PAGEOPH, Vol. 121, No. 2, 1 9 8 3
0033-4553/83/020339-1451.50 + 0.20/0 © 1983 Birkh~iuser Vertag, Basel
Indian Summer Monsoon and E1 Nino D. A. MOOLEYl and B. PARTHASARATHY 1
A b s t r a c t - The associations between strong to moderate E1 Nino events and the all-lndia and subdivisional summer monsoon rainfall is examined for the period 1871 to 1978. The significance of the association is assessed by applying the Chi-square test to the contingency table. The analysis indicates that during 22 E1 Nino years the Indian monsoon rainfall was mostly below normal over most parts of the country. However, the association between E1 Nino and deficient rainfall or drought is statistically significant over the subdivisions west of longitude 80°E and north of 12°N. During the five strong E1Nino years-- 1877, 1899, 1911, 1918, and 1972-- many areas of India suffered large rainfall deficiencies and severe droughts. There are four moderate E1 Nino years-- 1887, 1914, 1953,and t976--when the sufferingwasmarginat. The relationship between E1 Nino and the Indian monsoon rainfall is expected to be useful in forecasting large-scale anomalies in the monsoon over India.
Key words: Indian summer monsoon; Droughts/floods; E1 Nino; Contingency techniques.
1. Introduction Abnormalities in the Indian summer monsoon, manifested as floods and droughts, have disastrous effects on agriculture, industry, and generation of hydroelectric power, causing severe strain to the national economy. The summer monsoon gives 75 percent to 90 percent of the annual rainfall for most parts of the country, between June and September. The increase in population has resulted in an increasing need for higher food production and hydroelectric power, making the link between the Indian economy and the summer monsoon critical (MooL~v et al., 1981; 1982a,b). Awareness of the need to understand and predict the monsoon over India have recently generated much interest in the possible relationships between the amount and distribution of Indian monsoon rainfall and antecedent regional and global circulation features. The most important need in monsoon forecasting is to pick out, with a reasonable degree of success, the years of low rainfall and, if possible, the drought years. Atmospheric circulation and Sea Surface Temperature ( S S T ) in the equatorial Pacific region exhibit remarkably coherent and persistent patterns. Several other atmospheric and oceanic features vary in association i Indian Institute of Tropical Meteorology, Pune-411 005, India.
340
D.A. Mooleyand B, Parthasarathy
with those in the equatorial Pacific (BJERKNES, 1969; EGGER el al., 1981; RASMUSSON and CARPENTER, 1982; NEWELLet aL, 1982). Anomalies in the monsoon rainfall over India are associated with positive SST anomalies over the eastern and central equatorial Pacific during the summer and a u t u m n months (KHANDEKAR,1979; ANGELL,1981; MOOLEY and PARTHASARATHY, 1983b). In this paper we propose to examine statistically the relationship between E1 Nino on one hand and all-India monsoon rainfall, as well as subdivisional rainfall in the monsoon season, on the other.
2. Details of the data Indian summer monsoon rainfafl The validity of any statistical analysis depends primarily on the quality of the data used. The rain gauge network over India varied by two orders of magnitude, from about 50 in 1850 to 5,000 in 1970. This variability in the network vitiates the areal average rainfall for different meteorologic subdivisions of the country. In view of this, 306 stations, one from each of the districts in the plain regions of India and distributed fairly uniformly over the country, were selected to form the network of rain gauge stations. These stations have rainfall data from 1871 on. The relevant rainfall data for the selected raingauge stations were collected from the records of the Deputy Director General of Meteorology (Climatology and Geophysics), Poona. The areal representativeness of a rain gauge in a hilly area is small and the rain gauge network in most hilly areas of the country is inadequate; therefore the hilly areas of the country, shown as hatched areas in Figure 1, have not been considered. The area considered measures 2.88 million km 2, which is about 90 percent of the total area of the country. Hereafter the area considered will be referred to as India or the country. Seasonal (summer monsoon season, June to September) area-weighted rainfall series were prepared for India (to be referred to as aft-India series) and for 29 meteorological subdivisions (as shown in Fig. 1) of the country by assigning a weight to each rain gauge station. Statistical tests indicate that the all-India and the 29 subdivisional seasonal rainfall series considered in this study for the period 1871 to 1978 are homogeneous and Gaussian distributed. The tests also indicate that there is no significant persistence in the series.
El Nino years E1 Nino may be defined as an anomalous oceanic and meteorological event characterized by the sudden appearance of abnormally warm surface water on a scale of a thousand kilometers offthe coast of Peru and Eucador (South American coast between the equator and 12°S). E1 Nino generally sets in around March or April and may last a year or more, attaining maxi-
Indian Summer Monsoon and E1 Nino
6¢E
72"
80"
76"
84"
341
92"
88"
96"E
N
/ 0
24
,..~
20
•
.i h, ~ ?,...,,o., : o,-, /
l-,; ~,. "'"'
.... " z6 '.. ",;;,.w ': z~
/ . ~ (-(
>
L "L
o~. - . J ~.o~H~,,
"~"
'~' x '/
~-" ,,/,J./
7~
t
"-'~ " k
t !
\
!
"
°
12*
\ 72°E
2
ARUNAE;HAL
76 °
PRADESH
3 NORTH ASSAM 4
SOUTH
ASSAM
5 SUB-HIMALAYAN WEST BENGAL 6
13 HARAYANA
88 °
Z4 MADHYA MAHARASHTRA
14 PUNJAB
25 MARATHWADA
15HIMACHAL PRA{:)ESH
26 VtDAR8HA
I 6 J A M M U AND KASHMIR
Z7 COASTAL` ANDHRA PRAOESH
GANGETtC WEST BENGAL 18 EAST
RAJASTH AN
2B TELANGANA
t9 WEST MADHYA
BIHAR PLATEAU
9 BIHAR
PLAINS
PRADESH 20EAST
--8
92"E
RAJASTHAN
IO EAST UTTAR PRAOESH II
84 =
i 7 WEST
70RISSA 8
" 80 °
Z9 R A Y A L S E E M A SO TAMIL NAOU
MADHYA PRAOESH 31 COASTAL` KARNATAKA
21 GUJARAT
3 2 NORTH KARNATAKA
WEST UTTAR PRADESH
22 SAURASHTRA • KUTCH
] 3 SOUTH KARNATAKA
PLAINS
23 KONKAN
3,4 K E R A L A
t 2 WEST UTTAR PRADESH HILLS
Figure L Meteorological subdivisions (regions) of India. Area considered: Contiguous India, excluding hilly areas (hatched).
342
D.A. Mooley and B. Parthasarathy
mum sea surface temperatures late in December. It usually appears to weaken temporarily between June and September, as the positive sea surface temperature anomalies and the South Pacific trade winds decrease in strength. There are a number of important physical questions about E1 Nino that remain unanswered; these include such topics as the initiating mechanism, the source of the warm water, and the dynamics of the communication between the eastern and central equatorial Pacific Ocean circulations. The surface air pressure difference between Easter Island (27°10'S; 109°25'W) and Darwin (12°26'S; 130°52'E) was shown by QUXNN and BURT (1972) tO be statistically related to the E1 Nino phenomenon and was shown to be a good index for prediction of the occurrence of the phenomenon. QUINN et al. (1978), on a thorough examination of all the available sea surface data and historical records, prepared a chronology of E1 Nino events since 1726 and have classified them as strong (S), moderate (M), weak, and very weak. The strong and moderate E1 Nino events as given by QUINN for the period 1871 to 1978 have been used in this study. The classification of E1 Nino-type events by intensity is highly subjective, since no two cases are exactly alike; however, strong E1 Nino events are recognized as positive sea surface temperature anomalies along the Peruvian coast in excess of 3 °C, and moderate events as those in excess of 2°C (QOINN et al., 1978). The effects of a moderate E1 Nino on the Peruvian anchoveta fishery, are considerable but less serious than the strong category. A total of 22 strong and moderate events occurred during the 108-year period studied (see Table 1), an average of one event every 4.9 years. 3. Relationship between Indian summer moonsoon rainfall and Et Nino events
SmKA (1980) has shown a general association between E1Nino events and deficient rainfall. RAMAOE (1983) also found a strong tendency for a belownormal summer monsoon during El Nino years, but he noted exceptions during individual years. RASMUSSON and CARPENTER (1983) showed a strong tendency for below-normal summer monsoon during E1 Nino years. In view of this, we have made a detailed statistical examination of the association between monsoon rainfall in India and in its component subdivisions and E1 Nino events. All-India rainfall series
In order to identify the years of abnormal monsoon rainfall the following procedure was adopted. It was already shown that the all-India and 29 meteorological subdivisional rainfall series are Gaussian distributed, so criteria based on probability considerations will be most appropriate. The standard deviate (ti = ( R i - R)/o') value of _+ 1.28 and ___0.84 are, respec-
I n d i a n S u m m e r M o n s o o n a n d E1 N i n o
343
Table 1 Details of El Nino Years and All-lndia Summer Monsoon Rainfall. Normal Seasonal Rainfall ~= 85.31 cm. Standard Deviation (a) = 8,29. )
Series
Intensity of El Nino
Year
1
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
All-India seasonal rainfall
1871
M
t877 1880 1884 1887 1891 1896 1899 1902 1905 1911 1914 1918 1925 1929 1939 1941 1953 1957 1965 1972 1976
S M S M S M S M M S M S S M M S M S M S M
1
I L
Averages
Amount (cm)
% departure from normal
Standardized value
84.57 60.37 81.68 92,91 89.65 78.90 82,4 l 62,81 79,I 1 71.51 73.29 89,89 64,81 80,27 81.93 78,86 72.90 91,97 78,43 70.68 65,32 85,47
0.9 -- 29.2 4.3 + 8.9 + 5.1 7.5
-- 0.09 --3.01 - - 0.44 +0,92 +0,52 --0.77 -0.35 --2.71 --0,75 --1.66 -1,45 +0.55 -- 2.47 --0,61 --0,41 -0.78
78,08
-
3.4 26.4
7.3 -16.2 -14,1 + 5.4 --24.0 5.9
--
-
4.0
7.6 -14.6 + 7,8 8.1 -17.2 -23.4 + 0.2
-
-
8.8
-
1.50
+0.80 -0.83 1.76 --2.41 +0.02 --0.87
tively, the 10 percent and 20 percent points of the standard normal distribu-_ tion (one-tailed value), where R,. is the monsoon rainfall of the ith years, R is the mean (normal monsoon) rainfall, and cr is the standard deviation of the monsoon rainfall. These percentage points have been utilised to categorise the monsoon rainfall as drought condition, deficient, below normal, above normal, excess, or flood, as indicated below: (i) (ii) (iii) (iv) (v) (vi)
Drought: Deficient: Below normal: Above normal: Excess: Flood:
< < < > > >
-- 1.28 --0.84 but > - 1.28 0.0 but > - 0 . 8 4 0.0 but < 0.84 0.84 but < 1.28 1.28
A similar classification was used by MOOLEY and PARTHASARATHY (1983a) tO identify droughts and floods over different meteorological subdivisions of India. Figure 2 shows the all-India summer monsoon rainfall for the period of t871 - 1978, and the drought or flood years.
344
D.A. Mooley and B. Parthasarathy
z < :11 I 1
06 O
T-~o
I
e
o
.=-
o
C
0
o
!
-
o
C
C~ la
m
0
C 0
; Z 7 o
E E o
®
I
Indian Summer Monsoonand El Nino
345
Table 1 gives the details of strong and moderate E1 Nino years, along with all-India rainfall and the seasonal rainfall departures from the mean, expressed as percentage and standardized values. The average departure of the all-India rainfall during these 22 E1 Nino years works out to be - 8.8 percent a n d - 0.87 standard units. There are 17 years (i.e., 77% of the total number of El Nino years) of below-normal rainfall, out of which eight years show drought situations. Of the five E1 Nino years with above-normal monsoon rainfall, one year shows excess rainfall. Some of the strong El Nino y e a r s - namely 1877, 1899, 1918, and 1972 m s h o w high monsoon rainfall deficiencies (standardized variate less than -2.5). The E1 Nino event data and the rainfall series have been grouped to form contingency tables. The four groups of years are (i) E1 Nino and drought or deficient rainfall, (ii) E1 Nino and no drought or deficient rainfall, (iii) no E1 Nino and drought or deficient rain, and (iv) no E1 Nino and no drought or deficient rainfall. Taking the null hypothesis of independence, the contingency tables were tested for significant association by applying the Chi-square test. The results of the Chisquare test are given in Table 2. It is seen from this table that for the all-India series the value of the Chi-square test statistic is 4.97, which is significant at the 5% level for t df. The association between E1 Nino and drought or deficient rainfall is, therefore, significant. India as a whole (all-India) is perhaps too large an area to be treated as a single unit. Moreover, intercorrelation coefficients between all-India rainfall and all meterologic subdivision rainfall series are not highly significant. Figure 3 shows that for seven subdivision rainfall series the correlation coefficients (CC) are not significant at the 1 percent level; these subdivisions are mostly in the northeastern part of India. High CC values (0.6) are noticed over central and northwestern parts of India. In view of these points we have examined the association between E1 Nino events and the summer monsoon rainfall of the 29 subdivisions of India.
Rainfall seriesfor meteorologic subdivisions Figure 4 gives details of the meteorologic subdivisions considered in this s t u d y m their areas, expressed as percentages of the country's area, and rainfall statistics such as amount of summer monsoon rainfall, this rainfall expressed as percentage of annual amount, standard deviation and coefficient of variation--and the rainfall characteristics are depicted for each of the subdivisions during 22 strong or moderate E1 Nino years. The smallest subdivisions are sub-Himalayan West Bengal and coastal Karnataka, each of whose area is less than one percent of the country. The biggest subdivisions are West and East Madhya Pradesh, each of whose area is about 8 percent of the country. The amount of seasonal rainfall during the four months June to September is more than 70 percent of the annual rainfall in almost all subdivisions except some in southern peninsular India, the least percentage
03 04 05 06 07 08 09 t0 Il 13 14 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34
8 6 7 7 2 5 5 5 7 11 10 11 9 10 9 8 9 8 10 7 10 11 7 10 6 6 10 11 11 7
Deficient rainfall or drought conditions, as defined in text.
All-India North Assam South Assam Sub-Himalayan W.B. Gangetic W.B. Orissa Bihar Plateau Bihar Plains East U.P. West U.P.(Plains) Harayana Punjab West Rajasthan East Rajasthan West M.P. East M.P. Gujarat Saurashtra and Kutch Konkan and Goa Madhya Maharashtra Marathwada Vidarbha Coastal A.P. Telengana Rayalaseema Tamilnadu Coastal Karnataka North Karnataka South Kamataka Kerala
Subdivision
Drought, a El Nino
~ Significant at 5% level.
13 17 18 12 23 17 14 13 12 10 11 15 9 9 14 13 13 10 5 11 I1 13 18 t6 15 14 9 12 17 15
Drought, No El Nino
13.55' 8.35' 2.24
74 69 71
11 11 15 * Significant at 1% level.
4.97 b 0.59 1.17 3.86 b 3.06 0.09 0.50 0.73 3.86 b 16.45¢ 11.92" 10.14" 11.67c 14.76c 6.32 b 5.04 b 7.19 ~ 7.7@: 22.97 c 4.56 b 11.92c 12.32~ 1.17 6.90~ 1.08 1.73 14.80"
Calculated Chi-square value
73 69 68 74 63 69 72 73 74 76 75 71 77 77 72 73 73 76 81 75 75 73 68 70 71 72 77
No Drought, No El Nino
14 16 15 15 20 17 17 17 15 I1 12 11 13 12 t3 14 13 14 t2 15 12 11 15 12 16 16 12
No Drought, El Nino
Frequency of occurrence of
Results of Contingency Analysis of the Occurrence of Drought or Deficient Rainfall Conditions in the Monsoon Season During 22 Strong or Moderate El Nino Events, 1871-1978
Table 2
t:r
t~
o_
.o .> K O
4~ O~
Indian Summer Monsoon and El Nino
~gOp
68* t
72 ° ' ~
76 °
80 ° ~_~
84* '
347
88 ° '
9 2 ° ..... '
96* '
136 °
52* --
TIBET
--..-,-,-.';/-,'""°""Y
J
-4z4 °
-'0.74: :" : , ~ : ...... : ~ . . t t : : . ' . : t : r : : 0.58 BURMA
20°1-
,g;P.::::..:'.- E.::.b'.ee ::£:::'./:::::o:3z :::i' ,aRABIAN
16° -
SEA
:~'.v:.'.'.:" . . ~ - " . ' V E ~ . r i : . j / " "ii"£%~';':':~ x
~ B A Y V
~::~L".':.'.'""
0
-~ 20°
160
0"58
12" O,24 ":'."
72 °
76 ° CC
sm
80 ° SIGNtFICIENT
84 ° AT
88 ° I%
92"E
LEVEL
Figure 3. Correlation coefficients between all-india rainfall and the different meteorologic
subdivision rainfall series for the summer monsoon season (June to September)for the period 1871- t978.
being for Tamilnadu. It is seen from Figure 4 that high rainfall is received over the west coast and northeastern parts o f the country, and low rainfall over northwest India and Tamilnadu. Central parts of India receive moderate rainfall. Low coefficients o f variation (CV) are observed over high rainfall areas, and high CV over low rainfall areas o f the country. We have adopted the same procedure discussed in the earlier section to examine the association between E1 Nino events and the seasonal rainfall o f different subdivisions. It is generally seen from Figure 4 that during the 22 E1 Nino years rainfall was below normal in about 68 percent o f the years, and in
2'~10
,C~nl~loti¢Wetl 8enQQ
0,1,,o
lllih~l~PtoI~'¢I~
oi~
06
o~
08
os
Kw~. YypeM Irl N I 4 e ( l e $ t r q ,
3-z4
iI)~
5ol
~
~z
600
k
Roildlll
O~u~! ~
kulll > - i 28
~
*
e
eX
X @X
X
x
x !O
x
x ~)
~ 0 ie X * I(~
X 8 iO
xiO
mJzo :' zo I H
xx
7
0
3819 9
5
a
5
6 zo
io;8
z
,~2o'~o
~,
,
,7!
9 ] 1 9 1 ~ 3 { 2 ~ I~
;,~,,x,.~ele,o
!* O O O . • x 0 x ee[x o b e e . i e e . i . O .
o 0 • oeo,, xO e!eeoe,
x i0
• x oo
*O'OOxlE~- 1 8 0 @ ~O Oex!o x @x xe o O @ O O e , o @ , xO, o o o o e x x x * * x O" ,,0, e , , e 6 e e e e . . , o l ~ • • o o . e . o e e * * xiOx x .O..ex x;!.O].
• Ox Ox • 0. 00*
•e,. 1 9 2 7 21 16 141~4 ~0 24 2~ 24 23 9 ~4 19 ~I ~ 2 2 5
•
X X
~i X
Oi e X O! X X
,] • e x M $ M S
e]
zo
*
X
X
0
1= '
~
O ~) • X e X O 0 0 ~ e@ X ~ O X x o 0 0 X jO x;*
O
O
_
|E.#
oeoxloo o.eox x @ *@~ * *e O e!xo*!oo!'o!o°!o * 4) • x O x l O O . O O , O e , "00. @ e,oeioo;xoox!xexxoe. • ~1~1@ • @ ! o e i x .I. * o @ i . c e e i o e i e X
x iO!. * ~)i x
X C
• C @ C
X (~ •
~
•
•
~
20
18
2;3
28'
23
Z2
20
27
21
20
~4
.,
16
l?
~9
~0
23
~2
19
14
,4 0
IS
15
12
II O
"-_*_~:_
o o e o , • *'oix x x e o x i x o e I !~ 0 x OQ • ex Oe[O. x x Oei" 0 x ex 0 .X 0 0 0 1 0 4 • Oil" x, x x e" 0 x X ~I X X x x~ix x x [O OO x x xO 0 @ @ iX OC x x @x D~" x! xOx O X X e i* x e x x xeex x x e O o e e e ~ 0 'x 0 OOO*G x@ x x * x ~x xO O [O • x O* x x x xO. x . e O x O e x @ eooxloeoo.oxoxoie,. 0 ~
~ I
....
i .=!
_=:
i~l ( - 0 64 Oehc,~n,
~7z .........
1oz
119
SI4
S9
It9
162
t09
I77
187
122
471
,oe
z.z
1811
171
I00
137
177
zoo
i~,~
156
,~o
168
ZIPS
174
It
E i°
-- - i~-
=. E
(%)
,,
,~
,,
,~
t7
,,
I~ 1,
6 ,6 16 ,, 1= J4
,.
13 1, ,, 14 ,z 16 13 i~
~
,,
~
,
Ii
,o
I '
~ •
,
5
~
4
1o ~
~ 6 ~ ~ ,o e ~
a 10 6 9 ~ ,
,
, ,
,, 9 1o ,o , 9 ,
~ ,o
s 6
7 , ~
-,'
"'"°"'*
'
,
3
,
z
z z
,
i
z 4
6 6
,
4 s
,.,
El Nino years and summer monsoon (June to September) rainfall position in different meteorologic subdivisions oflndia.
II~ I l l - I ) i v l l i l l
M llllb*011rl~III~FJ0ed 48HI I r l c e l l Roinfcdl
Figure 4.
i
.'_:_'-:; Ol 111"1~11 lliUllI)I'11411I14 D l f l l i l e l I t l i l l h l l l
s4
117 72
Zl4l
33
59
80
52
2"77
:~011
422
714
0"64
4-S2
:'.40
394
504
~
~'39 2"25
64
6119
2'24 87
79
94
,, ~3
88
9;~
88
8,
87
~
e~
11(2
~,
¥5
SO
66
5~0
:'371
,4
z..
400
IZJ~,
646
25?
45,
774
'7 ?9
S'lt
i,ss IlJoMed~ele)
~
1~,
~,~'~.~,.
Ceestld K w. ./.a. l. a k a MMIk Kl~lloktl
3t
Vld4~bhe
T,=-"T"=~IU
MefQthwelle
116
Royekleeema
Me4klm Mlthe~llekt~n
19
Ko~k~m & Ooo
25
24
!0
=.~=.. s~,,,,~,t,,, .. ©
=1 zz
Telengen4
[081 M(~dhya P r ~ i h
~0
Ceetl4# Andh~e P,ed~',llh
1"18
Elml R = j a s t h g n
tl
26
ssz
Wt,! Ruiaslha.
17
27
,'59
Horoyon4
,5 6-7?
5.~
Will Ut)O, P,mdltuh(Ploin,)
go?
~-oe
II
~o3.3
~.z7
Plain
tt07
,,,z
1146
IIHIS
~ :o. u.., p,.~,,~
2"7"I'
~41
0"~5
O~ [ Sub HimoJoya. W,~! ~ I
1411
65
166
1454
I'H 4"211
NOrth A,l~m
~
~,,,,.
.,,.
p.,. s---.. ~ r" ,,......, ..~,, mm ~
0 4 ; $ovlh Alllom
03
=......,-,..,-,--
s,,~-
::r
::r
e~
O O
>
9
Indian Summer Monsoon and E1Nino
349
56 percent of the years deficient rainfall or drought conditions were reported. For eight subdivisions the monsoon rainfall was below normal in more than 80 percent of the E1 Nino years, the highest percentage being for South Karnataka. Eleven subdivisions reported 50 percent or more years of E1 Nino with deficient or drought conditions. In 18 out of 22 E1 Nino years, below normal rainfall was observed in about 70 percent of the subdivisions. The five years 1877, t899, 1911, 19t8, and 1972 stand out in excellent agreement with bad monsoon years and strong E1 Nino events. In view of this we have further examined the statistical significance of the associations of E1 Nino with drought or deficient rainfall. The results are given in Table 2. Table 2 shows that the associations are significant at the 5 percent level for 19 contiguous subdivisions, and at the 1 percent level for 14 subdivisions. The association is thus highly significant for about half of the subdivisions. Figure 5 shows the subdivisions for which the association between E1 Nino and drought or deficient rainfall is significant at 5 percent and 1 percent levels. It is observed from Figure 5 that an excellent relationship (i.e., significance at 1% level) exists for contiguous parts of Northwest India, central parts of peninsular India, and northern parts of the west coast. There is no significant association for the portion east of 80°E or for Kerala or Tamitnadu. From this it can be seen that the drought or deficient rainfall over three-fourths of India is significantly associated with the E1 Nino phenomenon. Correct prediction of E1 Nino will be helpful in the foreshadowing of drought or deficient rainfall conditions over India. We feel that the variability of the monsoon circulation and the associated large-scale anomalies in the performance of the Indian monsoon over the India subcontinent have to be regarded as an important manifestation of a planetary-scale phenomenon. The E1 Nino event, Southern Oscillation (SO), and Sea Surface Temperatures (SST) in the equatorial Central Pacific Ocean are interrelated to the overall evolution of the circulation features of the tropical region. Strong coupling between these features is well reflected in the behaviour of the Indian monsoon rainfall; i.e., strong E1 Nino, weak SO resulted in low monsoon rainfall during the years 1877, 1899, 1911, 1918, and 1972. However, during the moderate E1Nino years of 1887, 1914, 1953, and 1976, the coupling between Et Nin.o, SST, and SO has apparently been weak, or perhaps there was a general breakdown of the association and the failure of the Indian summer monsoon was not of a large-scale character in these years. Therefore it is essential to understand the possible feedback between these phenomena as an important mechanism for cause and effect relationships between SST, SO, El Nino, and the Indian summer monsoon rainfall. It is also necessary to examine whether any factor or factors has played an important role in influencing the monsoon performance in years of poor association with E1 Nino. Studies by QUINN and BURT (1972), QUINN (1974), WYRTKI el al. (1976), QUINN et al. (1978), DONGUY and HENIN (1980), RAMAGE and HORI
350
D. A. Mooley and B. Parthasarathy
&
I
L
1
3z*-
I
t
I
1
~.
C9 z¢
z~ L
! P
BURMA
,1
zdBAY OF BENGAL
ARABIAN SEA 12 a
L _ _ _ _ _ _ _ _ _ t _ ~ _ _ _ _ _ _ ~~ AN'K'~ 72* 76* 80* SIGNIFICANT
AT
8 4 ('
1% L E V E L
~. . . . 88*
SIGNIFICANT
L.............. 92"E
A T ,.5% L E V E L
Figure 5. Association between El Nino and drought or deficient rainfall over 29 meteorologic subdivisions of India
(1981), and RASMUSSONand CARPENTER (1982) show that it is possible to predict or observe El Nino phenomena in advance. In view of this, Indian summer monsoon rainfall anomalies in different parts of the country can be foreshadowed, with due considerations to all these features.
Conclusions The analysis of moderate to strong E1 Nino occurrences and the characteristics of summer monsoon rainfall over India and over the different meteorologic subdivisions of the country during the period 1871 to 1978 brings
Indian Summer Monsoon and El Nino
35 1
out the following features: (i) During the 22 E1 Nino years the monsoon rainfall over most parts of the country was below normal. (ii) Association between El Nino events and deficient rainfall or drought conditions over India is statistically significant. (iii) The worst years for the Indian summer monsoon are generally Et Nino years. (iv) Deficient rainfall or drought conditions over 19 meteorologic subdivisions show significant association with the E1 Nino phenomenon. The association is significant at the 1 percent level for Northwestern India, the northern west coast, and the central peninsula. E1 Nino, the Southern Oscillation, and equatorial Pacific Sea Surface Temperature are coupled. Studies of many research workers show that it is possible to predict E1 Nino several months in advance. This would provide an additional factor for foreshadowing the general behaviour of the Indian summer monsoon.
Acknowledgment The authors are grateful to Dr. Bh. V. Ramana Murty, Director, Indian Institute of Tropical Meteorology, Pune, for the facilities to pursue this work, and to the Deputy Director General of Meteorology (Climatology and Geophysics), Pune, for making available the necessary rainfall data. They are also thankful to Mrs. N. A. Sontakke and Mr. D. R. Kothawale for assistance in computation, and to Mrs. S. P. Lakade for typing the manuscript of the paper. REFERENCES ANGELL, J. K. (198 t), Comparison of variations in atmospheric quantities with sea surface temperature variations in the equatorial eastern Pacific, Monthly Weather Rev. 109. 230-243. BJERKNES, J. (1969), Atmospheric teleconnections from the equatorial Pacific, Monthly Weather Rev. 97, 163- 172. DONGUY, J. R., and HENIN, C. (1980), Climatic teleconnections in the western South Pacific with El Nino phenomenon, J. Phys Ocean. 10, 1952-1958. EGGER, J., MEYERS, G., and WRIGHT,P. B. (1981), Pressure, wind and cloudiness in the tropical Pacific related to the Southern Oscillation, Monthly Weather Rev. 109, 1139 - t 149. KHANDEKAR,M. L. (t 979), Climatic teleconnections from the equatorial Pacific to the lndian monsoon, analysis and implications, Arch. Meteorol, Geophys. Biokl. A 28, 159-168. MOOLEY, D. A., and PARTHASARATHY,B. (1982), Fluctuations in deficiency of the summer monsoon over India, and their effect on economy, Arch. Meteorol. Geophys. Biokl. B 30, 383-398. MOOLEY,D. A., and PARTHASARATHY,B, Droughts andfloods over lndia in summer monsoon seasons: 1871-1980. In: A. S. Perrot, M. Beran, and R. Ratchiffe (eds.), Variations of the Global Water Budget. (Dordecht, Holland: Reidel, 1983a), pp. 239-252.
352
D.A. Mooley and B. Parthasarathy
MOOLEY,D. A., and PARTHASARATHY,B. (1983b), Variability of the Indian summer monsoon and tropical circulationfeatures, Monthly Weather Rev. 111, 967 -978. MOOLEY, D. A., PARTHASARATHY,B., and SONTAKKE,N. A. (1982), An index of summer monsoon rainfall excess over India and its variability: t871-1978, Arch. Meteorol. Geophys. Biokt. B 31, 301-311. MOOLEY, D. A., PARTHASARATHY,B., SONTAKKE,N. A., and MUNOT, A. A. (1981), Annual rainwater over India, its variability and impact on the economy, J. Clirnatol. 1, t 6 7 - 186. NEWELL,R. E., SELKmK,R., and EBISUZAKt,W. (1982), The Southern Oscillation:Sea surface temperature and wind relationships in a lO0-year data set, J. Climatol. 2, 357-374. QUINN, W. H. (1974), Monitoring and predicting El Nino invasions, J. Appl. Meteorol. 13. 825-830. QUINN, W. H., and BUR'r,W. V. (1972), Use of the Southern Oscillation in weatherprediction, J. Appl. Meteorol. 1t, 616-628. QUINN, W. H., ZOPF, D. O., SHORT,K. S., and Kuo YANG, R. T. W. (1978), Historical trends and statistics of the Southern Oscillation, El Nino. and Indonesian droughts, Fish. Bull. 76, 663-678. RAMAGE,C. S. (1983), Teteconnections and the siege of time, J. Climatot. 3, 223-231. RAMAGE,C. S., and HORI, A. M. (1981), Meteorological aspects of El Nino, Monthly Weather Rev. 109, 1827-1835. RASMUSSON,E. M., and CARPENTER,T. H. (1982), Variations in tropicalsea surface temperature and surface wind field associated with the Southern Oscillation~El Nino, Monthly Weather Rev. 110, 354-384. RASMUSSON,E. M., and CARPENTER,T. H. (1983), The relationship between eastern equatorial Pacific Sea Surface Temperature and rainfall over India and Sri Lanka, Monthly Weather Rev. 111. 517-528. SIKKA,D. R. (1980), Some aspects of the large-scalefluctuation of summer monsoon rainfall over India in relation tofluctuations in the planetary and regional scale circulation parameters, Proc. Ind. Acad. Sci. (Earth & Planet. Sci.) 89, 179-195. WYRTKI, K., SrRouv, E., PATZERT,W., WILLIAMS,R., and QUINN, W. H. (1976), Predicting and observing El Nino, Science 191, 343-346. (Received 10th May 1983, revised 24th August 1983, accepted 15th October 1983)
