TAO, Vo1.6, No.

4, 591-606, December 1995

Pre-Monsoon Thunderstorm Activity Over Bangladesh

From 1983 to 1992
MD. ABDUL MANNAN CHOWDHURyl , 2 and UTPAL KUMAR DEI

(Manuscript received 13 September 1994, in final form 21 November 1995)

ABSTRACT

Daily and 3-hourly weather information in code figures and daily rainfall
data from 19 stations in Bangladesh for the months of March through May,
from 1983 through to 1992, have been analysed to give some understanding
to the space/time distributions as well as to the nature of the thunderstorms
throughout the country. The Statistical characteristics of these disturbances
are presented.
It turns out that over Bangladesh thunderstorm (TS) occurrence has a
wide variation in terms of both space and time. Whereas in Sylhet TS. days
make up 50% of the premonsoon season, in Maijdee they comprise only 10%.
However, in Maijdee precipitation per TS day is abnormally high. In central
Bangladesh, the most dominant time of TS occurrence is either late afternoon
or early evening, unlike that in the north where the corresponding time is
either night or early morning. In general, the frequency of TSs as well as the
amount of precipitation are greatest in the month of May almost all over the
country with the average rainfall per TS day being quite high though dry TS
days are still quite numerous.
TAO

(Key words: Pre-monsoon thunderstorm, Nor'wester, Correlation co-efficient,

Spatial Contour)

1. INTRODUCTION

Bangladesh is a land of rivers and branching of these rivers and the canals is widespread
throughout the country (Figure 1) leading to the conclusion that, to a large extent, they control
the weather and climate of the country. The usual wind patterns in the Bangladesh region are
north-east trade winds on the surface and the upper air westerlies. Nonetheless,Bangladesh
also falls within the Indian southwest monsoon system, and the surface winds during the
monsoon time are SW/S/SE, though the upper wind remains westerly. The premonsoon phase

I Physics Department, Jadavpur University, Calcutta-700032, India

2 Permanent address: Department of Physics, Jahangir Nagar University, Savar, Dhaka, Bangladesh

591
592 TAO, Vol.6, No.4, December 1995

oI 25 SOMile
I !
N

o 25 5'[Link]

OF BEN GAL

Fig. 1. Rivers and canals of Bangladesh.

i.e. March through May, is a transition phase, when many a violent thunderstorm (TS) affect
the region. During this period, the subtropical high shifts to the Bay of Bengal with two other
heat lows falling around Bangladesh. One is in the Assam region in the east and the other
is dominantly in the Madhya Pradesh region of India to the west. In their wake, the surface
winds are, in general, SW/S in south Bangladesh, while the surface winds in the northern
region are easterlies due to the topography. During this period, a strong westerly jet appears
over the country at the upper tropospheric level which generates associated atmospheric
TAO

instabilities.
This phase is characterized by high .surface temperatures and convective developments
because of copious moisture support and the presence of atmospheric instabilities. In some
developments, giant cumulonimbus clouds are produced resulting in severe thunderstorms
which cause huge losses in terms of property and lives. However, these thunderstorms also
provide principal rainfall patterns during the period and are the main source of rains until the
southwest monsoon sets in.
In fact, the TS activity in the Bangladesh region is supposed to have three distinct
causes:
In the southern part, as stated previously, the surface wind during the premonsoon period
is SW/S in general. With this wind comes sea moisture. Besides this, the southern land area
is covered by a large number of rivers, creeks and water bodies in various forms. All these
give enough moisture support for TS activity.
In the adjacent western part of the country lies the Gangetic plain of West Bengal.
There, the cause of the TS activity is mainly the instabilities produced in the hilly tracts of the
 Md. Abdul Mannan Chowdhury & Utpal Kumar De 593

adjacent states of Bihar and Orissa. As the downdrafts meet the sea moisture-laden SE surface
winds of the region, successive TS developments occur. It appears as if the disturbances are
propagating from the NW direction, resulting in the premonsoon TS activity in Gangetic
west Bengal being popularly known as Nor'westers (local name of SLS). Incidentally, in the
middle region of Bangladesh many of the downdrafts of Nor'westers arrive and, with the
help of moisture from large rivers like Padma and Meghna, these disturbances regenerate
into TS activity.
The northern part of Bangladesh is greatly influenced by the wind system of the foothill
zones of the Himalayas. The surface winds are usually easterly here, and due to the orogra-
phy, these easterlies frequently cause TS developments. Additionally, sometimes a western
disturbance strengthens and its influence supposedly makes TS developments occur.
The average surface pressure and wind patterns for the month of April are presented in
Figure 2. However, a sharp demarcation of the three zones for all three months cannot be
expected.
Various aspects of TS activity have been studied and published in journals by such
authors as Rai Sircar (1953), Koteswaran and Srinivasan (1958), Sen and Gupta (1961),
De and Sen (1961), Nandi and Mukherjee (1966), Mukherjee and Bhattacharya (1972) and
Mukherjee et al. (1977). In all these studies, a number of important results have been
obtained with regards to the formation,structure, timing and devastations of Nor'westers.
However, almost all of these studies focus on India. In fact,as for Bangladesh, only a
limited number of research works on thunderstorms done by the Bangladesh Meteorological

60 70 80 90 100
11
H 1012 LEGEND
1014
-CALM ~ OSkts -,10kts
1010
1 """,,15 kls
"""" 20kts .., SOkis
1012 - Wind direction al 03 GMT
0-Wind direction at 12 GMT

30 30

TAO

60 70 80 90

Fig. 2. Pattern of mean pressure (mb) and surface winds in April. B indicates
Bangladesh.
594 TAO, Vol.6, No.4, December 1995

Department was presented in the seminar on "SEVERE LOCAL STORMS" organized by the
South Asian Association for Regional Co-operation (SAARC) countries.
The present study is the first of its kind in the context of Bangladesh, and it is believed
that the results (of this study) will be useful in making objective forecasts of TSs in the
country.
In the present study, 3-hourly interval surface observational data as well as daily rainfall
data have been analysed over a ten-year interval (1983-92) for 19 stations scattered all over
Bangladesh. The study has essentially been confined to the three-month period before the
onset of monsoons i.e. March, April and May.

2. GEOGRAPHICAL LOCATION OF BANGLADESH

Bangladesh extends from 20° -45'N to 26° -40'N and 88° -05'E to 90° -40'E. Most of
the country is flat except for hills in the southeastern parts. In fact, the country is surrounded
by the Assam Hills to the east and by the Meghalaya Plateau (i.e. Shillong Plateau) to
the north, with the lofty Himalayas beyond. The Bay of Bengal lies to the south of the
country. To the west is the contiguous Plain of west Bengal and the Gangetic Plain of India.
The geographical setting of Bangladesh and the location of the observational stations in the
present study with their heights are shown in Figure 3.

3. DATA AND STATION LOCATION

Daily and 3-hourly weather information data in code figures for the months of March
through May, and daily rainfall data for the same months, for the period from 1983 to 1992
for 19 stations in Bangladesh were obtained from the Bangladesh Meteorological Department
(BMD). It should be noted that the present study is confined to surface observations which
include pressure,dry bulb temperature, dewpoint temperature, wind velocity and direction,
the present weather as well as that in the past three hours at every three-hour interval starting
from 00 GMT. The stations have been chosen in such a way as to represent almost every
zone of the country.
Starting from the north of the country, the following stations are considered for the
TAO

study:
Rangpur (RNP 859), Dinajpur (DNP 862), Bogra (BGR 883), Rajshahi (RJSH 856),
Ishurdi (ISDI 907), Mymensingh (MYMS 886), Sylhet (SLT 891), Comilla (CML 933),
Dhaka (DCA 923), Faridpur (FDP 929), Jessore (JSR 936), Khulna (KLN 947), Bansal (BSL
950), Patuakhali (PTKI 955), Khepupara (KPRA 984), Maijdee (MJD 934), Chittagong (CTG
971), Rangamati (RMGT 943) and Cox's Bazar (C'BZR 992). The code name and the station
codes are given within brackets. The last three digits begin with 41.
Though the two stations, Srimangal and Satkhira are shown on the map, because of the
lack of data, they are excluded from this analysis.
4. ANALYSES
In the present work, the following statistical studies are presented to provide an un-
derstanding of the space/time distributions as well as the nature of thunderstorms over
Bangladesh.
(1) Monthly and biweekly frequency distributions of TSs during the premonsoon months.
 Md. Abdul Mannan Chowdhury & Utpal Kumar De 595

Sylhet
Mymensingh 035 25·
°19
:l>
III
III
l>
2( 3:
24·
-
:I:
r-
r-
III

23·
2'!

Fig. 3. Geographical setting of Bangladesh. The circles and numerical values

indicate the location and height (above mean sea level in meters) of the
stations used in this study. Shaded areas are in excess of 250 meters
TAO

above sea level. Some elevations are in excess of 1000 meters (Assam
Hills and the Meghalaya Plateau).

(2) Correlations of the number of annual TS days over premonsoon months of different
stations with three representative stations viz., DCA, RNP and KPRA.
(3) Percentage occurrence of TS at different 3-hourly time intervals per day.
(4) Average rainfall distribution per TS day over each biweekly interval.

4.1. TS Frequency Analysis

In this analysis, the identification of TS days has been made from weather information.
By international agreement, a "thunderstorm" day is defined as a local calender day on which
thunder is heard, regardless of the actual number of thunderstorms ( WMO, 1953).
Monthly and two-week frequency distributions of TS days are shown in Table 1. From
this and from Figure 4, it is evident that an unusually high number of TS days occur in Sylhet
596 TAO, Vo1.6, No.4, December 1995

Monthly and two-week frequency distributions of TS days are shown in Table 1. From
this and from Figure 4, it is evident that an unusually high number of TS days occur in Sylhet
(SLT). SLT is just at the foot of the Meghalaya Hills, and this orographical feature plays an
important role in enhancing TS days.

25~0

24·58

t 24·05

QI
-g.... 23·53
:;:;
c
....J 23000

22·48

21·96

21.43 L.....I---I-..LL....J-..J......,;L...ll--l--1-....I...-.L......I---I--L....J-..L....Jl....L.J......J....-1-J
88.60 89·11 89·63 90·14 90·66 91·17 91·69 92·20
Longitude - -
Fig. 4. Frequency distribution of TSs over Bangladesh in the pre-monsoon season
from 1983-1992.
TAO

The lowest frequency of TS is in Maijdee. In twelve other stations, 20% or more of

the days considered agree well with the average TS days. For all the stations, the monthly
TS frequency increases systematically from March to May. This is quite expected as sun
becomes more overhead from March to May, thereby increasing the temperature as well as
evaporation. As all of Bangladesh is covered by innumerable water bodies, the moisture
support increases from March to May. This is distinctly seen in the present study.
From Table 1, it also clear that, in general, during March and April, the number of TS
days in the second two-week period is more than in the first. The exception occurs only in
five stations in March. However, in twelve stations TS frequency is highest in the first half
of May. Another notable feature, TS frequency sharply rises at RNP and MYMS from the
second half of April. Such an increase in TS days in the first half of May at most of the
stations is not unexpected, as the number of western disturbances is high during this period.
The number shows a decline with the approach of a monsoon.
The information of TS frequency on the two-week scale is important for Bangladesh
which is still predominantly an agrarian state.
 Md. Abdul Mannan Chowdhury & Utpal Kumar De 597

Table 1. Biweekly and monthwise frequency distributions of TS days. I, II and

T represent 1st half, 2nd half and the total of the monthly frequency
distributions, respectively.
Station March April May
I II T I II T I II T
RNP 1.2 1.5 2.7 1.5 5.0 6.5 7.6 7.0 14.6
DNP 0.6 0.5 1.1 0.9 2.0 2.9 3.3 3.7 7.0
BGR 1.0 1.5 2.5 2.1 3.4 5.5 5.7 5.6 11.3
RJSH 1.2 1.9 3.1 1.9 3.7 5.6 4.8 5.9 10.7
Ism 1.2 1.2 2.4 2.0 3.8 5.8 5.0 5.2 10.2
MYMS 1.2 1.9 3.1 1.9 5.7 7.6 7.2 6.5 13.7
SLT 3.9 6.1 10.0 9.7 10.4 20.1 10.7 10.1 20.8
CML 1.6 2.4 4.0 2.6 3.9 6.5 5.2 4.7 9.9
DCA 1.9 3.3 5.2 3.2 5.3 8.5 6.2 6.0 12.2
FDP 2.0 2.8 4.8 3.5 4.7 8.2 6.8 6.2 13.0
JSR 2.2 1.8 4.0 2.1 4.5 6.6 6.2 5.9 12.1
KLN 1.8 1.6 3.4 2.8 3.3 6.1 5.3 4.9 10.2
BSL 1.5 1.8 3~3 2.1 3.1 5.2 3.5 3.4 6.9
PTKI 1.8 1.7 3.5 1.9 3.7 5.6 5.6 4.4 10.0
KPRA 0.8 1.6 2.4 1.2 2.0 3.2 3.3 2.9 6.2
MJD 0.9 0.8 1.7 0.9 2.1 3.0 2.5 2.1 4.6
CTG 0.8 2.0 2.8 1.8 4.0 5.8 4.1 4.5 8.6
RMGT 1.2 1.4 2.6 2.7 3.5 6.2 3.7 3.8 7.5
C'BZR 0.5 0.7 1.2 1.2 2.5 3.7 3.5 4.2 7.7

4.2 Correlation Study of Annual TS Days Over Premonsoon Months With Three
TAO

Representative Stations
Dhaka, the capital of Bangladesh is considered the central point of the country by
the Bangladesh Meteorological Department and also representative of the middle region of
Bangladesh, where the origin of TSs may be [Link] to the same cause. This is strongly
supported by a similar frequency of occurrence of TS over a large region of the middle of
the country as discussed in Section 4.3. Thus, an attempt has been made to understand the
correlation of TS days between Dhaka and the rest of the stations of the country on the basis
of monthly occurrence over the ten-year period. In Figure 5, the spatial contour of correlation
co-efficient of TS days with respect to Dhaka is presented.
Pearson's correlation co-efficients between three representative stations, DCA, KPRA
and RNP of three regions of Bangladesh, and the rest of the stations are listed in Table 2.
If the correlation co-efficient for Dhaka and any station is greater or equal to 0.5 for all
three months, then that station may be considered well correlated with Dhaka. On the other
hand, when the correlation co-efficient for all three months is above 0.45, the station may be
taken as good correlation with Dhaka.
598 TAO, vu,e, No.4, December 1995

25.28

24.93

24.58

24.23

f 23.88
QI

" 23.53
:::J

21.43L-J>--'-~-"-.[Link]'-L.I.....ll..-C>....I.-L.....I--Jl--'-...!1.......J..l..L.....I--L--'-..J
88.60 88.94 89.28 89.63 89.97 90.31 90.66 91.00 91.34 91.68 92.03
Longitude _

Fig. 5. The spatial Contour of Correlation co-efficients of various stations with

Dhaka in the pre-monsoon season from 1983-1992. Dhaka is marked
with a dot.

With these rules, it turns out from Table 2 that BGR, RISH, SLT, CML, PDP and RMGT
are well correlated with Dhaka, whereas ISDI and MID have a good correlation. This simple
correlation study possibly indicates the extent of the middle region of Bangladesh from the
standpoint of TS activity over all three months. But this study also shows that in the month of
March, TS activity in all regions is highly interrelated. In fact March is a highly transitional
month when surface easterlies are slowly withdrawing from the lower part of the country TAO

and the two heat lows around the nation are not so prominent.
A similar correlation study can be carried out with a representative station of the northern
part of Bangladesh and another representative station in the southern part. Such studies are
conducted with RNP and KPRA. As expected, both these stations bear good correlation with
all stations of Bangladesh in the month of March. But in the case of KPRA, stations like
C'BZR, CTG bear correlation co-efficients greater than 0.5 for all three months. In contrast,
stations like ISR, KLN, PTKI and BSL bear good correlation for the months of March and
May, but the corresponding co-efficient is less in the month of April (Table 2). In the case
of RMGT, the co-efficient is very poor in the month of May, but it has a high value in the
other two months. For the, other stations, the correlation co-efficient with KPRA is erratic,
making them not worthy of presentation in Table 2.
In the case of RNP, the northern stations like DNP, BGR, RJSH, ISDI bear good
correlation co-efficients in the month of April also (Table 2). But the situation drastically
changes in the month of May. This shows the complexity of the wind pattern, moisture
availability and the role of local orography in the northern part as the monsoon approaches.
 Md. Abdul Mannan Chowdhury & Utpal Kumar De 599

Table 2. Correlation co-efficients between three representative stations, DCA,

KPRA and RNP and the rest of the other stations.
Station March April May
RNP-DNP 0.52 0.49 0.03
DNP-DCA 0.51 0.16 0.56
BGR-DCA 0.63 0.80 0.85
RJSH-DCA 0.61 0.53 0.52
ISDI-DCA 0.72 0.63 0.40
MYMS-DCA 0.53 -0.01 0.57
SLT-DCA 0.77 0.77 0.84
CML-DCA 0.97 0.63 0.65
FDP-DCA 0.91 0.88 0.72
JSR-DCA 0.79 0.45 -0.04
KLN-DCA 0.92 0.42 -0.34
BSL-DCA 0.79 0.87 0.16
PTKI-DCA 0.67 0.65 0.07
KPRA-DCA 0.55 0.44 -0.07
MJD-DCA 0.65 0.46 0.47
CTG-DCA 0.68 0.73 0.12
RMGT-DCA 0.61 0.59 0.75
C'BZR-DCA 0.51 0.40 0.10
JSR-KPRA 0.53 0.49 0.66
KLN-KPRA 0.80 0.27 0.52
BSL-KPRA 0.81 0.40 0.83
PTKI-KPRA 0.84 0.30 0.67
CTG-KPRA 0.61 0.60 0.66
TAO

RMGT-KPRA 0.76 0.91 0.17

C'BZR-KPRA 0.82 0.50 0.84
DNP-RNP 0.74 0.62 -0.07
BGR-RNP 0.73 0.58 -0.05
RJSH-RNP 0.72 0.68 0.18
ISDI-RNP 0.78 0.70 0.32

4.3 TS Occurrence at Different Three Hour Intervals of a Day

In this section the percentage of occurrence of TS in each 3-hour interval during the
months of March, April and May is studied for each station over the concerned period of ten
years. The time stated in Table 3A and 3B is local Bangladesh time i.e. six hours are added
to G.M.T. In this analysis, if the occurrence is 75% or above for a 12-hour period and if it
is 50% or above for a 6-hour period and if it is 25% or above for a 3-hour interval, then
600 TAO, Vol.6, No.4, December 1995

Table 3A. Percentage occurrence of TS in each three-hour interval.

STATION 00-03hrs. 03-06hrs. 06-09hrs. 09-12hrs. 12-15hrs. 15-I8hrs. 18-21hrs. 21-24hrs.

RNP 19.0 18.6 12.7 8.8 8.5 7.9 9.6 14.9
DNP 11.5 23.5 15.7 12.6 9.0 10.8 5.4 11.5
BGR 14.2 11.7 12.8 9.6 10.3 12.5 11.4 17.5
RJSH 4.1 5.1 7.3 6.6 13.9 27.8 23.1 12.1
ISm 7.1 5.0 7.7 7.1 13.4 22.0 20.7 17.0
MYMS 12.5 12.8 9.4 9.2 11.8 17.2 13.0 14.1
SLT 15.8 15.3 11.7 10.7 8.2 10.3 12.0 16.0
CML 7.1 7.5 6.9 8.6 14.3 21.8 20.3 13.5
DCA 8.7 6.4 5.4 5.2 12.3 23.5 23.2 15.3
FDP 6.5 5.1 5.7 4.9 14.1 27.1 23.5 13.1
JSR 6.6 5.3 3.7 4.8 13.8 23.0 27.8 15.0
KLN 8.0 5.5 6.1 5.8 12.2 24.3 25.0 13.1
BSL 6.2 3.5 5.8 5.1 14.8 27.2 26.5 10.9
PTKI 7.2 7.2 6.0 12.0 14.8 22.0 18.1 12.7
KPRA 7.9 10.8 8.4 10.8 17.7 17.3 16.8 10.3
MID 4.6 7.7 4.7 16.3 10.9 29.4 15.5 10.9
CTG 16.7 14.1 10.6 9.4 10.0 11.4 10.5 17.3
RMGT 9.3 9.2 9.0 12.0 14.2 17.6 15.1 13.6
C'BZR 10.7 14.5 15.9 11.2 11.7 11.2 14.5 10.3

Table 3B. 12 hrs, 6 hrs and 3 hrs of maximum TS activity with respective percentage
of occurrence.

Station Maximum periods Total hours Maximum % of

of occurrence occurrence
TAO

RNP: 21 hrs.- 09 hrs. 12 hrs. 65.2

00 hrs.- 06 hrs. 6 hrs. 37.6
00 hrs.- 12 hrs. 3 hrs. 19.0
DNP: 00 hrs.- 12 hrs. 12 hrs. 63.3
03 hrs.- 09 hrs. 6 hrs. 39.2
03 hrs.- 06 hrs. 3 hrs. 23.5 (V)
BGR: 15-18,21-24,00-03 12 hrs. 56.9
& 06 - 09 hrs.
21 hrs.- 03 hrs. 6 hrs. 31.7
21 hrs.- 24 hrs. 3 hrs. 17.4
RJSH: 12 hrs.- 24 hrs. 12 hrs. 76.9 (I)
15 hrs.- 21 hrs. 6 hrs. 50.9 (I)
15 hrs.- 18 hrs. 3 hrs. 27.8 (I)
 Md. Abdul Mannan Chowdhury & Utpal Kumar De
601

Table 3B. (Continued.)

Station Maximum periods Total hours Maximum % of
of occurrence occurrence
Ism: 12 hrs.- 24 hrs. 12 hrs. 74.1 (V)
15 hrs.- 21 hrs. 6 hrs. 43.7
15 hrs.- 18 hrs. 3 hrs. 22.0 (V)
MYMS: 15-24, 3-6 hrs. 12 hrs. 57.1
15-18,21-24 hrs. 6 hrs. 31.4
15 hrs.- 18 hrs. 3 hrs. 17.2
SLT: 18 hrs.- 06 hrs. 12 hrs. 59.1
21 hrs.- 03 hrs. 6 hrs. 31.8
21 hrs.- 24 hrs. 3 hrs. 16.0
CML: 12 hrs.- 24 hrs. 12 hrs. 69.9
15 hrs.- 21 hrs. 6 hrs. 42.1 (V)
15 hrs.- 18 hrs. 3 hrs. 21.8 (V)
DCA: 12 hrs.- 24 hrs. 12 hrs. 74.3 (V)
15 hrs.- 21 hrs. 6 hrs. 46.7 (V)
15 hrs.- 18 hrs. 3 hrs. 23.5 (V)
FDP: 12 hrs.- 24 hrs. 12 hrs. 77.9 (I)
15 hrs.- 21 hrs. 6 hrs. 50.6 (I)
15 hrs.- 18 hrs. 3 hrs. 27.1 (I)
JSR: 12 hrs.- 24 hrs. 12 hrs. 79.6 (I)
15 hrs.- 21 hrs. 6 hrs. 50.8 (I)
18 hrs.- 21 hrs. 3 hrs. 27.8 (I)
KLN: 12 hrs.- 24 hrs. 12 hrs. 74.7 (V)
15 hrs.- 21 hrs. 6 hrs. 49.4 (V)
18 hrs.- 21 hrs. 3 hrs. 25.0 (I)
BSL: 12 hrs.- 24 hrs. 12 hrs. 79.4 (I)
15 hrs.- 21 hrs. 6 hrs. 53.7 (I)
15 hrs.- 18 hrs. 3 hrs. 27.2 (I)
PTKI : 12 hrs.- 24 hrs. 12 hrs. 67.5
15 hrs.- 21 hrs. 6 hrs. 40.1 (V)
15 hrs.- 18 hrs. 3 hrs. 22.0 (V)
KPRA: 09 hrs.- 21 hrs. 2 hrs. 62.6
12 hrs.-18 hrs. 6 hrs. 35.0
TAO

12 hrs.- 15 hrs. 3 hrs. 17.7

MJD: 09 hrs.- 21 hrs. 2 hrs. 72.1 (V)
09-12,15-18 hrs. 6 hrs. 45.7 (V)
15 hrs.- 18 hrs. 3 hrs. 29.5 (I)
CTG: 00-06,15-18 & 12 hrs. 59.5
21 hrs.- 24 hrs.
21 hrs.- 03 hrs. 6 hrs. 34.0
21 hrs.- 24 hrs. 3 hrs. 17.3
RMGT: 12 hrs.- 24 hrs. 12 hrs. 60.5
15 hrs.- 21 hrs. 6 hrs. 32.7
15 hrs.- 18 hrs. 3 hrs. 17.6
C'BZR: 03-09, 12-15 & 12 hrs. 56.5
18 hrs.- 21 hrs.
03 hrs.- 09 hrs. 6 hrs. 30.4
06 hrs.- 09 hrs. 3 hrs. 15.9
602 TAO, Vo1.6, No.4, December 1995

these periods are termed intensive periods of thunderstorm activity. On the other hand, if
the coresponding percentage are above 70, 40 and 20% respectively, then these periods are
termed vulnerable periods of thunderstorm activity.
Table 3A presents the percentage of occurrence of TS in each 3-hour interval. Table 3b
presents 12 hours, 6 hours and 3 hours of maximum thunderstorm activity, with the respective
percentage of occurrence and indicates whether it is intensive, vulnerable or otherwise.
It is observed that ten out of 19 stations have 12 hours maximum thunderstorm activity
from 12.00 hours to 24.00 hours. These ten stations lie mostly in central Bangladesh. These
ten stations, namely RJSH, ISDI, DCA, PDP, JSR, KLN, BSL, CML and PTKI all have the
identical six hours of maximum thunderstorm activity i.e. 15 hours to 21 hours. However,
of the ten stations,8 have the same 3 hours of maximum thunderstorm activity i.e. from 15
hours to 18 hours. The two exceptions, namely JSR and KLN have a different 3 hours of
maximum thunderstorm activity i.e. 18 hours to 21 hours. The notable point about these ten
stations is that all the different periods of maximum thunderstorm activity fall either into the
intensive or vulnerable class with few exceptions.
Of the 9 stations having different 12 hour maximum thunderstorm activity, five, (namely
DNP, RNP, BGR, MYMS and SLT) are located in north Bangladesh. On the other hand,
the other four stations namely MJD, CTG, C'BZR and KPRA are located mostly in the
extreme southern part. Stations CTG and C'BZR are located on the Hill tract and, as already
mentioned, station SLT is just at the foot of the Meghalaya Plateau. Hence, orography plays
an important role in the thunderstorm activity of these three stations. However, KPRA is
located on a plain.
Of the five northern stations above, four have a 3-hour period of maximum thunderstorm
activity either at night or in the early morning. The only exception is MYMS where a 3-hour
period of maximum activity is from 15 hours to 18 hours, like most stations in the central
region.
Of the four southern Bangladesh stations,each have a distinct 3-hour period of maximum
thunderstorm activity. In CTG, night time (21 hours to 24 hours) thunderstorm is active,
but morning (06 hours to 09 hours) thunderstorm is dominant in C'BZR. In KPRA the
corresponding period is just after noon (12 hours to 15 hours). The most notable thing is in
TAO

the case of MJD, where the corresponding period is in the afternoon (15 hours to 18 hours)
under intensive class with nearly 30% of all thunderstorms occurring within this short span.
In fact, the 9 stations in northern and extreme southern areas have hardly any period of
thunderstorm activity which might fall in either the intensive or vulnerable class.
The notable exception is MJD where all 3 periods of different durations come either
under intensive or vulnerable class. The only exception is DNP where the 3 hours of max-
imum thunderstorm activity in the early morning hours (03 hours to 06 hours) are included
in [Link] category.
Now, when the correlation study and the different periods of maximum thunderstorm
activity are compared, it seems that the origin of most of the thunderstorms for DCA, RJSH,
ISDI, CML and PDP is identical. Apparenly RMGT may also be included in this family, but
many of the thunderstorms there also have a different origin as the thunderstorms there are
more evenly distributed over a day. On the other hand, it can be concluded that most of the
thunderstorms of JSR and KLN are expected to have a common origin.
 Md. Abdul Mannan Chowdhury & Utpal Kumar De 603

4.4 Rainfall Per TS Day

In the present study, any rainfall from March to May is assumed to be due to TS activity
only. In Table 4, rainfall in mm per TS day over each two-week period and averaged over
the ten-year interval is presented.
It is evident that the tendency for rainfallffS days in general increase from March to
May even biweekly, though there are many departures from this. In fact, such a uniform rise
over each successive two-week is observed only in DCA and DNP.
In 14 out of 19 stations, the maximum rainfallffS occurred in the second half of May.
In 3 stations this occurred in the first half of May. But station C'BZR has some exceptional
behaviour since the maximum rainfallffS occurred in the first half of March. There the
rainfallffS day in the second half of March is very low compared to the rainfallffS day in
other halves.
A notable feature occurs in the case of MJD where the rainfallffS day is always very
high and it increases up to 82.8 mmfTS day in the second half of May. This amount is
almost twice that during highest rainfallffS day as that of any other station. On the other
hand, MJD has the lowest frequency of TS occurrence,. which shows that TS occurrence is
comparatively a rare phenomena but whenever it does occur, it generates huge precipitation.
Lastly, it should be pointed out that, in principle, the rainfallffS day is greater in the
second half of a month, but it is obvious from Table 4 that there are many departures.

Table 4. Rainfall (mm) per TS day.

Station March April May

I II I II I II
RNP 9.5 13.4. 11.5 18.6 15.2 26.8
DNP 14.6 15.4 18.2 19.4 26.4 40.7
BGR 6.2 3.5 4.8 14.7 28.2 17.0
RJSH 6.5 3.4 8.4 8.7 13.2 15.8
ISm 11.0- 16.1 8.6 0.6 17.8 21.1
MYMS 13.9 9.4 23.2 11.8 25.1 34.2
TAO

SLT 8.7 27.0 17.0 20.8 23.8 34.8

CML 8.3 22.8 21.1 16.9 30.4 33.4
DCA 6.6 16.4 19.2 20.8 23.9 30.0
FDP 7.8 19.6 13.5 17.1 17.9 27.4
JSR 10.3 14.0 7.3 8.1 8.4 18.1
KLN 12.9 9.9 4.7 19.2 19.5 22.7
BSL 11.2 10.9 16.8 18.7 34.4 35.4
PTKI 18.7 6.9 18.7 13.6 38.0 25.3
KPRA 27.3 9.4 13.6 27.8 29.9 33.8
MJD 28.1 39.3 53.7 48.4 71.8 82.8
CTG 13.6 12.0 35.2 23.5 27.1 33.4
RMGT 13.4 22.5 33.0 25.1 42.8 32.3
C'BZR 40.4 6.9 39.8 23.6 29.5 28.5
604 TAO, Vo1.6, No.4, December 1995

4.5 Percentage Occurrence Of Different Types Of TS

The occurrence of TSs may be classified according to the WMO code. The TS with
codes 17 and 29 may be considered dry, where those with codes 91-99 are termed wet. Any
code from 96-99 may also be taken as a severe TS.
Accordingly, all TSs occurring at a particular station over a ten-year period are grouped
by the percentage of occurrence under three columns in Table 5.

Table 5. Different types of TS occurrence in Bangladesh (in % form).

Station TS TS TS
17 & 29 91 - 95 96 - 99
RNP: March 70.5 25.0 4.5
April 58.4 33.3 8.3
May 61.0 36.4 2.6
DNP: March 78.6 14.3 7.1
April 64.1 23.1 12.8
May 60.2 34.5 5.3
BGR: March 53.7 43.9 2.4
April 70.5 25.6 3.9
May 64.8 29.6 5.6
RJSH: March 68.6 27.2 4.2
April 61.3 32.5 6.2
May 65.5 32.4 2.1
ISm: March 58.7 27.1 14.2
April 78.4 18.0 3.6
May 68.6 25.9 5.5
MYMS: March 78.4 19.6 2.0
April 74.2 24.2 1.6
May 58.6 39.8 1.6
SLT: March 70.3 25.9 3.8
April 65.1 32.2 2.7
May 53.6 45.5 0.9
CML: March 60.8 32.6 6.6
April 49.6 48.6 TAO
1.8
May 58.0 37.1 4.9
DCA: March 47.0 36.1 16.9
April 59.3 28.0 12.7
May 45.1 37.9 17.0
FDP: March 71.1 22.9 6.0
April 70.5 25.9 3.6
May 60.4 33.9 5.7
JSR: March 67.2 26.6 6.2
April 56.6 38.1 5.3
May 58.2 38.3 3.5
KLN: March 60.9 28.2 10.9
April 68.9 24.3 6.8
May 51.5 36.8 11.7
BSL: March 76.5 19.6 3.9
April 68.3 31.7 0.0
May 61.0 36.6 2.4
 Md. Abdul Mannan Chowdhury & Utpal Kumar De 605

Table 5. (Continued.)
Station TS TS TS
17 & 29 91 - 95 96 - 99
PTKI: March 47.5 44.0 8.5
April 58.9 33.7 7.4
May 47.4 39.6 13.0
KPRA: March 66.7 30.9 2.4
April 61.0 39.0 0.0
May 59.8 39.2 1.0
MID: March 58.8 35.3 5.9
April 26.1 54.3 19.6
May 43.9 53.0 3.1
CTG: March 43.5 56.5 0.0
April 38.1 51.4 10.5
May 36.4 57.6 6.0
RMGT: March 66.7 31.1 2.2
April 49.2 36.9 13.9
May 59.7 25.5 14.8
C'BZR : March 60.0 40.0 0.0
April 37.3 59.7 3.0
May 38.9 55.7 5.4

In general, the dry TS has dominance over the wet type, but there are 7 stations where
the situation differs. Five of the seven stations are in the southern part where moisture support
from the sea is copious and two other stations, namely DCA and CML, get enough moisture
support from nearby big rivers. In fact, CTG is the only station where wet TSs dominate
over dry TSs in all three months, though the severe type dominates only in April.
At four out of the seven stations, namely DCA, PTKI, MJD and C'BZR, in two out of
the three months, wet TSs predominate over dry varieties. In MJD, the dry variety nearly
reaches only 26% in April, when the severe type wet TS shoots up to nearly 20%. Such a
high incidence of the severe variety is not found in any other place. In DCA, the occurrence
of severe TSs is equally dominant over all three months. This is unique for DCA. However,
in C'BZR the occurrence of the severe variety is very poor all three months.
TAO

, In RMGT and CML, wet TSs dominate over the dry type in only one month Le. in
April. However in RMGT the incidence of the severe type is quite appreciable during both
April and May. ISD is the place where the severe type has a large occurrence only in March.
In DNP and KLN, there are periods when the % of occurrence of the severe type exceeds
10%.

5. CONCLUSIONS

In conclusion, it may be said that though Bangladesh is not a large country, the
space/time distributions of TSs vary greatly. Among the 19 stations studied, station SLT
has the most numerious premosoon TS occurrence: more than 50% the premonsoon season
are TS days. On the other hand, corresponding figure drops to 10% in the case of MJD. In
the majority of the stations, TS days contribute around 20% to 25%. In contrast in MJD,
though the occurrence of TS is comparatively rare, whenever it does occur, the possibility of
huge precipitation is very strong.
606 TAO, Vo1.6, No.4, December 1995

In central Bangladesh, the most likly time of TS occurrence is early afternoon or evening
while in the north it is either night or early morning. In southern areas, there is a wide
variation as to the intense period of TS activity.

Acknowledgements One of the authors (Md. Abdul Mannan Chowdhury) wishes to

express thanks to the Indian Council for Cultural Relationship (I.C.C.R) for offering him a
fellowship to pursue this work. The other author (U.K.D) thanks the India Meteorological
Department for giving him a research project on thunderstorm studies, and this work forms
a part of that. Both the authors wish to express their appreciation to Mr. S.N. Sengupta,
Director(Rtd.) of the India Meteorological Department for many helpful discussions.

REFERENCES

Ahmed, R, and S. Karmakar, 1993: Arrival and withdrawal dates of summer monsoon in
Bangladesh. Int. J. Climat., 13, 727-740.
Chowdhury, M. H. K. (Ed.), 1986: Proc. of the seminar on severe local storms, Bangladesh
Meteorological Department, Govt. of the People's Republic of Bangladesh, Dhaka.
Court, A, and J. F. Griffiths, 1982 : Thunderstorms climatology, In: Kessler, E. (Ed.),
Thunderstorm Morphology and Dynamics, 2, U.S. Department of Commerce, NOAA,
U.S.A
De, A. C., and S. N. Sen,1961: A radar study of premonsoon thunderstorms (nor'westers)
over Gangetic West Bengal. Ind. J. Met. Geophys., 12, 51-60.
Fischer, F. E., 1973 : Fundamental Statistical Concepts, Harper and Row, San Francisco,
Publishers Inc. 76-82.
Koteswaram, P, and V. Srinivasan, 1958: Thunderstorms over Gangetic West Bengal in the
premonsoon season and the synoptic factors favourable for their formation. Ind. J. Met.
Geophys.,9301-312.
Mukherjee, A. K., and P. B. Bhattacharrya, 1972: An early morning tornado at Diamond
Harbour on 21 March, 1969. Ind. J. Met. Geophys., 23, 223-226.
Mukherjee, A K., S. Kumar, and G. A Krishnamurthy, 1977: A radar study of growth and
TAO

decay of thunderstorms around Bombay during the premonsoon season. Ind. J. Met.
Geophys., 28, 475-478.
Mobassher, A, and M. H. Rashid, 1991: Some climatological characteristics of mean sea
level pressure in Bangladesh. Mausam, 42, 195-200.
Nandi, J. and A K. Mukherjee, 1966: A tornado over north-west Assam and adjoining West
Bengal on 19 April, 1963. Ind. J. Met. Geophys., 17, 421-426.
Rai, Sircar. N. c., 1953: On the use of tephigrams for forecasting Norwes'ters. Ind. J.
Met. Geophys., 4, 269-278.
Sen, S. N., and J. R Sen Gupta, 1961: A premonsoon thunderstorm with tornadic character-
istics at Calcutta. Ind. J. Met. Geophys., 12, 41-48.
Srinivasan, V., 1954: Thunderstorms in Kutch Saurastra. Ind. J. Met. Geophys., 5,
249-252.
World Meteorological Organisation (WMO), 1982: International Weather Code.