. 
















PRESSURE (mb) 


40 



W * -(Deg. La t.)-> E 



(b) 

v 


Fig. 14. East-west cross section of zonal (diagram a) and meridional 

(diagram b) winds in m/s, averaged with respect to the center 
of all seven North Indian Ocean tropical cyclones in 1979 at 
their first 122 time period. 


























41 


6. ANALYSIS OF INDIVIDUAL CASES OF NORTH INDIAN OCEAN TROPICAL 
CYCLONES DURING FGGE 

We will now discuss the characteristics of each of the seven 
tropical cyclones that formed within the N.I.O. during the 1979 FGGE 
year. The observational information during this year was larger and its 
accuracy greater than in any period before or after 1979. This is 
because of the special network established during this period. 

6.1 TO 17-79 

TC 17-79 was the strongest tropical cyclone (V = 85 kts) in 

max 

N.I.O. in 1979 (see Fig. 15.). It was the most destructive cyclone in 

India since TC 22-77 (Nov. 1977) which coincidentally followed a similar 

track and did great damage to the Madras area. TC 22-77 claimed 14,000 

lives, while TC 17-79 caused about 700 deaths. The smaller death total 

with TC 17-79 is believed due to better preparation and to better 

heeding of the storm warnings as a consequence of the destruction of the 

previous cyclone. More than 300,000 people were evacuated in advance 

(DeAngelis, 1979) of TC 17-79. TC 17-79 was also responsible for 

sinking the United Vanguard ship before it could take evasion action. 

At 1200Z on May 06, observations from ships participating in FGGE 

defined a cyclonic circulation near 7.0°N, 88.0°E with reported wind 

speeds of 20-25 kts. After an erratic and looping track, 21/2 days 

later at 00Z May 09, TC 17-79 attained typhoon intensity of V >65 

max - OJ 

kts. 

During this period another tropical cyclone was located in the 
South Indian Ocean about 750-800 n mi. to the southwest of TC 17-79. 


42 



Fig. 15. Best track of TC 17-79 (JTWC ATR 1979). 


The Southern Hemispheric tropical cyclone tracked slowly towards the 
southeast. Following a northwest track. TC 17-79 weakened a little in 
the next 24 hours and then reinten3ified to its maximum intensity on 12Z 
May 11 when maximum winds were estimated to be 85 knots and the center 
pressure 967 mb. TC 17-79 then rapidly lost its intensity and 
dissipated when it moved over the Indian subcontinent near Madras. 

Time Series of Wind Parameters . The time series of estimated V 

max 

(in knots) and the low-level outer circulation (700-1000 mb V at 6° 
radius in m/s) are shown in Fig. 16. It is interesting to note that the 

































43 



low level outer 6 radius circulation (V^) was strengthening prior to 
JTWC's first report at 12Z May 05. From the 5th to the 7th of May, the 

^max increased from 15 kts to 35 kts without a significant increase in 
the low-level outer circulation, however. The low-level outer 
circulation attained its peak magnitude at 00Z May 10, two days before 
TC 17-79 reached its maximum intensity. 

The third curve in Fig. 16 shows the mean upper tropospheric 
outflow (V r ) between 150 and 250 mb at 6° radius. This outflow (V ) was 
obtained by averaging the radial wind at the 150, 200, and 250 mb 
layers. A 3-point running mean (between 12-hour interval data) was 
taken to smooth out diurnal and other short time-scale variations. 


(150-250 mb) at 6° (m/s) 






44 


(This 3-point running mean was applied to all time series with 

significant 00Z and 12Z variation.) The upper-level outflow of TC 17-79 

increased significantly right before the system was officially reported 

by JTWC. Another major increase occurred between 12Z Hay 07 and 12Z May 

08 when the system was undergoing its maximum intensification rate. 

There was a noticeable break in the increase of V between 00Z May 09 

max 

and 00Z May 10 associated with the decrease of the uppei—level outflow. 

Early Stage . Between 00Z May 03 and 00Z May 05, the ITCZ (Inter 
Tropical Convergence Zone) was located between the equator and 10°N from 
60°E to 90°E with loosely organized cloud clusters spread over the 
entire region. The only significant change in the ITCZ strength during 
this stage was an increase of the low-level outer circulation of the 
pre-TC 17-79 disturbance. Figure 17 shows this low-level tangential 
wind out to the 14° radius about the position of 6.4°N, 90.6°E at 12Z 
May 03, 12Z May 04, and 12Z May 05. This is the location of TC 17-79 at 
12Z May 05. The heavy curves are the 2 m/s tangential isotach. 

At 12Z May 03, a maximum V of 12 m/s is located at 4-6° radius 
south and southwest from the center. North of the center, tangential 
winds are only 2-3 m/s. The 2 m/s isotach does not penetrate through 
the west side (V^ is negative here) to form a closed pattern around the 
center. By 12Z May 04, however, a closed pattern of 2 m/s isotach is 
found around the center, V to the north has increased to 5 m/s. The 
maximum V fc to the south and southwest of the center is maintained at 12 
m/s. Negative is still present at 12° radius to the south and 10° to 
the north and east of the center. However, positive V is spread 


over 



45 



Pre-TC 17-79 
700-1000 mb 

v t 

ii 

6 —i-1 

0 4 8 

Deg. Radius 


(C) 


Fig. 17. Plan view of low-level circulation . (700-1000 mb V^) of TC 17-7 9 

about 6.4°N, 90.6°E or the center of the pre-TC 17-7 9 disturbance at 
(a) 12Z May 03, (b) 12Z May 04. and (c) 12Z May 05. Values in m/s. 




















46 


most of the region around the cyclone by 12Z May 05. The maximum mean 
V around the cyclone increased from 12 to 14 m/s during this period.. 

Figure 18 shows the low-level large-scale circulation pattern at 
00Z May 05. A westerly wind maximum is located on the equator between 
80°E and 90°E. The pre-TC 17-79 disturbance is located to the northeast 
or cyclonic side of this low-level wind maximum region. In the south 
Indian Ocean, TC 16-79 is located at around 7°S, 84°E. TC 16-79 
attained tropical storm strength on May 04 and intensified to 50 kts and 
60 kts on the 5th and 7th respectively (DeAngelis, 1979). The increase 
of TC 16-79's intensity appears to have helped produce these strong 
westerly winds near the equator. The cross-equatorial flow along the 
east Africa coast does not seen to play a significant role in producing 
this strong low-level coupling between TC 17-79 and TC 16-79. 

At upper levels, an anticyclonic center is located around 15°N, 

90°E, north of the pre-cyclone cloud cluster region where TC 17-79 will 
form. An upper-level trough is located 20° to the west - northwest of 
the anticyclone center. The flow is mostly northeasterly and easterly 
near the vicinity of the cluster region. No significant change in 
upper-level flow occurs during the three-day period of genesis except 
for an upper-level trough moving slowly eastward. 

Developing Stage . At 12Z May 05, a well organized cloud pattern is 
found centered at 6°N, 91°E. Maximum wind at this time is estimated at 
20 kts (JTWC ATR, 1979). A significant developmental feature appears to 
be the establishment of an upper tropospheric outflow channel to the 
north. This can be seen from the increase of the 150-250 mb radial wind 
from 12Z May 04 to 00Z May 05 in Fig. 16. The radial circulation 



MAY 5, I2Z 850 mb 


47 



850 rnb streamline field at 12Z May 05. Circles denote the center of TC 17-79 
and TC 16-79. 

















































48 


increases slightly in the next 48 hours with the maximum wind going from 
20 to 40 kts at 12Z May 07. 

Radial-height cross sections of azimuthally averaged tangential 
wind (V^.) at 12Z May 05, 12Z May 07, and 12Z May 09 are shown in 
diagrams a, b and c of Fig. 19. Cyclonic circulation extends up to 
300-400 mb with maximum anticyclonic wind near 175 mb. This circulation 
pattern is very similar to those of developing tropical cyclone wind 
patterns in the northwest Pacific as discussed by McBride, 1981. Note 
how the outer cyclonic tangential circulation from the surface to 400 mb 
shows a steady increase while the upper-level anticyclone circulation 
shows little change until the maximum intensity is reached. Diagram a 
of Fig. 20 depicts the difference in the tangential wind between 12Z May 
07 and 12Z May 05. The tangential wind increases over the whole depth 
of the troposphere between these two time periods. 

A stronger intensification rate is noted between 12Z May 07 and 00Z 
May 09. The circulation pattern at 12Z May 09 is shown in diagram c of 
Fig. 19 and the tangential wind change from 12Z May 07 to 12Z May 09 is 
shown in diagram b of Fig. 20. A comparison with diagram a of Fig. 17 
shows that the tangential wind increase is mostly at larger radii from 
12Z May 05 to 12Z May 07, but at smaller radii from 12Z May 07 to 12Z 
May 09. In other words, the cyclonic circulation increases at all 
levels as the cyclone is in its early developmental phase from 20 to 40 
knots but then has an increased inner tangential wind concentration in 
the later developmental stages from 40 to 60 knots maximum winds. The 
level of maximum tangential wind increase also shifts to a lower level 
in the later developing stages. 


P {100 mb) P(IOOmb) P(IOOmb) 


49 



(a) 05I2Z 


(b) 0712Z 


(c) 09I2Z 


Radius-height vertical cross sections of azimuthally averaged 
tangential wind (in m/s) for TC 17-79 at (a) 12Z May 05, V max = 20 kt 
12Z May 07, V max = 40 kts, and (c) 12Z May 09, V max = 60 kts. 


Fig. 19. 

















(100mb) P (100mb) 


50 





Radius ( deg. lat.) 

Fig. 20. Radial-height cross-section of 2-day tangential wind increases 

for TC 17-79 in m/s: (a) 12Z May 05 to 12Z May 07; (b) 07 12Z to 
09 12Z. Plus sign means an increase in cyclonic circulation 
or a decrease in anticyclonic circulation. 












51 


During this tine period, the 200 mb upper-level trough moves closer 
to the longitude of the cyclone center (Fig. 21). An upper-level 
outflow channel is established to the north of the cyclone center. 

Figure 22 shows the plan view of the 150-250 mb radial winds in a moving 
coordinate relative to the center of TC 17-79 at 12Z May 07. A major 
outflow channel is found to the north of the center. A broader outflow 
region is also located to the southeast of the cyclone center.• The 
looping track of TC 17-79 starts around this time period. This looping 
is likely due to changes of the steering flow current resulting from 
fluctuations in the north vs. south strength of the surrounding monsoon 
trough as has been discussed by Xu and Gray, 1982. 

Although the upper-level outflow started to weaken after 12Z May 09 
the V max still continued to show a general increase from 12Z May 09 to 
12Z May 11. This appears to be related to the passage of the upper- 
level trough system. Convection gradually concentrated at inner radii 
especially on the east and southeast sides of the center. Figure 23 
shows the satellite portrayal of this cyclone at 12Z May 09 and 12Z May 
11. An upper tropospheric outflow has been established to the west and 
southwest of the center. Figure 24 shows the upper-level large-scale 
circulation pattern at 00Z May 10 and 00Z May 11. The upper-level 
radial wind (in the moving coordinate system relative to the cyclone 
center) is shown in Fig. 25. A pronounced outflow channel is located to 
the west of the cyclone center. 

Decaying Stage. TC 17-79 makes its landfall at 06Z May 12 and 
maintains its intensity of 80 kts for another 6 hours or so. Its 
circulation patterns are well-maintained until 12Z May 12. It then 
rapidly dissipates as it moves inland. 



52 



OOZ 200 mb 


20° N 


{Q) 



20° N 


IIO°E 


(b) 


\ 

Fig. 21. ECMWF analyzed 200 mb wind field associated with TC 17-79 for (a) 
OOZ May 07, and (b) OOZ May 08. The large solid circle denotes 
the center of TC 17-79. Length of arrows is proportional to wind 
speed. 































































53 



Deg. Radius 


#—i—i 

0 4 8 


Kl 


A 


Fig. 22. Plan view of upper-level radial wind for TC 17-79 (150 - 250 mb. 


V r ) at 12Z May 07, V max = 40 kts, in a moving coordinate 
relative to the cyclone center. TC 17-79 is moving toward 275° 
at a speed of 3.1 m/s. Shaded area is negative or inflow 
region. 


6.2 TC 18-79 

TC 18-79 began as a monsoon depression in the Arabian sea at 14Z 
June 17. Maximum wind speed at that time was estimated to be 25 kts. 
TC 18-79 followed a westward track throughout its life. It dissipated 
over the Oman coast (Fig. 26). Its sustained wind increased to a 
maximum value of 50 kts in 36 hours. It maintained this magnitude for 
24 hours until shortly before it hit the coast. Figure 27 shows the 

time series of V^, the low-level outer circulation, and the upper- 
level outflows at 6° radius. 








54 



{b) 

Satellite picture of TC 17-79, 
May 09, (b) May 11. 


center is circled, at 12Z on (a) 


Fig. 23 














55 




Fig. 24. 200 mb flow" field relative to TC 17-79 (circled) for 00Z May 10 

(diagram a) and 00Z May 11 (diagram b). Length of arrows is 
proportional to wind speed. 



















































56 



A 

In 

#—4—i 

0 4 8 

Deg. Radius 


Fig. 25. Plan view of upper level radial wind (150 - 250 mb average V r ) 

for TC 17-79 at 12Z May 11. (V max = 85 lets) in a moving coordinate 
system. The storm is moving toward 295° at 3.4 m/s). 



Fig. 26. Best Track of TC 18-79. (From JTWC ATR, 1979.) 


























57 



Prior to TC 18-79's formation, a frontal system moved eastward in 
the South Indian Ocean. A surge type flow propagated equatorward behind 
this front. This brought about a strengthening of the low-level wind 
maximum along the Somali coast. This low-level band of strong winds 
extended eastward across the entire Arabian Sea. Although this low 
level wind maximum is climatologically always along the Somali coast 
during this season, its strength was greater than normal strength during 
the period of TC 18-79's formation. This appeared to aid in the 


(150-250 mb) at 6° (m/s) 







formation of TC 18-79 by increasing the westerly winds on the southern 
side of the pre-cyclone disturbance and generally enhancing its low- 
level circulation. 

This strengthening of the low-level jet can be seen from the plan 
view of the low level (700-1000 mb) tangential wind field around the 
center of TC 18-79 at 12Z June 17 when the center was located at 17°N, 
68°E. Figure 28 shows these plan views at 12Z June 15, 12Z June 16, and 
12Z June 17. At 12Z June 15 (48 hours before the beginning of TC 18- 
79), a low-level wind speed maximum of 16-17 m/s was located 10° south 
of the pre-cyclone center. Cyclonic circulation is found only on the 
south side of the center. This is part of the prevailing southwesterly 
flow from the low-level' Somalia jet. The pre-cyclone disturbance center 
is located just on the poleward (or cyclonic shearing side) of this 
low-level wind speed maximum. To the north side of the pre-cyclone 
center only a weak anticyclonic circulation is present. 

By 12Z June 16, 24 hours later, the maximum wind to the south 
increases slightly. The most distinctive new alteration of the low- 
level flow is the spreading of the cyclonic circulation to the east and 
north sides of the developing system. Note how the 2 m/s isotach has 
now formed a closed pattern around the center. The low-level 
anticyclonic region has greatly decreased. The maximum wind increased 
to 23 m/s by 12Z June 17 and moved closer to the system center. The 
cyclonic wind spreads over almost the whole region. This gradual 
building-up of the low-level circulation is similar to that of TC 17-79. 


59 



TC 18-79 
700-1000 mb 

Vf 

A 

N 

§—i—1 
0 4 8 

Deg. Radius 


Fig. 28. Plan view of low-level circulation (700-1000 mb V t ) in m/s 

relative to the center of TC 18-7 at 12Z June 15 (diagram a), 
12Z June 16 (diagram b) and 12Z June 17 (diagram c). 



















60 


The gradual strengthening of the average low-level 6° radius 

cyclonic circulation around the entire cyclone system in the early 

stages can also be observed in Fig. 27. At 12Z June 15, the azimuthal 

average of tangential wind at 6° radius was only 1.7 m/s. However, by 

12Z June 17 (two days later), the wind speed at this same radius has 

increased to 10.7 m/s. Compared to that of TC 17-79, this magnitude is 

much greater (at the formation stage) even though TC 18-79 never reached 

the same intensity level that TC 17-79 did. The cyclonic circulation of 

TC 18-79 was not only strong, but it also extended over a very broad 

region. Diagrams a and b of Fig. 29 show vertical cross sections of 

azimuthally averaged V. at 12Z June 17 and 12Z June 19 when V are 

t . max 

estimated to be 25 kts and 50 kts, respectively. Again, the cyclone 
circulation patterns are quite similar to those found in northwest 
Pacific tropical cyclones. However, the outer and environmental 
circulation of TC 18—79 is much stronger at this early stage, compared 
to the average of the other cases. A lower tropospheric azimuthally 
averaged wind speed of 7-8 m/s at 10° radius as is present at 12Z June 
17 is quite large. It is interesting to note that the upper- and 
lower-level circulation beyond 4° radius does not change much during 
this 48-hour period, but the maximum inner core cyclone intensity 
doubles its magnitude during this time interval. 

It should be noted that the 6° radial outflow increases greatly 
prior to the beginning of TC 18-79's formation but does not increase 
during the period when the maximum winds increase the most between 12Z 
June 17 and 00Z June 19. However, the 200 mb outflow concentrates in 
channels to the west and southwest of the center as shown in Fig. 30 


P (100 mb) P (100 mb) 


61 


1 

z 

3 

4 

5 

6 

7 

8 

9 

10 

2 4 6 8 10 12 i4 

Radius (deg. lat.) 





(b) 


Fig. 29. Radial cross sections of the azimuthally averaged tangential 

wind (V fc ) in m/s for TC 18-79 at 12Z June 17 (diagram a) and 
12Z June 19 (diagram b). 

















62 


for the period of 00Z June 17 and 00Z June 18. An examination of the 
upper-level circulation pattern shows that the center of TC 18-79 is 
located south of an upper-level anticyclone center. This results in a 
strong prevailing easterly wind over the vicinity of TC 18-79. This 
prevailing easterly wind and westward stretching of the Southern 
Hemisphere anticyclone have eliminated the outflow to the east and 
enhanced the outflow channel to the west and southwest. Figure 31 shows 
the plan view of the upper-level radial wind at 12Z June 17 in which a 
major outflow channel can be seen toward the southwest. It is this 
outflow channel that might have caused the intensification of TC 18-79. 
Documentation of outflow channel increase and cyclone intensification 
has been previously discussed by Chen and Gray (1984). A physical 
hypothesis for this process has been advanced by Holland and Merrill 
(1984) of our project. 

Although TC 18-79 was still increasing in inner-core intensity from 
00Z June 18 to 00Z June 19, the upper-level outflow began to weaken at 
this time (a feature similar to that of TC 17-79). This was due (we 
believe) to the increase of the upper-level inflow on the eastern side 
produced by the strong westward directed flow at this level. Besides 
causing outflow decreases, this strong upper-level easterly flow appears 
also to have acted to bring about a shearing off of convection. This 
brought about a rapid weakening of TC 18-79 before it reached the Oman 
coast. Figure 32 shows the satellite IR picture of TC 18-79 at 18Z June 
19, when TC 18-79 starts to weaken. The clear region and sharp edge of 
the cloud pattern on the east side indicate an inflow to the storm 
system from the east. Another reason for TC 18-79's rapid weakening is 


63 




JUNE 18, OOZ 200mb 


50°E 


80°E 


I IO°E 


(b) 


Fig, 30, 200 mb flow field for TC 18-79 at OOZ June 17 (diagram a) and 
OOZ June 18 (diagram b) . The large dot shows the center of 
TC 18-79, The length of the arrows is proportional to wind speed. 












































04 



A 

N 

#—I—l 

0 4 8 

Deg. Radius 


Fig. 31. Plan view of upper-level radial wind (150-250 mb V r ) for TC 18-79 
at 12Z, June 17, in a moving coordinate (277° and 4.0 m/s). 



Fig. 32. Satellite, pictures of TC 18-79 at 18Z, June 19. Center is circled. 
















65 


its moving close to the dry Arabian Peninsula. Similar weakening of 
tropical cyclones can occur off the northwest Australian Coast if dry 
desert air should be advected into the storm system. 

Although TC 18-79 weakens rapidly after 18Z June 19, both the low- 
level and upper-level outer circulation patterns are well maintained 
until 12Z June 20 (V fflax = 20 kts) as seen in diagram a of Fig. 33. 
Despite the zero line of this figure moving to a lower altitude, the 
low-level cyclonic circulation still has a magnitude of 5-7 m/s at 10° 
radius. Diagram b of Fig. 33 indicates that a closed cyclonic 
circulation is still present over a broad region at this time even 
though the inner maximum wind strength has greatly decreased. Maximum 
winds have also shifted to the south and southeast of the center. This 
is due to a westward movement of TC 18-79 away from the strong low-level 
ITCZ westerly flow which held its position at the same location, 
DeAngelis (1980c) has speculated that TC 18-79 helped bring about the 
advance of the southwest monsoon over the western portions of the Indian 
subcontinent. Krishnamurti, et al. (1981) have presented evidence that 
a low-level tropical cyclone vortex often forms in the Arabian Sea just 
prior to the advancement of the summer monsoon over India in June. 

6.3 TC 23-79 

TC 23-79 was first reported as a monsoon depression at 02Z Sept. 

18. Prior to this time, JTWC had traced the pre-TC 23-79 tropical 
disturbance for more than two days. It tracked westward across the 
southern end of the Indian subcontinent (Fig. 34) and into the Arabian 
Sea. Loosely organized convection was associated with this disturbance. 
The upper-level prevailing wind was from the east. The plan view of the 


66 



Fig. 33. 


The cross section of average tangential wind (V.) of TC 18-79 at 
12Z June 2Q (diagram a) and the plan view lcw-level 1000 to 
700 mb average tangential wind (V ) at 122 June 20 (diagram 
b). Wind speeds in m/s. 












67 


■ 



£ ►* xi 

55 I;1 5| 

r: Or ? J* w* < 

« i u un * ' 

=i ml m i 

E S ^ooirt jii! 

□ 0S0 ,fJ<< Q--2 > 

— t a • 

^ *i 2;Oa.fli»^^l3Xl 

O ~ L - . . — jh j* u* ( 


* * * 

"?* o- 2 § 

£J ^ o 

o £ Q 2 

fc! *■ 

y S> it 

^ tn > 

y«u s 

2 ?a i 

** EL. ^ S 
k- UJ X “ 
u*i *" < 

LM -* IS 

a ci 


O* lo 

hs 

— O 
a* ± 

S3 


L±J X 

tn < 

C* 3 


On 

r* 

o\ 


cn; 

£- 

< 

O 

£ 


ON 

r- 

l 

CN 

CN 

CJ 

E- 

TJ 

C 

CO 

On 


cn 

CN 

CJ 

H 

<w 

O 

.X 

a 

CO 

t- 

xJ 

4 J 

CO 

a? 

m 


TT 

cn 

h0 

•»-H 

Ci_ 


















































68 


upper-level radial wind at 12Z September 17 showed an inflow on the 
eastern side and an outflow to the western side of the center (diagram a 
of Fig. 35) or a westward blow-through at upper levels. This outflow 
shifted with time (diagram b - September 18th) and, on the 20th (diagram 
c), there is a double outflow to the southwest and the northeast. 

Double outflows are, in general, more conducive to cyclone 
intensification. 

The time series of the V , the low-level outer circulation, and 

Ul&X 

the upper-level outflow are shown in Fig. 36. Like those of TC 17-79 
and TC 18-79, the low-level outer circulation increases (from 3.2 m/s to 
6.4 m/s) prior to the formation of TC 23-79. Diagrams a and b of Fig. 

37 show the plan view of the low-level circulation at 12Z September 16, 
12Z September 17, and 00Z September 18 in a moving coordinate relative 
to TC 23-79's center. As with the two previously discussed cyclone 
systems, the initial cyclonic circulation is concentrated in the 
southern part of the region; while in the northern part, the flow is 
hardly cyclonic at all. The maximum tangential wind is located 8-10° 
south of the center with a magnitude of 16-17 m/s. On the synoptic 
chart, the pre-TC 23-79 disturbance is located north-northeast to a 
low-level westerly wind maximum which is likely driven by the cross- 
equatorial monsoonal flow from the east coast of Africa. 

Note how the flow on the north side of the TC 23-79 system becomes 
more cyclonic and extends out to 6-8° radius. This results in an 
increase of the low-level outer circulation despite the decrease of the 
maximum cyclonic circulation to the south. At 00Z September 18, the 


69 





TC 23-79 
150-250 mb 

V r 

A 

hU 

i—I 

0 4 8 

Deg. Radius 


Fig. 35. Plan view of upper-level mean radial wind (V ) in m/s between 

(150-250 mb levels relative to the moving center of TC 23-79 at 
12Z September 17 (diagram a), 12Z September 18 (diagram b), and 
12Z September 20 (diagram c) moving coordinate. 




















70 



s-\ 

o 

<£> 


JO 

E 

O 

o 

o 

i 

o - 

g 

\>-\ 


0.5 


0.0 




1.5 £ 

o 

<0 

1.0 o 


-Q 

E 

O 

to 

CM 

I 

o 

to 


- 0.5 !>*" 


Fig. 36. Time series of cyclone intensity (V max ) and low-level outer 

circulation (V.) and upper-level outflow (V ) at 6° radius for 
TC 23-79. r 

maximum cyclonic circulation increases to 16 m/s and moves in closer to 

the center. The cyclonic circulation to the north and northeast of the 

center also shows a gradual increase of the circulation about TC 23-79. 

Unlike TC 17-79 and TC 18-79, the development of TC 23-79 is rather 

slow after the initial organizing of the system on 00Z September 18. 

Intensity increases only from 15 kts to 25 kts in three days. The low 

level outer circulation increases only slightly from the 18th to the 

20th and then decreases. The most peculiar feature in the time series 

is the early peaking of the 6° radius upper-level outflow at 00Z 

September 17 and then the decrease of the outflow between 00Z September 

17 and 00Z September 18. A much greater decrease occurs between 










71 




Plan view of low-level circulation (700-1000 mb V t ) of TC 23-79 
at (a) 16 12Z, (b) 17 12Z, and (c) 18 00Z of September for 
P^e-TC 23-79 disturbance in a moving coordinate. 


Fig. 37. 






















72 


12Z September 19 and OOZ September 21. This is likely a result of the 
inward concentration of the convection at the expense of the outer 
convection and a change in the surrounding flow patterns. 

The IR satellite pictures of TC 23-79 at 06Z September 17, 18Z 
September 18, and 18Z September 20 are shown in Fig. 38. The 
corresponding radial outflows at these three time periods are shown as 
"X" marks in Fig. 36. Cloudiness at 06Z September 17 is spread over a 
much larger domain than on the 18th or the 20th. The prevailing 
easterly flow causes a major outflow to the western side of the 
convection as is seen in diagram a of Fig. 35. A northeast outflow jet 
is established on the 18th. This is due to the approaching of the 
upper-level trough to the northwest of the system. 

At 18Z September 20, the cirrus coverage becomes greatly reduced. 
The outflow to the northeast side increases while the outflow to the 
west, southwest and south sides reduces and becomes confined to a much 
narrower region. These features can be seen from both the satellite IR 
image (diagram c of Fig. 38) and the upper-level radial winds (diagram c 
of Fig. 35). The prevailing upper-level flow has undergone a change to 
be more from a southerly as opposed to an easterly direction as TC 23-79 
moves from the south side to the west side of the upper-level 
anticyclone center. The ECMWF upper-level 200 mb flow fields at OOZ 
September 21 and OOZ September 22 are shown in Fig. 39. TC 23-79 is now 
located to the west side of the anticyclone. This 200 mb anticyclone 
does not change its location through the life cycle of TC 23-79. TC 
23-79 moves from the south to the west side of the anticyclone. 

TC 23-79 experiences a much stronger intensification rate after 06Z 
September 21. The upper-level outflow also increases slightly at this 


73 



caption on following page)) 














































74 



Fig. 38. Satellite picture of TC 23-79, center is circled, at 06Z September 
17 (diagram a), at 18Z September 18 (diagram b), and at 18Z 
September 20 (diagram c). 


time but decreases greatly after 00Z September 22. This outflow was 
much stronger 2-3 days before this major intensification period, 
however. This stronger intensification rate on the 22nd and 23rd is 
believed due to the approach of the upper-level trough from the 
northwest. This creates a poleward outflow jet as seen in diagram c of 
Fig. 35. The upper-level circulation patterns at 00Z September 21 and 
00Z September 22 (Fig. 39) show this interaction between upper-level 
trough and the cyclone. As discussed by Chen and Gray (1984) and Sadler 
(1976), such an increase in outflow jets is usually favorable for 
tropical cyclone intensification. 









75 



SEPT 21, OOZ 200mb 


40° 


20 ° 


20°E 50°E 80°E 




20° N 


20°S 


50° E 


80°E 


IIO°E 



Fig. 39. 200 mb flow field at OOZ of (a) September 21 and (b) September 

22. Circle is the center of TC 23-79. Length of arrows is 
proportional to wind speed. 
































































Although named before TC 23-79, TC 22-79 formed in the Bay of 
Bengal after TC 23-79. It was officially named as a tropical cyclone at 
02Z Sept. 20 with estimated maximum sustained winds of 20 kts. The best 
track for TC 22-79 is shown in Fig. 34. Since TC 22-79 formed very 
close to the Indian subcontinent, it did not last long (only three days) 
before it hit the land. No significant intensification was observed. 

TC 22-79 formed east of the region where TC 23-79's initial cloud 
cluster had originated. Both systems were much affected by the westerly 
winds of the summer monsoon. Besides the equatorial flow off the east 
coast of Africa,another cross-equatorial flow also existed between 75- 
80°E or southwest to the center. (See Fig. 40). This cross-equatorial 
flow is likely a result of the strengthening of the circulation around 
Southern Hemisphere Tropical Cyclone TC 21-79 centered at 6°S, 82°E. 
DeAngelis (1980a) indicates that TC 21-79 formed on the 15th of 
September and attained its maximum intensity of 45 kts on the 22nd. 

This was the same period during which TC 22-79 was developing. The 
increase of the low-level outer circulation of TC 22-79 is also 
portrayed in Fig. 41. The plan view of the low-level tangential wind 
(figures not shown) indicated that this increase (as with the three 
previous systems) occurred mostly in the southwest quadrant at radii of 
6-10° where the influences of the Southern Hemisphere tropical cyclone 
and the westerly monsoon flows merged. 

At the upper level, an anticyclonic center is located near 20°N, 
100°E causing a prevailing easterly wind over the vicinity of the pre-TC 
22-79 cloud cluster. The increase of the upper-level outflow is a 
result of the increase of convection on the western side of the system. 


SEPT 19, OOZ 850 mb 


77 



O O 

O LlI 

00 


Fig. 40. 850 mb flow field at 00Z # September 19. Circles are the centers of TC 22-79 and TC 23-79. 










































78 




O 

CO 




1.0 



Fig. 41. Time series of cyclone intensity (V max ) and low-level outer 

circulation (V. ) and upper-level outflow (V ) at 6° radius for 
TC 22-79. r 

Figure 42 shows the satellite picture at 182 September 19 (this time 

marked by an X in Fig. 41). which indicates that the deep Cb convection 

and outflow is located mainly at the western side of the system. The 

plan view of the upper-level radial wind (not shown) also indicates an 

increase of outflow in this region. This further demonstrates that 

upper-level divergence which is associated with Cb convection around a 

tropical cyclone does not necessarily result in a significant 

intensification of the system. Our project's research in recent years 

has clearly shown (Arnold. 1977) that the amount of deep convection in a 

tropical system is not well related to the intensity or the intensity 

change of that system. It is the concentrated nature of the outflow 

into narrow channels that is the important ingredient of 


(150-250mb) at 6°(m/s) 









79 


Fig. 42. Satellite picture of TC 22-79 at 18Z, September 19. Centers of 
TC 22-79 and TC 23-79 are circled. 

intensification. This is not to say that the amount of deep convection 
is not better related to the strength of the cyclone as defined in 
section 4. 

On the 22nd of September, the upper-level trough, which caused a 
major intensification of TC 23-79, moves eastward to the north of TC 
22-79. However, no intensification is observed with TC 22-79 during 
this time period. One possible reason for this lack of TC 22-79's 
intensification is its close proximity to land at this time. Another 
reason, which is likely more important, is that the trough was too far 
away from the cyclone center. As can be seen from Fig. 43, the outflow 
associated with the trough to the north of TC 22-79 is not very strong 
and does not penetrate close to the storm center. This is more evident 
when a comparison is made with Fig. 35(c) which shows the upper-level 



















80 




Kl 
#— I- 

0 4 


8 


Deg. Radius 


Fig. 43. Plan view of upper-level radial wind (150-250 mb V r ) for TC 22-79 
at 12Z September 22 in a moving coordinate. 

outflow pattern of TC 23-79 associated with the passage of the upper- 
level trough. The outflow is not only much stronger with TC 23-79 but 
it extends inward very much closer to the cyclone center. 

6.5 TC 24-79 

Prior to the formation of TC 24-79 in the Bay of Bengal, loosely 
organized cloud clusters were spread all over the North Indian Ocean 
between the equator and 15° north. JTWC post-analysis identify the 
pre-TC 24-79 cloud cluster at 00Z October 27. This cloud cluster 
tracked north-northwestward on the first day (see Fig. 44). At 00Z 
September 28, the pre-TC 24-79 disturbance moved to the southwest side 
(cyclonic shear side) of the low-level southeasterly trade winds. It 









81 



Fig. 44. Best track of TC 24-79 (JTWC ATR, 1979). 


then turned on a northwest track and started to organize its convection. 
At 00Z September 29, it was named TC 24-79 with estimated maximum wind 
speed of 20 kts. The 850 mb flow at this time is shown in Fig. 45. 

Note in Fig. 45, that the low-level wind speed maximum is likely a 
consequence of the beginning influence of the winter season 
northeasterly monsoon flow. The organization of this tropical cyclone 
is thus distinctly different from that of the four previous systans 
where the initial cyclonic flow was initiated on the equatorward side 
from westerly winds. Following this low-level prevailing easterly flow, 
TC 24-79 tracked westward throughout most of its life and gradually 
increases its intensity before striking the Indian coast near Madras. 

The low-level cyclonic circulation of TC 24-79 is much 





















04 


















































83 


weaker than that of the other systems, even at its maximum intensity 
(Fig. 46) . The low-level circulation was also much shallower in 

comparison with the other TC systems. In fact, TC 24-79 is a relatively 

small (in view of the radial domain) syston especially in the coverage 
of its satellite cloudiness. The mechanism which maintains TC 24-79's 
slow developing appears to be the coupling between the low-level 
vorticity and the upper-level anticyclonic circulation. TC 24-79 had a 
strong upper level anticyclonic surrounding circulation throughout its 
life. 

6.6 TC 25-79 

TC 25-79 was a large tropical system. Gusts up to 35 kts were in 

evidence out to 300 miles from its center (DeAngelis, 1980b). This was, 

however, a very strange tropical cyclone for two reasons. First, it did 
not have a recognizable cloud pattern — something almost always 
associated with any tropical cyclone that forms. Even at its maximum 
intensity (40 kts) at 06Z November 16, convection was still not well- 
organized near the center. Secondly, TC 25-79 followed a basically 
straight northward track as shown in Fig. 47. To understand why it 
moved straight northward, we have to look at both upper- and low-level 
flows. At upper levels, a trough was located over the Arabian Peninsula 
with strong south-southwesterly flow existing over the Arabian Sea 
region. At low levels a southeast monsoon flow was found over the 
Indian sub-continent. TC 25-79 was located on the southwest side 
(cyclonic shear side of a low-level southeast monsoonal wind speed 
maximum). The northward track of TC 25-79 is a result of the 
combination of the upper—level southwesterly and low-level southeasterly 
steering flows. The net zonal steering component was near zero. 


84 



Radius ( deg. lat.) 


Fig. 46. Cross section of tangential wind for TC 24 79 at 12Z, October 31. 



Fig. 47. Best track of TC 25-79 (JTWC ATR, 1979). 


























Since TC 25-79 does not have a well organized cloud pattern, the 

post-analysis by JTWC does not have it identified until 12Z November 14. 

However, its low-level circulation was well-established at 00Z November 

13. Figure 48 shows the time progression of the maximum wind (V ), 

max 

the low-level outer circulation, and the upper-level outflow. Note that 
the low-level circulation increases greatly from 00Z November 12 to 00Z 
November 13. It is likely that JTWC post-analysis should have named TC 
25-79 a little earlier. Also, note that, as with the previous system, 
the upper tropospheric outflow (V ) peaks before the maximum intensity 
of the cyclone is reached. 

The upper-level outflow is very strong until 00Z November 16. Two 
significant increases of upper-level outflow occur between 12Z November 
12 and 00Z November 14 and between 12Z November 15 and OOZ November 16. 

The former one is due to the enhanced convection to the west of the 
center of the pre-TC 25-79 cluster. (No significant intensification was 
associated with this increase of outflow.) The nighttime IR satellite 
images at 18Z November 13 is shown in Fig. 49. An upper-level trough is 
located to the northwest of this convective region. Note that this 
cloud pattern largely dissipates in 24 hours and leaves a less defined 
cloud pattern. The upper-level outflow to the north and northeast is 
also very pronounced, however. 

The second increase of upper-level outflow is a result of the 
approaching of the upper-level trough that strengthens the northeast 
outflow jet. Figure 50 shows the 200 mb flow features at OOZ November 
15. The northeast outflow jet is found mainly to the north and 
northeast of TC 25-79. It is this outflow jet that probably leads to 
the intensification of TC 25-79 between 12Z November 15 and 18Z November 16. 


86 


60 


^ 40- 

w 


TC 25-79 


/? 


X 

o 

> e 20 




' ♦ y : \ 
W-r .* : 



A 


' 150-250 mb V t at 6° (m/s) \ 

1.8 3.0-0.I -3.4-5.1 -7.1 -66-8.0-8.1 -8.4 -8.1 -8.7 

J I L 


12 13 14 15 16 17 18 

—*~DATE (Nov.) 


12 - 


0 

(D 


8 


JD 

£ 

O 

o 

4 Q 

i 

o 


l> 


2.0 


s 

o 

(£> 


1.0 


£ 

o 

to 

OJ 

I 

o 

to 


l> 


Fig. 48. 


Time series of cyclone intensity (V max ) and low-level outer 
circulation (V t ) and upper-level outflow (V ) at 6° radius for 
TC 25—79. Also shown is 150 mb — 250 mb tangential wind at 6° 
radius. 



Fig. 49. Satellite picture of TC 25-79 at 18Z, November 13. Center is 
circled. 














87 



Fig. 50. 200 mb flow field at 00Z, November 15. Center of TC 25-79 is 

circled. Length of arrows is proportional to wind speed. 


6.7 TC 26-79 

TC 26-79 developed in the Bay of Bengal in late November with 
easterly trade wind flow on its north side (Fig. 51). It was very 
similar to TC 24-79 in both its track and the influence of the low-level 
trades on its poleward side. Figure 52 shows the 850 mb flow pattern at 
00Z November 21 when TC 26-79 starts its organization. Note that TC 
26-79 is located just to the southwest (cyclonic shear side) of the 
low-level wind speed maximum. 

Although named at 12Z November 20, TC 26—79 does not have a clearly 
defined cloud pattern until 00Z November 23. Its low-level circulation 
pattern was also not well-established when it was named. However, the 
low-level vorticity was still large near the system during its early 
stages. At upper levels, an anticyclone center was located over the 
north border of the Bay of Bengal, with prevailing southeast wind over 




















88 


the cyclone region. The upper-level anticyclonic circulation of TC 26 
79 experienced a steady increase. No apparent strong outflow channels 
or significant intensification was observed with this system. 

Using the results of the analysis of these seven individual cases 
the next section will attempt to synthesize some of the important 
physical processes that seen to be occurring in N.I.O. tropical 
cyclones. 



BEST TRACK 
23 NOV-25 NOV 197V 
MAX SFC WIND 30 KT5 


MINIMUM SLP 995 MBS 


Fig. 51. Best track of TC 26-79 (JTWC ATR 1979). 





















NOV 21.12Z 850mb 


89 



850 rob flow fields at 12Z, November 21. Center of TC 26-19 is circled. 































7. 


DISCUSSICN OF RELEVANT N.I.O. TROPICAL CYCLONE FEATURES 


The previous section analyzed the large-scale circulation patterns 
for all seven tropical cyclones that formed in the North Indian Ocean 
during the 1979 FGGE-year period. Although there may be some 
inadequacies in the FGGE Ill-b data and analysis techniques, these data 
are still the best information available for any cyclone season. We 
believe the ECMWF analysis shows most of the important large-scale wind 
characteristics about these cyclone systems, particularly the rotational 
component of the wind field and the large-scale divergence patterns of 
the upper troposphere. We will now discuss the important general 
physical processes related to these tropical cyclones in terms of their 
influence on the genesis, intensification, structure and motion. It is 
recognized, however, that discussions based only on 1979 cyclone 
information may not always strictly apply to tropical cyclones occurring 
in this region for other years. 

7.1 Characteristics of the Large-Scale Circulation Patterns During 

Tropical Cyclone Formation and Development 

Like tropical cyclones in the northwest Pacific, N.I.O. tropical 
cyclones form within the monsoon trough region. To show this similarity 
to northwest Pacific cyclones, we have performed a similar north-south 
cross section averaging of zonal wind for the initial stages of all 39 
systems that later became tropical cyclones in the northwest Pacific 
during the same FGGE year. Results are shown in diagram a of Fig. 53. 

Note the general similarity when comparison is made to the same average 
of the seven N.I.O. pre-tropical cyclones as seen in diagram b of Fig. 53. 


91 



( 0 ) 

Northwest 

Pccific 



(b) 

North 

Indian 

Ocean 


Fig. 53. North-South.cross section of zonal winds (U), averaged for all 
39 northwest Pacific tropical cyclones (a) and 7 North Indian 
Ocean tropical cyclones (b) in 1979, at their first 00Z time 
periods. 















92 


A few differences are to be noted, however. The easterly trade winds on 

the poleward side are much broader in the northwest Pacific; but the 

\ 

N.I.O. trade winds are of a comparable magnitude and are closer to the 
disturbance center. 

The N.I.O. cases have a much stronger and broader monsoon-generated 
westerly wind component on their equatorward side in comparison to that 
of the northwest Pacific cases. Not only is this westerly monsoon flow 
stronger than in the Pacific, it is also much deeper. This implies that 
the equatorward westerly wind influences are more important to N.I.O. 
tropical cyclone formation than are equatorial westerly wind influences 
in the west Pacific. 

The average wind fields for all the seven N.I.O. cases in 1979 
reveal a very deep monsoon-generated cyclonic circulation around the 
center of the tropical cyclone formation region. An important feature 
from our previous case analysis is that there is always a build-up of 
the low-level cyclonic circulation prior to the formation of the 
tropical cyclone system. This feature is similar to the ITCZ type 
tropical cyclones of the northwest Pacific as has been observed by Lee 
(1984). Our analysis also shows that this low-level circulation build¬ 
up is mainly a result of the cross-equatorial monsoonal surges or other 
features that strengthen the westerly wind on the equatorward side of 
the pre-cyclone disturbances. Love (1982) has also observed similar 
equatorial side westerly wind increases about Pacific and Australian 
region tropical disturbances from opposite hemispheric middle-latitude 
cold surge penetrations into the tropics as illustrated in Fig. 54. The 
N.I.O. tropical cyclone forecaster should always be aware of such cold 
surge penetrations occurring in the Southern Hemisphere and their 


93 . 


potential for enhancement of the N.I.O. monsoon trough and TC formation. 

In the latter part of the autumn season (mid-October to December), 
however, the needed low-level cyclonic circulation -increase for TC 
formation may occur more on the poleward than the equatorward side of 




LONGITUDE OF GENESIS 
40*W 30 20 ' 10 _ 

” 1 t t r m. 


HOT* 


20*N 


I0*N 


IO*S 


Fig. 54. Highly idealized gradient level streamline chart for the 

three days preceeding tropical storm genesis in the Southern 
Hemisphere. Subscripts -3, -2, -1 refer to the position of 
the system that number of days before genesis. The momentum 
burst has been shaded (from Love, 1982). 


the pre-cyclone N.I.O. systems. It is thus possible for the main low- 
level cyclonic circulation build-up of N.I.O. systems to occur on either 
the equatorward or the poleward sides. The poleward trade wind 
circulation build-up may result from beginning East Asian winter season 
surges or other processes that act to enhance the trade winds. 






94 


Another special feature of the N.I.O. basin is the role of dual 
cyclones on either side of the equator. The strengthening of the 
equatorial circulation about a tropical cyclone in one hemisphere can 
cause a similar increase in the equatorial low-level circulation of the 
other hemisphere. Two of the seven 1979 cases were so affected. Keen 
(1982) and others have observed cross-equatorial tropical cyclone pairs 
to occur in both the Indian Ocean and in the Pacific. 

A strong low-level positive vorticity field is thus very important 
to tropical cyclone formation in the N.I.O. as it is in most monsoon 
trough genesis regions as previously discussed by the second author Gray 
(1968, 1979). It is to be noted that the initial enhancement of low- 
level circulation about the pre-cyclone disturbance is typically on only 
one side of the disturbance (i.e., equatorward side or poleward side) 
and that the tip-off to early cyclone formation and later growth rests 
with the gradual wrapping around of the tangential wind field to the 
other sides of the cyclone. 

It is to be noted by the time series of the individual storm 
profiles of V max> ? T (700-1000 mb) and V r (150-250 mb) at 6° radius - 
see Figs. 16, 27, 36, 41, 48 - how little related are these parameters. 

V can decrease or not decrease as increases. These observations 

show how difficult it is to relate the maximum intensity of a N.I.O. 
tropical cyclone to the strength of the storm's outer circulation. 

These observations also show how little the magnitude of the cyclone's 
net upper tropospheric radial outflow (i.e. V^) is related to the 

cyclone's intensity. It is interesting, however, how the net upper- 
level radial outflow (and likely also the net deep cumulus convection) 
of most of these N.I.O. system peaked about 2 days before the date of 


95 


the observation of the maximum wind. A similar measurement has also 
been reported by E. Rogers (personal communication) in like-time series 
analysis of 20 Atlantic tropical cyclones. These observations would 
indicate that a blow-up or enhancement of deep cumulus convection (as 
might be measured by the satellite) in a tropical disturbance or weak 
tropical storm may be indicative of that system's future 1-3 day 
intensification. 

Upper-level Flow Characteristics . McBride (1981), McBride and Zehr 
(1981) and others have emphasized the importance of the upper-level 
anticyclonic vorticity (besides the low-level positive vorticity) for 
tropical cyclone genesis.in the northwest Pacific and Atlantic. We do 
not see the crucial role of the upper—level anticyclonic vorticity in 
the early stage genesis of our N.I.O. systems. Upper-level anticyclonic 
flow features might play an important role in the later intensification 
stages of N.I.O. cyclones, however. 

We are finding that the increase of maximum winds (V ) in a tp i 

max 

not well related to the net upper—level outflow (V ) occurring around 
the tropical cyclone, but rather is more related to the concentration of 
this outward motion into narrow outflow channels (see Holland and 
Merrill (1984) and Chen and Gray (1984) for more discussion). Upper- 
level anticyclonic vorticity is required for the establishment of these 
narrow and strong outflow channels. The net radial outflow increase 
which is produced by an increase in the enhanced deep Cb convection 
within the tropical disturbance or tropical cyclone appears to be more 
related to the change in cyclone strength or the 1-3° mean tangential 
wind of the cyclone. 

Chen and Gray have discussed the variety of upper-tropospheric 



96 


outflow patterns which can occur with tropical cyclones undergoing 
central-core intensification. An idealization of the variety of these 
outflow patterns is shown in Fig. 55. The type of outflow patterns 
associated with most N.I.O. tropical cyclones are those of pattern a, b, 
c, and d. A southwest outflow channel is sometimes established by the 
strengthening and westward movement or stretching of an upper-level 
anticyclone in the South Indian Ocean. Poleward outflow channels are 
typically produced by an eastward propagating upper-level trough to the 
poleward side of the cyclone. It is observed that a tropical cyclone 
can undergo an even more rapid intensification if double outflow 
channels are present to both the poleward and equatorward sides of the 
cyclone. 

We have to emphasize that these outflow channels are the result of 
the interaction between the tropical cyclone and its surrounding 
environmental flow features which may, to a large extent, be independent 
of the particular internal processes of the cyclone. 

7.2 Comparison of the Structural Characteristics of N.I.O. Tropical 

Cyclones with Western North Pacific Tropical Cyclones 

Many cross sections of tangential winds for individual N.I.O. 
tropical cyclones have been shown. These tangential circulation 
patterns are in most respects quite similar to the tangential 
circulation patterns of the individual northwest Pacific cyclone 
systems. To verify this statement we have averaged the tangential wind 
field for three time periods for N.I.O. TC 17-79 just before it attained 
its maximum intensity of 85 kts. Result are shown in Fig. 56 - diagram 
a. Three weak typhoon systems in the northwest Pacific, Typhoon Ellis 
(85 kts), Typhoon Cecil (80 kts), and Typhoon Irving (90 kts) are 
selected for comparison with TC 17-79 (which was of an intensity close 



Fig. 55. Variety of outflow patterns associated with tropical cyclone 
intensification (from Chen and Gray, 1984). 

to these systems). The same averaging was performed and is shown in 


Fig. 56 (diagram b). The results inside 3° radius are not reliable and 
have not been shown. Note how similar these two profiles are. The 
circulation about TC 17-79 is very much like that of the three minimal 
northwest Pacific typhoons. We can also make a comparison with the 10- 
year northwestern Pacific typhoon rawinsonde composite of Frank (1977), 
shown in diagram c of this figure. This latter rawinsonde composite 
result is for all northwest Pacific tropical cyclones of typhoon 
intensity (2 65 knots) or greater. These were of considerably stronger 
intensity than the average of our seven N.I.O. cases. 

























( qui ooi) d < qm 001) d ( quu QOI) d 


98 



(a) TC-17-79 
(N. 1.0.) 


(b)3 NWP 
SYSTEMS 


(C) WPD4 


Fig. 56. Diagram a: vertical cross section of tangential wind 

averaged at three time periods right before the tropical 
cyclone N.I.O. TC 17-79 attained its maximum intensity. Diagram 
b: average of three northwest Pacific minimal typhoons during 
the FGGE year. Diagram c: the 10-year northwest Pacific typhoon 
rawinsonde composite of Frank (1977). 











99 


It is interesting to see that the average results for the three 
weak typhoon cases and the 10-year typhoon composite closely resemble 
the three time-period composite of TC 17-79. Although results in all 
three cases are similar, there are some differences. TC 17-79 appears 
to have a stronger outer circulation as compared to those of northwest 
Pacific systems. We believe this to be due to the stronger equatorial 
monsoon trough that is present in the N.I.O. as compared with the 
Pacific. We also find that the outer circulation of N.I.O. systems (TC 
17-79, TC 18-79, TC 23-79, and TC 25-79) was generally stronger than 
that of most individual Pacific systems. The outer radius low-level 
cyclonic circulation also extends over a greater vertical extent for 
these four N.I.O. systems. Cyclonic winds extend up to 300-400 mb in 
the N.I.O. systems and even higher on their equatorial sides. It 
appears that N.I.O. tropical cyclones experience a stronger equatorward 
monsoonal influence and have a stronger large-scale surrounding low- 
level cyclonic circulation than similar tropical cyclones of the 
northwestern Pacific and the Atlantic. N.I.O. systems that do 
experience a more poleward than equatorwards cyclonic wind influence (as 
occurs in the late autumn periods of November and December) generally 
have a weaker, shallower and less well-defined low-level cyclonic 
circulation. 

Maximum Intensity o f N.I.O. Tropical Cyclones . Since the tropical 
cyclones in the N.I.O. form in relatively small semi—enclosed ocean 
basins, these systems generally have shorter lifetimes compared to those 
in the other ocean basins especially to the tropical cyclones of the 
northwest Pacific and north Atlantic ocean basins. There is a much 
higher probability that N.I.O. cyclones will move inland before 



100 


favorable large-scale conditions have sufficient time to initiate a 
substantial intensification. As discussed by Sadler and Gidley (1973), 
most N.I.O. systems dissipate due to landfall. Almost none take on 
extra-tropical characteristics. This is one of the major reasons that 
less than one-third of N.I.O. tropical cyclones reach typhoon intensity. 
(In the northwest Pacific and the Atlantic, roughly 60-70 percent of the 
tropical cyclones that form reach typhoon intensity.) Only a few N.I.O. 
systems are weakened or die due to being sheared off by strong 
baroclinic westerly wind patterns as often occurs at more polar 
latitudes in the other ocean basins. 

Another reason for the weaker tropical cyclones in the N.I.O. is 
the lack of a TUTT (Tropical Upper Tropospheric Trough), which is almost 
always present in the northwest Pacific during summer and often present 
in the Atlantic during the summer season. Many researchers (Sadler, 

1978; Holliday and Thompson, 1979) have emphasized the importance of the 
TUTT in initiating both the formation and the intensification of 
tropical cyclones through the development of favorable upper 
tropospheric poleward outflow channels. Because the TUTT is a mid- 
oceanic and primarily a mid-summer phenomenon it is not to be expected 
that it would be present during the spring and autumn transitional 
seasons when N.I.O. cyclones occur. The Asian continent to the north is 
also an inhibiting factor to the development of this oceanic upper 
trough system.. 

7.3 Characteristics of Motion of N.I.O. Tropical Cyclones 

We have described the tracks of FGGE year N.I.O. cyclones in 
section 4. FGGE year N.I.O. cyclones mostly moved on a northwesterly 
track except for TC 18-79 which went straight northward. TC 17-79 had 


101 


an erratic and looping track due the passage of an upper-level trough to 
its north. The average northwestward track for these FGGE year systems 
is the result of the prevailing easterly steering current. N.I.O. 
systems typically do not recurve or have much of an easterly track 
component because they occur at a lower latitude and move over land and 
die before they have a chance to be caught up in the westerly 
circulation. The Himalayan Mountain Chain also helps protect the N.I.O. 
from the influences of strong and deep westerly currents which, if they 
were to impinge on these storm systems, would recurve them to the east. 

The average speed of motion of these FGGE-year cyclones was about 
8-10 kts. This speed is .comparable to those of northwest Pacific 
tropical cyclones before they recurve but slower than Atlantic cyclones 
moving in the trade winds. Sadler and Gidley (1973) state that North 
Indian tropical cyclones move slower compared to the cyclones of the 
western Pacific. But Sadler and Gidley do not distinguish between the 
motion of northwest Pacific storms before and after they recurve. It is 
also to be noted that the variations of speed among different N.I.O. 
cyclones is much less than those of northwestern Pacific or western 
Atlantic systems. N.I.O. tropical cyclones occur over a smaller areal 
domain than do Pacific and Atlantic cyclones. They encounter fewer 
changes in their steering flow components. By contrast, tropical 
cyclones in the northwest Pacific occur over a very wide areal domain 
and throughout most of the year. Pacific systems thus encounter a 
greater variety of large-scale general circulation patterns. 

Consequently, the track and speed variability of Pacific cyclones is 
also greater. 

Matsumoto (1984) has recently developed a new 1-3 day TC 


102 


statistical track prediction scheme for the Atlantic, northwest Pacific 
and N.I.O. This scheme has shown itself, in early testing, to be very 
skillful. This is the only statistical scheme that has been developed 
in the N.I.O. which combines synoptic, persistence, and steering flow 
parameters. Matsumoto found that his prediction errors in the N.I.O. 
are 10-20% smaller than in the northwest Pacific and Atlantic. This is 
attributed to the generally slower motion of N.I.O. cyclones. Matsumoto 
also found that persistence is a better predictor in the N.I.O. and 
steering and synoptic data less skillful predictors in comparison with 
the northwest Pacific and the Atlantic. It is planned that the new 
statistical scheme undergoes good testing to determine its feasibility 
as an operational forecast aid. 


103 


8. ACKNOWLEDGEMENTS 

This research has been supported by the Naval Environmental 
Prediction Research Facility (NEPRF) under the encouragement of Samson 
Brand (Grant No. N00228-83-3122). We have utilized the computing 
facilities of the National Center for Atmospheric Research (NCAR) for 
ECMWF data analysis from the NSF Grant No. ATM-82-14041. NCAR is 
supported by the National Science Foundation. 

We would like to thank Captain Roger Edson for a critical review of 
this manuscript and to Cindy Schrandt, Patti Nimmo and Barbara Brumit 
for very competent support in manuscript preparation. 


104 


9. REFERENCES 


Arnold, C. P., 1977: Tropical cyclone cloud and intensity 

relationships. Dept, of Atmos. Sci. Paper No. 277, Colo. State 
Univ., Ft. Collins, CO, 154 pp. 

Askue, C., 1984: Varying structure and intensity change characteristics 
of four western north Pacific tropical cyclones. Dept, of Atmos. 
Sci., M.S. Thesis, Colo. State Univ., Ft. Collins, CO, 97 pp. 

Askue, C. and C. S. Lee, 1984: A study of the genesis and 

intensification of two 1979 northwest Pacific supertyphoons. 

Paper presented at the 15th Technical Conference on Hurricanes and 
Tropical Meteorology, January 9-13, Miami, Florida, 243-248. 

Bansal, R. K. and R. K. Datta, 1972: Certain aspects for 

intensification of tropical storms over Indian Ocean area. Indian 
J. Meteo . Geophvs .. 23, 4, 503-506. 

Chen, L. and W. M. Gray, 1984: Global view of the upper level outflow 

patterns associated with tropical cyclone intensity changes during 
FGGE. Paper presented at the 15th Technical Conference on 
Hurricanes and Tropical Meteorology, January 9-13, Miami, Florida, 


224- 

231. 







DeAngelis, 

325. 

D., 

1979: 

Hurricane alley. 

Mariners Weather Loc. 

23, 

322- 

DeAngelis, 

28. 

D., 

1980a: 

Hurricane alley. 

Mariners Weather 

Log, 

24, 

24- 

DeAngelis, 

110. 

D. , 

1980b: 

Hurricane alley. 

Mariners Weather 

Log, 

24. 

107- 

DeAngelis, 

429. 

D.. 

1980c: 

Hurricane alley. 

Mariners Weather 

Log, 

24, 

424- 

Dvorak, V. 

F., 

1975: 

Tropical cyclone intensity analysis 

and 




forecasting from satellite imagery. Mon . Wea. Rev ., 103 . 420-430. 

Frank, W. M., 1977: The structure and energetics of the tropical 

cyclone, I: Storm structure. Mon . Wea . Rev . 105 . 1119-1135. 

i 

Gray, W. M., 1968: Global view of the origin of tropical disturbances 
and storms. Mon. Wea. Rev., 96, 669-700. 


























105 


Gray, W. M., 1975: Tropical cyclone genesis. Dept, of Atmos. Sci. 

Paper No. 234, Colo. State Univ., Ft. Collins, CO, 121 pp. 

Gray, W. M., 1979: Hurricanes: t'heir formation, structure and likely 

role in the tropical circulation. Supplement to Meteorology Cver 
the Tropical Cceans . Published by RMS, James Glaisher House, 
Grenville Place, Bracknell, Berkshire, RG 12 1BX, D. B. Shaw, ed., 
155-218. 

Holland, G. J., 1984: On the climatology and structure of tropical 

cyclones in the Australian/southwest Pacific region, I: Data and 
tropical storms. Aust . Met . Mag., Vol 32, 1, 1-16. 

Holland, G. J., 1984: On the climatology and structure of tropical 
cyclones in the Australian/southwest Pacific region, II: 
Hurricanes. Aust . Met . Mag., Vol. 32, 1, 17-32. 

Holland, G. J., 1984: On the climatology and structure of tropical 

cyclones in the Australian/southwest Pacific region. III: Major 
hurricanes. Aust . Met . Mag .. Vol 32, 1, 33-46. 

Holland, G. J. and R. T.^Merrill, 1984: On the dynamics of tropical 
cyclone structural changes. Quart . J. Roy. Meteor . Soc., 110, 
723-745. 

Holliday, C. H. and A. H. Thompson, 1979: Climatological 

characteristics of rapidly intensifying typhoons. Mon . Wea. Rev ., 
107 , 1022-1034. 

Joint Typhoon Warning Center, 1979: Annual Typhoon Report. NAV0CEAN- 
C0MCEN/JTWC. Guam, 191 pp. 

Keen, R. A., 1982: The role of cross-equatorial tropical cyclone pairs 
in the southern oscillation. Mon . Wea . Rev .. 110 . 1405-1416. 

Krishnamurti, T. N., P. Greiman, Y. Ramanathan, R. Pasch and P. Ardanuy, 
1980: Quick look summer M0NEX atlas. Part I: Saudi Arabia 
Phase. FSU Report No. 80-4, 71 pp. 

Krishnamurti, T. N., P. Ardanuy, Y. Ramanathan, and R. Pasch, 1981: On 
the onset vortex of the summer monsoon. Mon. Wea. Rev., 109, 2. 
344-363. 

Lee, C. S., 1984: Large-scale influences on individual cases of 

tropical cyclone genesis during the FGGE year. Paper presented at 
the 15th Technical Conference on Hurricanes and Tropical 
Meteorology, January 9-13, Miami, Florida, 281-288. 

Love, G., 1982: The role of the general circulation in western Pacific 
tropical cyclone genesis. Dept, of Atmos. Sci., Paper No. 340, 
Colo. State Univ., Ft. Collins, CO 80523, 215 pp. 


















106 


Matsumoto, C., 1984: A statistical method for one- to three-day 

tropical cyclone track prediction. Dept, of Atmos. Sci. Fh.D. 
Dissertation, Colo. State Univ., Ft. Collins, CO, 200 pp. 

McBride, J. L., 1981: Observational analysis of tropical cyclone 

formation. Part III. Budget Analysis. J. Atmos . Sci . , 3_8, 1117- 

1131. 

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 , 1132-1151. 

Mooley, D. A., 1980: Severe cyclonic storms in the Bay of Bengal, 
1877-1977. Mon. Wea. Rev. , 108 . 1647-1655. 

Sadler, J. C., 1978: Mid-season typhoon development and intensity 
changes and the tropical upper tropospheric trough. Mon . Wea . 

Rev.. 106 . 1137-1152. 

Sadler, J. C. and R. E. Gidley, 1973: Tropical cyclones of the North 
Indian Ocean. ENVPREDRSCHFAC Tech. Paper No. 2-73, Naval 
Postgraduate School, Monterey, CA, 58 pp. 

Saha, K., F. Sanders, and J. Shukla, 1981: Westward propagating 

predecessors of monsoon depressions. Mon . Wea . Rev .. 109 , 2, 
330-343. 

Weatherford, C. and W. M. Gray, 1984: Relating typhoon intensity to 
outer 1-3° radius circulation as measured by reconnaissance 
aircraft. Paper presented at the 15th Technical Conference on 
Hurricanes and Tropical Meteorology, January 9-13, Miami, Florida, 
238-242. 

Xu, J. and W. M. Gray, 1982: Environmental circulations associated with 
tropical cyclones experiencing fast, slow and looping motion. 

Dept, of Atmos. Sci. Paper No. 346, Colo. State Univ., Ft. 
Collins, CO 80523, 111 pp. 











107 


COMSEVENTHFLT 

ATTN: FLT METEOROLOGIST 

FPO SAN FRANCISCO 96601-6003 


COMMANDER 

MIDDLE EAST FORCE 

FPO NEW YORK 09501-6008 


COMMANDING OFFICER 
USS SARATOGA (CV-60) 

ATTN: MET. OFFICER, OA DIV 
FPO MIAMI 34078-2740 


COMMANDING OFFICER 
USS KITTY HAWK (CV-63) 

ATTN: MET. OFFICER, OA DIV. 
FPO SAN FRANCISCO 96634-2770 


COMMANDING OFFICER 
USS CARL VINSON (CVN-70) 
ATTN: MET. OFFICER, OA DIV. 
FPO SAN FRANCISCO 96629-2840 


COMMANDING OFFICER 

USS NEW ORLEANS (LPH-11) 

ATTN: MET. OFFICER 

FPO SAN FRANCISCO 96627-1650 


COMMANDING OFFICER 
USS TARAWA (LHA-1) 

ATTN: MET. OFFICER 

FPO SAN FRANCISCO 96622-1600 


COMMANDING OFFICER 
NAVAL RESEARCH LAB 
ATTN: LIBRARY, CODE 2620 
WASHINGTON, DC 20390 


DISTRIBUTION 


COMSEVENTHFLT 

ATTN: NSAP SCIENCE ADVISOR 
BOX 167 

FPO SEATTLE 98762 


COMMANDER 

AMPHIBIOUS GROUP 1 

ATTN: METEOROLOGICAL OFFICER 

FPO SAN FRANCISCO 96601 


COMMANDING OFFICER 
USS CONSTELLATION (CV-64) 
ATTN: MET. OFFICER, OA DIV 
FPO SAN FRANCISCO 96635-2780 


COMMANDING OFFICER 
USS MIDWAY (CV-41) 

ATTN: MET. OFFICER, OA DIV. 
FPO SAN FRANCISCO 96631-2710 


COMMANDING OFFICER 
USS NEW JERSEY (BB-62) 

ATTN: MET. OFFICER, OA DIV. 
FPO SAN FRANCISCO 96688-1110 


COMMANDING OFFICER 
USS OKINAWA (LPH-3) 

ATTN: MET. OFFICER 

FPO SAN FRANCISCO 96625-1630 


COMMANDING OFFICER 
USS TRIPOLI (LPH-10) 

ATTN: METEOROLOGICAL OFFICER 
FPO SAN FRANCISCO 96626-1645 


COMMANDING OFFICER 
NAVWESTOCEANCEN 
BOX 113 

PEARL HARBOR, HI 96860 


COMSIXTHFLT 

ATTN: FLT METEOROLOGIST 
FPO NEW YORK 09501-6002 


COMMANDING OFFICER 
USS CORAL SEA (CV-43) 

ATTN: MET. OFFICER, OA DIV. 
FPO NEW YORK 09550-2720 


COMMANDING OFFICER 
USS ENTERPRISE (CVN-65) 

ATTN: MET. OFFICER, OA DIV. 
FPO SAN FRANCISCO 96636-2810 


COMMANDING OFFICER 
USS RANGER (CV-61) 

ATTN: MET. OFFICER, OA DIV. 
FPO SAN FRANCISCO 96633-2750 


COMMANDING OFFICER 
USS BELLEAU WOOD (LHA-3) 
ATTN: METEOROLOGICAL OFFICER 
FPO SAN FRANCISCO 96623-1610 


COMMANDING OFFICER 
USS PELELIU (LHA-5) 

ATTN: MET. OFFICER 

FPO SAN FRANCISCO 96624-1620 


OFFICER IN CHARGE 

U.S. NAVOCEANCOMDET 

FPO SAN FRANCISCO 96685-2905 


COMMANDING OFFICER 
U.S. NAVOCEANCOMCEN 
BOX 12, COMNAVMARIANAS 
FPO SAN FRANCISCO 96630-2926 



108 


DIRECTOR OF RESEARCH (2) 
U.S. NAVAL ACADEMY 
ANNAPOLIS, MD 21402 


COMMANDER (2) 

NAVAIRSYSCOM 
ATTN: LIBRARY (AIR-723D) 
WASHINGTON, DC 20361-0001 


AFGWC/DAPL 

OFFUTT AFB, NE 68113 


DIRECTOR 

NATIONAL HURRICANE CENTER 
NOAA, GABLES ONE TOWER 
1320 S. DIXIE HWY 
CORAL GABLES, FL 33146 


UNIVERSITY OF HAWAII 
METEOROLOGY DEPT. 
2525 CORREA ROAD 
HONOLULU, HI 96822 


DIRECTOR, JTWC 
BOX 17 

FPO SAN FRANCISCO 96630 


NAVAL POSTGRADUATE SCHOOL 
METEOROLOGY DEPT. 
MONTEREY, CA 93943 


COMMANDER 

NAVAIRSYSCOM (AIR-330) 
WASHINGTON, DC 20361-0001 


DIRECTOR (12) 

DEFENSE TECH. INFORMATION 
CENTER, CAMERON STATION 
ALEXANDRIA, VA 22314 


DIRECTOR 

FEDERAL EMERGENCY MANAGEMENT 
AGENCY (FEMA) 

WASHINGTON, DC 20472 


INSTITUTE FOR STORM RESEARCH 
UNIVERSITY OF ST. THOMAS 
3600 MT. VERNON 
HOUSTON, TX 77006 


BUREAU OF METEOROLOGY 
BOX 1289K, GPO 
MELBOURNE, VIC, 3001 
AUSTRALIA 


LIBRARY 

NAVAL POSTGRADUATE SCHOOL 
MONTEREY, CA 93943-5100 


USAFETAC/TS 
SCOTT AFB, IL 62225 


ACQUISITIONS SECT. IRDB-D823 
LIBRARY & INFO. SERV., NOAA 
6009 EXECUTIVE BLVD. 
ROCKVILLE, MD 20852 


COLORADO STATE UNIV. (6) 
ATMOSPHERIC SCIENCES DEPT. 
ATTN: DR. WILLIAM GRAY 
FORT COLLINS, CO 80523 


SCIENCE APPLICATIONS 
INTERNATIONAL CORP. (SAIC) 
205 MONTECITO AVE. 

MONTEREY, CA 93940 


DIRECTOR, INDIAN INST. OF 
TROPICAL METEOROLOGY 
RAMDURG HOUSE 
PUNE 411-005, INDIA 



L’219270