Dominant Synoptic Disturbance in the Extreme Rainfall at Cherrapunji,Northeast India, Based on 104Years of Rainfall Data (1902–2005)

FUMIE MURATA,a TORU TERAO,b HATSUKI FUJINAMI,c TAIICHI HAYASHI,d

HARUHISA ASADA,e JUN MATSUMOTO,f AND HIAMBOK J. SYIEMLIEHg

a Faculty of Science and Technology, Kochi University, Kochi, JapanbFaculty of Education, Kagawa University, Takamatsu, Japan

c Institute for Space-Earth Environmental Research, Nagoya University, Nagoya, JapandCenter for Southeast Asian Studies, Kyoto University, Kyoto, Japan

eNara Women’s University, Nara, JapanfTokyo Metropolitan University, Hachioji, and Japan Agency for Marine-Earth Science and

Technology, Yokosuka, JapangDepartment of Geography, North-Eastern Hill University, Shillong, India

(Manuscript received 5 June 2016, in final form 18 July 2017)

ABSTRACT

The characteristics of active rainfall spells (ARSs) at Cherrapunji, northeast India, where extreme high

rainfall is experienced, and their relationships with large-scale dynamics were studied using daily rainfall data

from 1902 to 2005 and Japanese 55-Year Reanalysis from 1958 to 2005. Extreme high daily rainfalls occur in

association with ARSs. The extremely large amounts of rainfall in the monsoon season are determined by the

cumulative rainfall during ARSs. ARSs start when anomalous anticyclonic circulation (AAC) at 850 hPa

propagates westward from the South China Sea and western North Pacific, and covers the northern Bay of

Bengal. The AAC propagates farther westward and suppresses convection over central India during ARSs at

Cherrapunji, and continues for 3 to 14 days. Consequently, a northward shift of the monsoon trough during

the ‘‘break’’ in the Indian core region occurs. The westerly wind, which prevails in the northern portion of the

AAC, transports moisture toward northeast India and enhances moisture convergence over northeast India

with southerly moisture transport from the Bay of Bengal, and greatly intensifies the orographic rainfall. In

the upper troposphere, the Tibetan high tends to extend southward with the onset of ARSs. A linear re-

lationship can be seen between the length and total rainfall of an ARS. Longer ARSs tend to result in greater

total rainfall. AACs with a greater zonal scale tend to produce longer and more intense ARSs. This study

provides evidence for the effect of western North Pacific AACs on the Indian summer monsoon.

1. Introduction

The Indian summer monsoon has remarkable intra-

seasonal variation. Rainfall periods with above and be-

low normal rainfall amounts have been called ‘‘active’’

and ‘‘break’’ spells, respectively. A detailed review of

studies of active and break spells was conducted by

Rajeevan et al. (2010). Traditionally, the active and

break spells of the Indian summer monsoon have been

based on the convective activity over northwest and

central India [referred to as the ‘‘Indian core region’’

following Rajeevan et al. (2010)]. Negative (positive)

rainfall anomalies over northeast India corresponded to

active (break) spells over northwest and central India.

The synoptic condition during break spells has been

described as a northward shift of the trough of low

pressure toward the foothills of the Himalayas.

Ohsawa et al. (2000) studied the intraseasonal rainfall

variation during the 1995 summer monsoon season, and

found that rainfall in Bangladesh increased when the

monsoon troughwas located at the foot of theHimalayas.

Shrestha et al. (2012) analyzed rainfall over the central

Himalayan region using 11 years’ worth of data from

the Tropical Rainfall Measuring Mission (TRMM)

Precipitation Radar, and showed that rainfall was heavy

over the southernmost hills of the Himalayas [called the

Siwalik Range or sub-Himalayas; approximately 500–

700m above mean sea level (MSL)] when the monsoon

trough was lying in the foothills of the Himalayas.

Goswami et al. (2010) analyzed the extreme rainfall overCorresponding author: Fumie Murata, fumie@kochi-u.ac.jp

15 OCTOBER 2017 MURATA ET AL . 8237

DOI: 10.1175/JCLI-D-16-0435.1

� 2017 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS CopyrightPolicy (www.ametsoc.org/PUBSReuseLicenses).

northeast India, and showed that these extreme events

occurred in association with monsoon synoptic events

rather than isolated thunderstorms. In general, studies

that have focused on intraseasonal variation over

northeast India have been limited.

Hartmann and Michelsen (1989) performed spectral

analysis of rain gauge data for 70 years from 3700 sta-

tions in India, and noted that several stations along the

southern flank of the Himalayas had spectral peaks of

nearly 15 days. The 40–50-day spectral peak was not

significant along the southern flank of the Himalayas,

whereas it was dominant at most stations elsewhere in

India. Ohsawa et al. (2000) detected a periodicity of

around 20 days in an analysis of the 1995 summer

monsoon season. Fujinami et al. (2011) analyzed rain

gauge data for 20 years in Bangladesh, and showed that a

7–25-day spectral peak was dominant, while a 30–60-day

spectral peak was not prominent. Fujinami et al. (2014)

further investigated the spatiotemporal structure of a

7–25-day mode over the Meghalaya–Bangladesh–western

Myanmar region.

A 10–20-day mode has been detected in various ele-

ments of the Indian summer monsoon system (e.g.,

Krishnamurti and Bhalme 1976). Past studies suggest

that the 10–20-day mode in the Asian summer monsoon

region originates over the equatorial west Pacific

(Annamalai and Slingo 2001), propagates westward

(Krishnamurti and Ardanuy 1980) with a double-cell

structure (Chen and Chen 1993), and is a manifestation

of the moist equatorial Rossby waves modified by the

climatological mean flow (Kikuchi and Wang 2009).

Fujinami et al. (2014) revealed that a westward-

propagating anticyclonic anomaly related to the equa-

torial Rossby wave from the western North Pacific

promotes rainfall over the Meghalaya–Bangladesh–

western Myanmar region.

This study analyzes active rainfall spells (ARSs) in the

monsoon season at Cherrapunji, which is a town located

in northeast India. Cherrapunji is a unique station be-

cause it holds the world record for the highest rainfall

in a calendar month and for longer time scales than a

calendar month (Jennings 1950; Kiguchi and Oki 2010),

as well as the highest daily rainfall records in India

(Guhathakurta 2007). Rainfall at Cherrapunji has clear

ARSs, and daily rainfalls with more than several hun-

dred millimeters are common during ARSs (Fig. 3). The

large daily rainfall at Cherrapunji during the ARSs is

suitable to detect the synoptic-scale events with high

sensitivity because it can minimize other noises caused

by few local cumulonimbus clouds.

Cherrapunji is located on the southern slope of the

Meghalaya Plateau (Fig. 1a) with a maximum elevation

of around 2000m above sea level. The plateau is divided

from the Himalayan ranges by the Brahmaputra Valley.

As the plateau is the first elevated terrain that warm,

moist air from the Bay of Bengal encounters, maximum

precipitation tends to occur (Houze 2012). It is con-

nected to the Indo-Burma Range in the east. The

southern margin of the Meghalaya Plateau is steep and

rugged due to the tectonic motion along the Dauki fault

at the southern boundary of the plateau. There are

several deep valleys that open toward the southwith steep

cliffs on both sides (Fig. 1b): the town of Cherrapunji is

situated near one such valley. Orographic amplification

is the major plausible cause for the extreme precipitation

at this location (e.g., Romatschke et al. 2010), although

other factors, such as a local front between the south-

west monsoon flow and east–northeast mountain winds

in the Brahmaputra Valley, may also have an influence

(Yoshino 1975).

Northeast India receives the heaviest rainfall in India.

As the region includes hilly areas and adjoins the great

Himalayas, orographic effects are important in rainfall

in northeast India (Goswami et al. 2010). Our study re-

veals the relationship between the local orographic rain-

fall and large-scale circulation patterns at intraseasonal

time scales. Understanding the rainfall patterns over

northeast India might be crucial in understanding the

hydrological processes over the Ganges–Brahmaputra–

Meghna River basin (Hofer and Messerli 2006). For ex-

ample, heavy rainfall duringARSs atCherrapunji directly

runs off toward Bangladesh and triggers flash floods.

Also, the high monsoon rainfall years at Cherrapunji

tended to correspond with severe flood years in Bangla-

desh (Murata et al. 2007).

Causes of the negative correlation between rainfall

anomalies in the Indian core region and northeast India

in the active–break cycle have not been discussed. The

ARSs at Cherrapunji can be representative in northeast

India because the extreme rainfall amount provides

noiseless signal for favorable synoptic-scale events. The

objective of this study is to answer the following questions

by understanding the characteristics of the Cherrapunji

ARSs (CARSs), and discuss the relationship with active–

break cycle in the Indian core region:

1) What is the relationship between CARSs and large-

scale dynamics?

2) What is the contribution of CARSs to the extreme

rainfall?

Section 2 describes the datasets used in this study.

The definition and characteristics of CARSs are de-

scribed in section 3. Section 4 gives the spatiotemporal

structure of CARSs. The synoptic condition during

CARSs, the relationship with the Indian core region, and

the contribution to the extreme rainfall are discussed

8238 JOURNAL OF CL IMATE VOLUME 30

in section 5. Finally, the results are summarized in

section 6.

2. Data

Data on the daily rainfall observed at Cherrapunji

observatory of the India Meteorological Department

(IMD) were obtained from the following different

sources: the Research Data Archive at the National

Center for Atmospheric Research (Computational and

Information Systems Laboratory; ds.480.0) for data

from 1902–70, and the Guwahati regional office of the

IMD for daily rainfall data from 1970 to 2003. Daily

rainfall data for 2003–05 were obtained directly from the

Cherrapunji Observatory of the IMD. Five years of daily

rainfall data at Cherrapunji are missing completely

(1965, 1982–84 and 1987), and 7 years (1946, 1960, 1961,

1963, 1964, 1972, and 1986) havemore than 10%missing

data for June–September. To study the relationship

between rainfall at Cherrapunji and that at other

regions, the IMD rainfall dataset (Pai et al. 2014) on a

0.258 latitude–longitude grid was used for the period

FIG. 1. (a) Topographical map of the region around Cherrapunji (shown with the star).

(b) Landscape of the southern cliff of the Meghalaya Plateau nearby Cherrapunji town.

15 OCTOBER 2017 MURATA ET AL . 8239

1902–2005, and the APHRODITE V1101 rainfall data-

set (Yatagai et al. 2009, 2012) on a 0.258 latitude–

longitude grid was used for the period 1958–2005. The

streamfunction, horizontal wind components and verti-

cally integrated water vapor flux components of the

Japanese 55-Year Reanalysis (JRA-55) (Ebita et al.

2011; Kobayashi et al. 2015) on a 1.258 latitude–longitudegrid were used to analyze the features of circulation

patterns and moisture transport during CARSs. Daily

interpolated outgoing longwave radiation (OLR) data

on a 2.58 latitude–longitude grid were used to show large-

scale convective activity, including ocean areas from

1979 to 2005 (Liebmann and Smith 1996). In the com-

posite analysis, the statistical significance of the differ-

ences from climatological values at each grid point was

estimated using the Student’s t test.

3. Features of ARSs at Cherrapunji

In this section CARSs are defined and described.

Figure 2 shows the 92-summer ensemble spectrum of the

Cherrapunji daily rainfall. The years for which more

than 10% of the June–September daily data are missing

were excluded; in total, 92 years were used, the first

three harmonics of the annual cycle were removed, and

the fast Fourier transform technique was used to cal-

culate the spectra. A 5-point running mean in the fre-

quency domain was applied to the raw spectra to reduce

estimate errors. Pronounced peaks appear around a

10–30-day period and a peak of about 11 days exceeds

the 95% confidence level. The spectrum is similar to that

of all Bangladesh daily rainfall (Fujinami et al. 2011)

and represents the dominance of the 10–20-day mode.

Fujinami et al. (2011, 2014) examined the characteristics

of the 10–20-day mode over the Meghalaya–Bangladesh–

westernMyanmar region. This study focused on the timing

of onset, length, and total rainfall amount of CARSs be-

cause they are the important hydrological characteris-

tics of this region.

Figure 3 shows examples of the daily rainfall from

21 May to 10 October in the years 1915 and 1974. The

annual rainfall in 1915 was almost mean value and in-

cluded four active spells. The annual rainfall in 1974 was

the highest between 1902 and 2005. The solid curve il-

lustrates the daily climatological rainfall: it shows the

average daily rainfall in each year, and smoothing by the

15-day running mean twice. The daily climatological

rainfall reaches a maximum in late June, which was

more than 1 month earlier than the rainfall over the

Indian core region (Rajeevan et al. 2010; Pai et al. 2016).

Heavy rainfall days (HRDs) are days with more than

1.5 times the daily climatological rainfall between June

and September. CARSs are defined as consecutive

HRDs on which the total rainfall exceeds 500mm. For

HRD periods that include 1 June or 30 September, the

number of consecutive HRDs extended into May or

October. There were only three (two) cases of 1-day

(2-day) intervals between the two adjacent CARSs.

Definitions of active–break spells can vary (Rajeevan

et al. 2010). Fujinami et al. (2014) defined active and

break monsoon spells according to filtered data. In this

study unfiltered data are used to define ARSs, with no

assumption regarding frequency. The shaded periods in

Fig. 3 represent ARSs in this study. There were nine

CARSs in 1974, which was the maximum number re-

corded during the analysis period. The CARSs repeated

periodically at intervals of 10–20 days, which is consis-

tent with the result of Fig. 2.

Table 1 shows the relationship between the total

rainfall on consecutive HRDs and the duration of con-

secutive HRDs. The numbers in the column in Table 1

indicate the periods of consecutive HRDs between 1902

and 2005. The number of periods decreases as the total

rainfall increases. Regarding the length of a period,

there was a greater percentage of consecutive 3-day

HRDs. A linear relationship is observed between the

number of consecutive HRDs and the total rainfall.

Greater rainfall is observed in the periods with more

consecutive HRDs. There are some extreme cases when

the total rainfall was between 3500 and 5000mm during

7- or 9-day consecutive HRDs. In this study, CARSs are

defined as events with more than 500mm total rainfall,

and the minimum length of a CARS is 3 days.

Variation between intra- and interannual scales has

been considered. For example, Gadgil and Joseph

(2003) showed that the interannual variation in all-India

FIG. 2. The 92-summer ensemble spectrum of the Cherrapunji

daily rainfall (mm). A red noise spectrum (dashed curve) and 95%

level of significance (solid curve) based on a lag-1 autocorrelation

are also shown.

8240 JOURNAL OF CL IMATE VOLUME 30

summer monsoon rainfall is related to the number of

days of rain breaks and active spells. Fujinami et al. (2011)

showed that more frequent and strong submonthly-scale

oscillations, and a greater number of rainy days, were

related to the wet monsoon years in Bangladesh. Figure 4

shows that the cumulative rainfall of CARSs is pro-

portional to total monsoon rainfall amount from June to

September. The cumulative rainfall during the period

without CARSs ranges from 2000 to 6000mm and has a

negative tendency with monsoon total rainfall because

the number of days without CARSs decreases when

CARSs increases.

Figure 5 shows the relationship between the cumulative

number of CARSs and the cumulative rainfall amounts

during CARSs over one year. As in the analysis of Fig. 2,

data for 92 years were used. The bottom and top of

the box are the first and third quartiles, respectively. The

thick band inside the box is the median. The ends of the

whiskers show the minimum and maximum. The number

of CARSs ranges from one to nine per year. The years

with four to nine CARSs over one year compose ap-

proximately half of the total sample years. In the range of

one to four spells per year, there is a tendency for total

rainfall to increase with the number of CARSs. However,

FIG. 3. Daily rainfall (histogram; mm) for 21May–10 October in (a) 1915 and (b) 1974, and the climatological daily

rainfall (solid curve; mm). The gray-shaded areas represent Cherrapunji active rainfall spells (CARSs).

TABLE 1. Relationship between the total rainfall of consecutive heavy rainfall days (HRDs; horizontal line) and the length of consecutive

HRDs (vertical line). Asterisks denote cells with more than 10 periods.

14-day 1 1

13-day 0

12-day 3 1 2

11-day 0

10-day 5 1 3 1

9-day 15 5 3 1 5 1

8-day 13 1 3 5 3 1

7-day 28 1 3 *13 4 3 2 1 1

6-day 57 1 *10 *30 *14 2

5-day 90 2 *46 *28 *11 3

4-day 103 *15 *70 *18

3-day 167 *103 *64

2-day 82 *82

1-day 68 *68

Total 632 272 194 98 38 15 11 1 1 1 1

Total , 500 , 1000 , 1500 , 2000 , 2500 , 3000 , 3500 , 4000 , 4500 ,5000

Total rainfall (mm)

15 OCTOBER 2017 MURATA ET AL . 8241

there is no clear trend toward an increase in rainfall when

there are more than four CARSs per year. It is not always

the case that years with a greater number of CARSs are

wet monsoon years. It implies that CARSs with greater

total rainfalls contribute towetmonsoon years in addition

to a number of CARSs.

Figure 6 shows a composite of the anomalous rainfall

from daily climatological rainfall over India during

CARSs. This map was produced by using the IMD grid

rainfall data for the period 1902 to 2005. Positive and

negative values show regions of excess and less rainfall

during CARSs, respectively. The hatched regions are

significant at the 95% confidence level. While excess

rainfall is observed over northeast India during CARSs,

less rainfall is observed over central India and the

Western Ghats. Usually, the active and break spells of

the Indian monsoon are defined based on the rainfall

over central India. Thus, it is important to note that

northeast India generally has the opposite tendency to

that of central India: the spatial distribution is very

similar to that of the break monsoon composition of the

Indian core region (e.g., Pai et al. 2016).

4. Relationship with large-scale circulation field

Figure 7 is the same as Fig. 6, but the APHRODITE

dataset is used to provide an extended viewover theAsian

monsoon region. Although the analysis period (1958–

2005) is shorter than that in Fig. 6, the distribution of high

and low rainfall areas over India is similar. The regionwith

significantly lower rainfall during CARSs extends longi-

tudinally over Pakistan, central India, the Indochina

Peninsula, and the northern Philippines. The region with

significantly excessive rainfall during CARSs lies south

and north of the relatively low rainfall area. The area of

relatively low rainfall is consistent with the location of the

monsoon trough or continental tropical convergence zone

in past studies (Gadgil 2003). Usually, typhoons and other

cyclonic disturbances, such as depressions or lows, move

westward along the monsoon trough zone and produce

rainfall over the monsoon trough area.

To explain the anomalous rainfall distribution during

CARSs (Fig. 7), data from composite anomalous stream-

function and composite anomalous wind (Figs. 8a,b) and

FIG. 4. Scatterplots of monsoon rainfall (total rainfall amount

from June to September; mm) and the cumulative rainfall with and

without CARSs (solid and open circles, respectively; mm). The

regression line is shown for within CARSs.

FIG. 5. The dependence of the cumulative total rainfall amounts

(mm) during CARSs over 1 yr on the number of CARSs. The

bottom and top of the box are the first and third quartiles. The thick

band inside the box is the median. The ends of the whiskers show

the minimum and maximum values. The data in parentheses next

to the number of CARSs on the horizontal axis show the number of

periods.

FIG. 6. Composite of anomalous rainfall (mm) from daily cli-

matological rainfall during CARSs over India. The India Meteo-

rological Department (IMD) grid dataset was used (1902–2005).

Hatched areas are statistically significant at the 95% level.

8242 JOURNAL OF CL IMATE VOLUME 30

composite vertically integrated anomalous moisture flux

and its divergence (Fig. 8d) were produced by using JRA-

55 reanalysis data. The analysis periodwas 1958–2005. The

composite of anomalous OLR is also shown (Fig. 8c) for

comparison with the convective activity, including ocean

areas. The analysis period is 1979–2005, due to the limited

satellite data. Anomalous anticyclonic circulation (AAC)

appears over central India and the northernBay ofBengal,

at both 200- and 850-hPa levels. Anomalous cyclonic cir-

culations appear over thewestern equatorial IndianOcean

and the Tibetan Plateau at the 200-hPa level. The positive

OLR anomaly (Fig. 8c) and anomalous moisture di-

vergence (Fig. 8d) show intense convective suppression

over the western Indian subcontinent. In the lower tro-

posphere (Fig. 8b), the AAC is extended farther east over

the South China Sea; this location is consistent with the

area of relatively low rainfall in Fig. 7 and the positive

OLR anomaly over the South China Sea and the Philip-

pines in Fig. 8c. The convergence of the anomalous

moisture flux over northeast India (Fig. 8d) corresponds

well with active convection over northeast India during

CARSs (Figs. 6 and 7), although negative OLR (Fig. 8c)

does not represent well the convective activity over

northeast India. The anomalous negative streamfunction is

located over the central equatorial Indian Ocean at the

850-hPa level, and represents a double-cell structure.

To further investigate how synoptic conditions pro-

duce moisture convergence over northeast India, a

composite of vertically integrated moisture flux and di-

vergence during CARSs (Fig. 9b) is compared with a

composite of climatological means throughout June to

September (Fig. 9a). CARSs occupy only 8.5% of total

analysis period, and the composite that excludes CARSs

is almost same as the climatological mean. There is a

large difference in the moisture flux field over 208–308N,

while there is almost no difference south of 208N, except

that the moisture flux is somewhat smaller during

CARSs. Climatologically, cyclonic circulation due to a

monsoon trough appears over central India, and mois-

ture transport is limited over this area (Fig. 9a). In

contrast, a strong westerly moisture transport appears

over central India, which intensifies moisture conver-

gence over northeast India during CARSs (Fig. 9b). The

location of moisture convergence over the foot of the

central Himalayas is shifted slightly northward and in-

tensifies during CARSs.

Figure 10 is the same as Fig. 8, but shows the com-

posites of the CARS onset days. In the lower tropo-

sphere (Fig. 10b), AAC is distributed over the northern

Bay of Bengal, the Indochina Peninsula, the South

China Sea, and the western North Pacific along 108–208N. Compared with Fig. 8b, the AAC does not cover

the Indian subcontinent. The location of the AAC is

almost consistent with the convective suppression rep-

resented by positive OLR anomalies (Fig. 10c) and the

divergence of the anomalous moisture flux (Fig. 10d).

The anomalous negative streamfunction is located to

the south of the AAC over the eastern Indian Ocean,

and shows the double cell structure. Another anomalous

negative streamfunction is located around Japan. In

the upper troposphere (Fig. 10a), the AAC is located

just over northeast India where convection is active

(Fig. 10c). There are two other anomalous negative

streamfunctions situated in the northeast and south of

theAAC. The negative OLR anomaly (Fig. 10c) and the

convergence of the anomalous moisture flux (Fig. 10d)

are distributed not only in northeast India but also ex-

tend to southern China along the 208–308N zone; this

location is consistent with the northern edge of theAAC

at the 850-hPa level (Fig. 10b).

Table 1 shows that greater total rainfall tends to be

observed during longer CARSs. The data were then

divided by the length of CARSs, and compared with

those in Fig. 11. The composite numbers of periods were

21, 40, and 9 for 3-, 5-, and 7-day CARSs, respectively.

Figure 11 shows the composite of anomalous stream-

function and anomalous wind data at 850 hPa for

3- (Figs. 11a,d,g), 5- (Figs. 11b,e,h), and 7-day CARSs

(Figs. 11c,f,i), respectively. The time evolution is also

shown from day22 to day12. On day22 (Figs. 11a–c),

an AAC forms over the South China Sea. This AAC

propagates westward, and the onset of CARSs

(Figs. 11d–f) occurs when the AAC covers the northern

Bay of Bengal. The AAC moves farther westward on

day12 (Figs. 11g–i). The westward propagation speed is

FIG. 7. Composite of anomalous rainfall (mm) from daily cli-

matological rainfall during CARSs over the Asian monsoon re-

gion. The APHRODITE dataset was used (1958–2005).

15 OCTOBER 2017 MURATA ET AL . 8243

in the range of 5–10m s21. On day22, the AAC extends

more zonally toward the western North Pacific in 7-day

periods than in 3-day periods. Furthermore, the AACs

are more intense in 7-day periods than in 3-day periods.

On day12, the westerly anomalies still remain over the

western North Pacific along 208N in the 5- and 7-day

periods. In the 7-day periods, the center of the AAC is

still located in the western North Pacific.

The time evolution of the vertical structure was ana-

lyzed. The composites of the anomalous streamfunction

and anomalous wind at 500 hPa were almost the same

as in Fig. 11 (figures not shown). Figure 12 shows

the composite of the anomalous streamfunction and

anomalous wind at 200 hPa in 5-day periods on day 22

(Fig. 12a), the onset of CARSs (Fig. 12c), and day 12

(Fig. 12e). They are compared with anomalous OLR on

day 22 (Fig. 12b), the onset of CARSs (Fig. 12d), and

day 12 (Fig. 12f) to examine the effect of diabatic

heating on the circulation field in the upper troposphere.

A significant AAC over the Tibetan Plateau tends to

move southward with the onset of CARSs, and lies over

the convectively active area. The comparison between

the onset day of the CARSs and day 12 shows a west-

ward propagation. The composite of 3- and 7-day

CARSs showed similar AAC movement toward the

south, but they did not reach significance (figures not

shown), possibly due to the relatively weak signal for

3-day CARSs and insufficient number of samples for

7-day CARSs.

5. Discussion

a. Synoptic conditions during CARSs

The composite of the CARS onset days implies that

the CARSs begin when the AAC in the lower tropo-

sphere is over the Bay of Bengal (Fig. 10b). This AAC

signal formed in the South China Sea and the western

FIG. 8. Composite of anomalous streamfunction (shading; 106m2 s21) and anomalous wind (vector; m s21) at

(a) 200- and (b) 850-hPa levels from daily climatological values during CARSs. The JRA-55 dataset was used

(1958–2005). (c) As in (a) and (b), but for outgoing longwave radiation (OLR) (shading; Wm22). (d) As in (a) and

(b), but for vertically integrated moisture flux (vector; kgm21 s21) and its divergence (shading; 1024 kgm22 s21).

Hatched areas are statistically significant at the 95% confidence level. Only anomalous wind vectors that showed

a 95% statistically significant difference are plotted.

8244 JOURNAL OF CL IMATE VOLUME 30

North Pacific propagates westward along 108–208N(Fig. 11). An anomalous negative streamfunction is

situated to the south of the AAC, and makes a double-

cell structure (Fig. 10b). During the CARSs, the AAC

propagates farther westward over central India (Fig. 8b),

and causes convective suppression (Fig. 8c). The AAC

over central India has a barotropic structure (Fig. 8a).

These features are consistent with the results of Fujinami

FIG. 9. Composite of vertically integrated moisture flux (vector; kgm21 s21) and its divergence (shading;

1024 kgm22 s21) for (a) June–September climatology and (b) during CARSs.

FIG. 10. As in Fig. 8, but for the onset day composites of each CARS.

15 OCTOBER 2017 MURATA ET AL . 8245

et al. (2014) and other studies in the 10–20-day mode

analysis. The AAC signal in the upper troposphere

moved southward from the Tibetan Plateau and covered

over central and northeast India (Fig. 12).

During CARSs, suppressed convective anomalies are

distributed over a zonally elongated band from southern

Pakistan, via central India, the northern Bay of Bengal,

the Indochina Peninsula, and the Philippines, to the

western North Pacific (Figs. 7 and 10c). A similar elon-

gated band, but with opposite sign, was shown by Moon

et al. (2013) during the active phase of the Indian sum-

mer monsoon, and by Takahashi et al. (2015) during

above-normal precipitation over the Indochina Penin-

sula. The zonally elongated band is formed due to the

westward propagation of AAC.

Wang andXie (1997) simulated the similar elongated

precipitation band over northwest India toward the

equatorial central Pacific. They used the extension of

the model developed by Wang and Li (1994). It is an

equatorial beta-plane channel model with a model

atmosphere consisting of a two-layer free troposphere

and a well-mixed boundary layer, and includes sur-

face evaporation in the moisture budget. The instabil-

ity generated by convection–frictional convergence

produced a low-frequency multiscale convective com-

plex in which the condensation heating coupled moist

Kelvin wave and moist Rossby wave with the gravest

meridional structure (n 5 1 mode). Wang and Xie

(1997) introduced the climatological July mean flow at

200- and 850-hPa levels and July mean surface specific

humidity over Asian monsoon region as the model’s

basic state.

The path, propagating speed, and horizontal scale of

the observed westward propagating AACs in this study

(Fig. 11) are similar to that of the simulated northern cell

of the n 5 1 equatorial Rossby waves in Wang and Xie

FIG. 11. As in Fig. 8b, but for the composites for 2 days before the onset of (a) 3-, (b) 5-, and (c) 7-day CARSs, respectively. (d)–(f) As in

(a)–(c), but for the composites of the onset day of CARSs. (g)–(i) As in (a)–(c), but for the composites of the 2 days after the onset

of CARSs.

8246 JOURNAL OF CL IMATE VOLUME 30

(1997). The structure of the simulated Rossby wave was

highly asymmetric about the equator, featuring the in-

tensified northern cell propagated from the South China

Sea to the Arabian Sea and the much weaker southern

cell that was located much closer to the equator over the

Indian Ocean sector. This happens because the vertical

easterly shear and high specific humidity are confined to

theNorthernHemisphere. Xie andWang (1996) showed

that an easterly shear can effectively destabilize Rossby

waves and the convective heating can further amplify

waves with the most unstable wavelength (about

4000km) for reasonably realistic parameters.

While the composite during CARSs shows the baro-

tropic structure over central India (Figs. 8a,b), the time

evolution of the 200-hPa AAC (Figs. 12a–c) shows that

the AAC in the upper troposphere moves not westward

but southward from the Tibetan Plateau on the onset of

CARSs. It corresponds to a southward extension of the

Tibetan high. Wang and Xie (1996) showed that the

vertical easterly shear tended to trap Rossby waves in

the lower troposphere. A plausible cause of the baro-

tropic structure over central India is that the upper

tropospheric AAC generated as a response to the dia-

batic heating over northeast India is superimposed on

the lower-tropospheric AAC that propagated from the

western North Pacific. It cannot be produced in the

model of Wang and Xie (1997) because topography was

not included in their model. The barotropic structure

over central India during CARSs has been described

during the break in the Indian summer monsoon

(Ramaswamy 1962). There are studies that considered

barotropic instability as amechanism to control a revival

of the Indian summer monsoon after the break events

(Rao 1971; Govardhan et al. 2017). Figures 8a and 10a

show a structure similar to a stationary Rossby wave

pattern (Hoskins and Karoly 1981). However, the mean

easterly flow in the upper troposphere over the south of

the Tibetan Plateau will not form a stationary Rossby

wave train.

Why do CARSs occur when the AAC is located south

of northeast India? Figures 8d and 10d show the

anomalous moisture convergence at the northern por-

tion of theAAC. It can be seen that the area is consistent

with the excessive rainfall area (Figs. 6 and 7). Figure 9b

suggests that the westerly moisture transport over cen-

tral India enhances water vapor convergence over

northeast India together with the moisture transport

from the Bay of Bengal during CARSs. In the climato-

logical condition, the westerly moisture transport does

not exist due to convergence over the Indian monsoon

trough that extends from the northern Bay of Bengal to

northwest India (Fig. 9a).

Murata et al. (2008) compared radiosonde soundings

at Dhaka, Bangladesh, with daily rainfall at Cherrapunji

during the monsoon in 2004, and showed that westerly

wind dominated over the Bengal plain during the

CARSs there. Figure 13 uses JRA-55 grid data corre-

sponding to the location of Dhaka during 1958–2005,

FIG. 12. As in Figs. 11b,e,h, but for 200 hPa for (a)–(c). (d)–(f) As in (a)–(c), but for the composite for OLR (shading; Wm22).

15 OCTOBER 2017 MURATA ET AL . 8247

and shows the correlation between the 925-hPa level

horizontal wind and daily rainfall at Cherrapunji. The

curve of the cumulative frequency distribution shows

that nearly half of the period was an easterly phase. This

situation corresponds with typical monsoon conditions

when the Indian monsoon trough is intensified along the

Ganges plain (Fig. 9a). However, the heavy daily rainfall

was concentrated in the westerly phase of 5–10ms21,

and occupies less than 20% of the overall period. For

the meridional wind component, the southerly wind

component was dominant throughout the monsoon pe-

riod, and the heavy daily rainfall appeared in the

southerly phase of 5–15ms21. This result implies that

the westerly phase or southwesterly wind is a favorable

condition for CARSs. As a small piece of additional

evidence, there is a short story in which a boatman living

in the Bengal plain recognizes a change in wind direction

from easterly to westerly as a sign of an ARS.1

b. Relationship with the break spell of the Indian coreregion

The composite of the ARSs at Cherrapunji (Fig. 6)

shows the opposite tendency to that of the composite of

the ARSs over the Indian core region (e.g., Pai et al.

2016). The AAC at the 850-hPa level (Fig. 8b) over

central India during CARSs weakens the airflow toward

the Western Ghats, as well as the monsoon trough over

the Ganges plain and the northern Bay of Bengal

(Fig. 9), and suppresses the generation of monsoon de-

pressions or lows, which are the main contributor to

rainfall over the Indian core region. This perspective

regarding the Indian monsoon break is similar to that of

Krishnamurti and Ardanuy (1980), and implies that the

onset of CARSs has the potential to predict break spells

over the Indian core region. The CARSs overlap with

61% of break spells and only 2% of active spells in the

Indian core region in the list shown in Pai et al. (2016).

However, Hartmann and Michelsen (1989) shows

that a 30–60-day periodicity is dominant over the Indian

core region, although years with a dominance of a

10–20-day periodicity are also included (Kulkarni et al.

2011). The 30–60-day mode exhibits remarkable north-

ward propagation from the equatorial Indian Ocean to

central India (Sikka and Gadgil 1980; Yasunari 1980).

Recently, Hatsuzuka and Fujinami (2017) investigated

the condition that low pressure systems (LPSs) with a

wavelength of around 6000km develop over Bangladesh.

The LPSs tended to occur when an anomalous cyclonic

circulation in 30–60-day mode was located over central

India and the northern Bay of Bengal, while an AAC in

10–20-day mode was located over the northern Bay of

Bengal. The westward propagation of the AAC from the

northern Bay of Bengal toward central India was unclear

in the LPS cases. Their result implies that combination of

10–20-day and 30–60-day modes produces the difference

in the dominant frequencies between northeast and cen-

tral India. CARSs may include more non-LPS cases as

they showed that convection over Meghalaya Plateau is

more active in the non-LPS cases.

c. Contribution of CARSs to extreme rainfall

Northeast India corresponds to the maximum rain-

fall area in the Indian subcontinent including not only

the Meghalaya Plateau but also the adjoining areas of

FIG. 13. Scatterplots of the dependence of daily rainfall (mm) on (a) zonal and (b) meridional winds (m s21) at the

925-hPa level JRA-55 grid data at 22.58N, 90.08E. The solid lines show the cumulative frequency distribution (%).

1 The short story ‘‘Boatman Tarini’’ written by Tarashankar

Bandyopadhyay, a twentieth-century Bengali novelist.

8248 JOURNAL OF CL IMATE VOLUME 30

the Himalayan range. Xie et al. (2006) pointed out the

importance of narrow mountain ranges in the Indian

subcontinent for the Indian monsoon circulation. In

general, diabatic heating due to convective activity is

an important factor in intensifying the monsoon

circulation.

Daily rainfall equaling the record-breaking levels at

Mumbai (Jenamani et al. 2006) is sometimes observed at

Cherrapunji (Figs. 3 and 13; Guhathakurta 2007). In-

terestingly, however, such extreme daily rainfall ex-

ceeding 500mm does not occur during on 1 or 2

consecutive HRDs and is included in longer CARSs

(Table 1). As in the findings of Goswami et al. (2010),

the fact implies that synoptic conditions during ARSs

are important for heavy rainfall at Cherrapunji. The

influence of the AACs on moisture flux convergence

seems to produce persistent orographic lifting and result

in large cumulative rainfall (Fig. 9). However, the

composite of OLR does not represent well the heavy

rainfall during CARSs (Fig. 8c). The orographic heavy

rainfall may be difficult to be detect by OLR. Hamada

et al. (2015) found that the extreme rainfall events with

maximum near-surface rainfall rates had relatively low

echo-top heights in the statistical analysis of TRMM

observations, and pointed out the weak linkage between

the heaviest rainfall and tallest storms.

The cumulative rainfall during CARSs has high cor-

relation with the total monsoon rainfall (Fig. 4). The

world record precipitation from 3 to 15 days was ob-

served at Réunion Island in the southwestern Indian

Ocean, and the records from 10 to 15 days were caused

by the three approaches of Cyclone ‘‘Hyacinthe’’ in

January 1980 (Chaggar 1984; Kiguchi andOki 2010). On

the other hand, the world record precipitation for more

than 1-month time scales at Cherrapunji is based on the

climatological long monsoon season during June–

September including high-frequency CARSs with the

dominance of 10–20-day periodicity.

Both total rainfalls during a CARS and frequency of

CARSs contribute to the cumulative rainfall (Fig. 5).

There is a linear relationship between the length and

total rainfall of an ARS (Table 1). Figure 11 shows that

AACs with a greater zonal scale tend to produce longer

and more intense CARS. The AACs over the western

North Pacific have been identified as an important

component because they exert a great influence on the

dynamics of East Asia monsoon through the produc-

tion of a Pacific–Japan pattern (Nitta 1987). Longer

persistent AACs, with a greater zonal scale, have been

observed as a specific feature in post–El Niño summers,

and various hypotheses have been proposed to explain

the dynamics, such as wind–evaporation–SST feedback

(Wang et al. 2000) or an Indian Ocean capacitor effect

(Xie et al. 2009). The findings of this study provide ev-

idence for the effect of AACs in the western North

Pacific on the Indian summer monsoon. Some severe

floods in Bangladesh (Hofer and Messerli 2006) oc-

curred in post–El Niño summers. Terao et al. (2013)

investigated the relationship between the rainfall over

Bangladesh and western North Pacific monsoon, and

focused on the post–El Niño summers of 1983, 1988, and

1998. Further investigations of the relationship between

the AACs over the western North Pacific and CARSs

are necessary for better forecasting of intense CARSs.

6. Conclusions

The synoptic conditions of active rainfall spells (ARSs)

at Cherrapunji (CARSs), northeast India, were in-

vestigated. The town is famous for the extreme high rain-

fall that occurs over time scales greater than 1 month. The

data that were predominantly used in this study were the

daily rainfall data from 1902–2005 and the JRA-55 re-

analysis data from 1958–2005. The CARSs correlated with

the monsoon break spells over the Indian core region.

Although several studies described excess rainfall in the

northern part of India during the Indian monsoon break,

research focusing on the intraseasonal variation over

northeast India have been limited.

The CARSs start when an AAC is over the Bay of

Bengal. The orographic rainfall at Cherrapunji in-

tensifies when a westerly wind prevails in the northern

portion of the AAC and enhances moisture flux con-

vergence over northeast India. The AAC propagates

westward from the South China Sea, and the sameAAC

suppresses convection over central India and the

Western Ghats. In the upper troposphere, the Tibetan

high tends to extend southward with the onset of

CARSs.While the synoptic conditions during the Indian

monsoon break have been described in terms of a

northward shift of the Indian monsoon trough, the

westward-propagating synoptic system plays an impor-

tant role in the occurrence of CARSs: this implies the

possibility of forecasting CARSs by monitoring the

AAC over the South China Sea and the western North

Pacific. A linear relationship was found between the

length of a CARS and the total rainfall during the spell.

Longer and heavier CARSs occurred when zonally

larger AAC existed over the South China Sea and the

western North Pacific. These findings provide evidence

that AACs in the western North Pacific control the

severity of CARSs and have an effect on the Indian

summer monsoon.

Cherrapunji is a calcium-rich area geologically, and

there are several limestone caves that have a potential to

reconstruct the paleo monsoon climate. For example,

15 OCTOBER 2017 MURATA ET AL . 8249

Myers et al. (2015) compared the stalagmite record of

recent 50 years at Mawmluh cave in Cherrapunji with

indices that represent Pacific decadal variabilities. The

fact that the rainfall at Cherrapunji occurs during the

break in the Indian core region may improve the in-

terpretation, although it is still uncertain whether the

same mechanism can be applicable on longer (in-

terannual to centennial) time scales.

Acknowledgments. The authors are grateful to the ed-

itor (Dr. Jason Evans) and anonymous reviewers for ap-

propriate and thoughtful comments and suggestions. The

authors thank to Drs. Ajit Tyagi, B. N. Goswami, G. S.

Bhat, and Someshwar Das for their supports to the au-

thors’ study in northeast India. Drs. Kazuo Ando and

Masayuki Onishi provided useful information about life in

the Bengal plain. This study was conducted with support

from a Grant-in-Aid for Scientific Research (S26220202

andA23241057) from the Japan Society for the Promotion

of Sciences (JSPS). Portions of this study were conducted

by the joint research program of the Institute for Space-

Earth Environmental Research, Nagoya University.

REFERENCES

Annamalai, H., and J. M. Slingo, 2001: Active/break cycles: Diagnosis

of the intraseasonal variability of the Asian summer monsoon.

Climate Dyn., 18, 85–102, doi:10.1007/s003820100161.

Chaggar, T. S., 1984: Reunion sets new rainfall records. Weather,

39, 12–14, doi:10.1002/j.1477-8696.1984.tb05440.x.

Chen, T.-C., and J.-M. Chen, 1993: The 10–20-daymode of the 1979

Indian monsoon: Its relation with the time variation of mon-

soon rainfall. Mon. Wea. Rev., 121, 2465–2482, doi:10.1175/

1520-0493(1993)121,2465:TDMOTI.2.0.CO;2.

Ebita, A., and Coauthors, 2011: The Japanese 55-year Reanalysis

‘‘JRA-55’’: An interim report. Sci. Online Lett. Atmos., 7, 149–

152, https://doi.org/10.2151/sola.2011-038.

Fujinami, H., and Coauthors, 2011: Characteristic intraseasonal

oscillation of rainfall and its effect on interannual variability

over Bangladesh during boreal summer. Int. J. Climatol., 31,

1192–1204, doi:10.1002/joc.2146.

——, T. Yasunari, and A. Morimoto, 2014: Dynamics of distinct

intraseasonal oscillation in summer monsoon rainfall over the

Meghalaya–Bangladesh–western Myanmar region: Covari-

ability between the tropics and mid-latitudes. Climate Dyn.,

43, 2147–2166, doi:10.1007/s00382-013-2040-1.

Gadgil, S., 2003: The Indian monsoon and its variability.

Annu. Rev. Earth Planet. Sci., 31, 429–469, doi:10.1146/

annurev.earth.31.100901.141251.

——, and P. V. Joseph, 2003: On breaks of the Indian monsoon.

J. Earth Syst. Sci., 112, 529–558, doi:10.1007/BF02709778.

Goswami, B. B., P. Mukhopadhyay, R. Mahanta, and B. N.

Goswami, 2010: Multiscale interaction with topography and

extreme rainfall events in the northeast Indian region.

J. Geophys. Res., 115, D12114, doi:10.1029/2009JD012275.

Govardhan, D., V. B. Rao, and K. Ashok, 2017: Understanding the

revival of the Indian summer monsoon after breaks. J. Atmos.

Sci., 74, 1417–1429, doi:10.1175/JAS-D-16-0325.1.

Guhathakurta, P., 2007: Highest recorded point rainfall over India.

Weather, 62, 349, doi:10.1002/wea.154.

Hamada, A., Y. N. Takayabu, C. Liu, and E. J. Zipser, 2015: Weak

linkage between the heaviest rainfall and tallest storms. Nat.

Commun., 6, 6213, doi:10.1038/ncomms7213.

Hartmann, D. L., and M. L. Michelsen, 1989: Intraseasonal peri-

odicities in Indian rainfall. J. Atmos. Sci., 46, 2838–2862,

doi:10.1175/1520-0469(1989)046,2838:IPIIR.2.0.CO;2.

Hatsuzuka, D., and H. Fujinami, 2017: Effects of the South Asian

monsoon intraseasonal modes on genesis of low pressure sys-

tems over Bangladesh. J. Climate, 30, 2481–2499, doi:10.1175/

JCLI-D-16-0360.1.

Hofer, T., and B. Messerli, 2006: Floods in Bangladesh. United

Nations Press, 450 pp.

Hoskins, B. J., and D. J. Karoly, 1981: The steady linear re-

sponse of a spherical atmosphere to thermal and oro-

graphic forcing. J. Atmos. Sci., 38, 1179–1196, doi:10.1175/

1520-0469(1981)038,1179:TSLROA.2.0.CO;2.

Houze, R. A., 2012: Orographic effects on precipitating clouds.

Rev. Geophys., 50, RG1001, doi:10.1029/2011RG000365.

Jenamani, R. K., S. C. Bhan, and S. R. Kalsi, 2006: Observational/

forecasting aspects of the meteorological event that caused a re-

cord highest rainfall in Mumbai. Curr. Sci., 90, 1344–1362.

Jennings, A. H., 1950: World’s greatest observed point rainfalls.

Mon. Wea. Rev., 78, 4–5, doi:10.1175/1520-0493(1950)078,0004:

WGOPR.2.0.CO;2.

Kiguchi, M., and T. Oki, 2010: Point precipitation observation

extremes in the world and Japan. J. Japan Soc. Hydrol. Water

Resour., 23, 231–247, doi:10.3178/jjshwr.23.231.

Kikuchi, K., and B. Wang, 2009: Global perspective of the quasi-

biweekly oscillation. J. Climate, 22, 1340–1359, doi:10.1175/

2008JCLI2368.1.

Kobayashi, S., and Coauthors, 2015: The JRA-55 reanalysis:

General specifications and basic characteristics. J. Meteor.

Soc. Japan, 93, 5–48, doi:10.2151/jmsj.2015-001.

Krishnamurti, T. N., and H. N. Bhalme, 1976: Oscillations of a

monsoon system. Part I. Observational aspects. J. Atmos.

Sci., 33, 1937–1954, doi:10.1175/1520-0469(1976)033,1937:

OOAMSP.2.0.CO;2.

——, and P. Ardanuy, 1980: The 10 to 20-day westward propa-

gating mode and ‘‘breaks in the monsoons.’’ Tellus, 32, 15–26,

doi:10.1111/j.2153-3490.1980.tb01717.x.

Kulkarni, A., R. Kripalani, S. Sabade, and M. Rajeevan, 2011: Role

of intra-seasonal oscillations in modulating Indian summer

monsoon rainfall. Climate Dyn., 36, 1005–1021, doi:10.1007/

s00382-010-0973-1.

Liebmann, B., and C. A. Smith, 1996: Description of a complete

(interpolated) outgoing longwave radiation dataset. Bull.

Amer. Meteor. Soc., 77, 1275–1277.

Moon, J.-Y., B. Wang, K.-J. Ha, and J.-Y. Lee, 2013: Tele-

connections associated with Northern Hemisphere summer

monsoon intraseasonal oscillation. Climate Dyn., 40, 2761–

2774, doi:10.1007/s00382-012-1394-0.

Murata, F., T. Hayashi, J.Matsumoto, andH.Asada, 2007: Rainfall

on the Meghalaya Plateau in northeastern India—One of the

rainiest places in the world. Nat. Hazards, 42, 391–399,

doi:10.1007/s11069-006-9084-z.

——, T. Terao, T. Hayashi, H. Asada, and J. Matsumoto, 2008:

Relationship between atmospheric conditions at Dhaka,

Bangladesh, and rainfall at Cherrapunjee, India.Nat.Hazards,

44, 399–410, doi:10.1007/s11069-007-9125-2.

Myers, C. G., J. L. Oster, W. D. Sharp, R. Bennartz, N. P. Kelley,

A. K. Covey, and S. F.M. Breitenbach, 2015: Northeast Indian

8250 JOURNAL OF CL IMATE VOLUME 30

stalagmite records Pacific decadal climate change: Implica-

tions for moisture transport and drought in India. Geophys.

Res. Lett., 42, 4124–4132, doi:10.1002/2015GL063826.

Nitta, T., 1987: Convective activities in the tropical western Pacific

and their impact on the Northern Hemisphere summer cir-

culation. J. Meteor. Soc. Japan, 65, 373–390, doi:10.2151/

jmsj1965.65.3_373.

Ohsawa, T., T. Hayashi, Y. Mitsuta, and J. Matsumoto, 2000: In-

traseasonal variation of monsoon activities associated with the

rainfall over Bangladesh during the 1995 summer monsoon

season. J. Geophys. Res., 105, 29 445–29 459, doi:10.1029/

2000JD900499.

Pai, D. S., L. Sridhar, M. Rajeevan, O. P. Sreejith, N. S. Satbhai,

and B. Mukhopadhyay, 2014: Development of a new high

spatial resolution (0.258 3 0.258) long period (1901–2010) dailygridded rainfall data set over India and its comparison with

existing data sets over the region. Mausam, 65, 1–18.

——, ——, and M. R. Ramesh Kumar, 2016: Active and break

events of Indian summer monsoon during 1901–2014. Climate

Dyn., 46, 3921–3939, doi:10.1007/s00382-015-2813-9.

Rajeevan,M., S. Gadgil, and J. Bhate, 2010:Active and break spells

of the Indian summer monsoon. J. Earth Syst. Sci., 119, 229–

247, doi:10.1007/s12040-010-0019-4.

Ramaswamy, C., 1962: Breaks in the Indian summer monsoon as a

phenomenon of interaction between the easterly and the sub-

tropical westerly jet streams. Tellus, 14, 337–349, doi:10.3402/tellusa.v14i3.9560.

Rao, V. B., 1971: Dynamic instability of the zonal current

during a break monsoon. Tellus, 23, 111–112, doi:10.1111/

j.2153-3490.1971.tb00552.x.

Romatschke, U., S. Medina, and R. A. Houze Jr., 2010: Regional,

seasonal, and diurnal variations of extreme convection in the

South Asian region. J. Climate, 23, 419–439, doi:10.1175/

2009JCLI3140.1.

Shrestha, D., P. Singh, and K. Nakamura, 2012: Spatiotemporal

variation of rainfall over the central Himalayan region re-

vealed by TRMM precipitation radar. J. Geophys. Res., 117,D22106, doi:10.1029/2012JD018140.

Sikka, D. R., and S. Gadgil, 1980: On the maximum cloud zone

and the ITCZ over Indian longitudes during the southwest

monsoon. Mon. Wea. Rev., 108, 1840–1853, doi:10.1175/

1520-0493(1980)108,1840:OTMCZA.2.0.CO;2.

Takahashi, H. G., H. Fujinami, T. Yasunari, J. Matsumoto, and

S. Baimoung, 2015: Role of tropical cyclones along the mon-

soon trough in the 2011 Thai flood and interannual variability.

J. Climate, 28, 1465–1476, doi:10.1175/JCLI-D-14-00147.1.

Terao, T., F. Murata, A. Habib, S. H. Bhuiyan, S. A. Choudhury,

and T. Hayashi, 2013: Impacts of rapid warm-to-cold ENSO

transitions on summer monsoon rainfall over the northeastern

Indian subcontinent. J. Meteor. Soc. Japan, 91, 1–21, doi:10.2151/jmsj.2013-101.

Wang, B., and T. Li, 1994: Convective interaction with boundary-

layer dynamics in the development of a tropical intra-

seasonal system. J. Atmos. Sci., 51, 1386–1400, doi:10.1175/1520-0469(1994)051,1386:CIWBLD.2.0.CO;2.

——, and X. Xie, 1996: Low-frequency equatorial waves in ver-

tically sheared zonal flow. Part I: Stable waves. J. Atmos.

Sci., 53, 449–467, doi:10.1175/1520-0469(1996)053,0449:

LFEWIV.2.0.CO;2.

——, and ——, 1997: A model for the boreal summer intra-

seasonal oscillation. J. Atmos. Sci., 54, 72–86, doi:10.1175/1520-0469(1997)054,0072:AMFTBS.2.0.CO;2.

——, R.Wu, andX. Fu, 2000: Pacific–East Asian teleconnection:

How does ENSO affect East Asian climate? J. Climate,

13, 1517–1536, doi:10.1175/1520-0442(2000)013,1517:

PEATHD.2.0.CO;2.

Xie, S.-P., H. Xu, N. H. Saji, Y. Wang, and W. T. Liu, 2006: Role

of narrowmountains in large-scale organization ofAsianmonsoon

convection. J. Climate, 19, 3420–3429, doi:10.1175/JCLI3777.1.——, K. Hu, J. Hafner, H. Tokinaga, Y. Du, G. Huang, and

T. Sampe, 2009: Indian Ocean capacitor effect on Indo–

western Pacific climate during the summer following El Niño.J. Climate, 22, 730–747, doi:10.1175/2008JCLI2544.1.

Xie, X., and B. Wang, 1996: Low-frequency equatorial waves in

vertically sheared zonal flow. Part II: Unstable waves. J. Atmos.

Sci., 53, 3589–3605, doi:10.1175/1520-0469(1996)053,3589:

LFEWIV.2.0.CO;2.

Yasunari, T., 1980: A quasi-stationary appearance of 30 to 40 day

period in the cloudiness fluctuations during the summer

monsoon over India. J. Meteor. Soc. Japan, 58, 225–229,

doi:10.2151/jmsj1965.58.3_225.

Yatagai, A., O. Arakawa, K. Kamiguchi, H. Kawamoto, M. I.

Nodzu, and A. Hamada, 2009: A 44-year daily gridded pre-

cipitation dataset for Asia based on a dense network of rain

gauges. Sci. Online Lett. Atmos., 5, 137–140.

——, K. Kamiguchi, O. Arakawa, A. Hamada, N. Yasutomi, and

A.Kitoh, 2012:APHRODITE: Constructing a long-term daily

gridded precipitation dataset for Asia based on a dense net-

work of rain gauges. Bull. Amer. Meteor. Soc., 93, 1401–1415,

doi:10.1175/BAMS-D-11-00122.1.

Yoshino,M.M., 1975:Climate in a Small Area. University of Tokyo

Press, 540 pp.

15 OCTOBER 2017 MURATA ET AL . 8251