Theor. Appl. Climatol. (2009) 95: 135–149 DOI 10.1007/s00704-007-0365-6 Printed in The Netherlands

1

Asian Disaster Preparedness Center, Asian Institute of Technology, Klong Luang, Pathumthani, Thailand Department of Geophysics, Banaras Hindu University, Varanasi, India 3 Centre for Atmospheric Sciences, Indian Institute of Technology, Hauz Khas, New Delhi, India 2

The evolution of mean conditions of surface meteorological fields during active==break phases of the Indian summer monsoon P. V. S. Raju1 , R. Bhatla2 , U. C. Mohanty3 With 11 Figures Received 19 May 2006; Accepted 30 October 2007; Published online 28 April 2008 # Springer-Verlag 2008

Summary The intra-seasonal variability of the Indian summer monsoon (ISM) is an important aspect that has been studied by various researchers, but still requires further examination to advance our understanding. A break monsoon situation, which is nothing more than a considerable reduction in rainfall over most parts of the Indian sub-continent, lasting from a few days to a few weeks, is one of the very significant periods of intra-seasonal variability of the ISM. In the present study, we examine the relationship between the break monsoon situation and changes in surface meteorological parameters using the National Centers for Environmental Prediction (NCEP) reanalysis datasets. Break monsoon situations persisting for a period of 3 days or more during 1968–1996 have been considered. An examination of average pentad data, commencing two pentads prior to the onset of the break situation to two pentads after the cessation of the break monsoon, is carried out in a sequential mode. The regions with statistically significant differences between prior and present break pentad are delineated by a Student’s t-test at the 95% confidence level. The aim is to examine the evolution of the break monsoon and its relationship with surface meteorological parameters and to identify possible precursors, that can assist in the identification of the initiation or cessation of the break monsoon situation over India. It is noted that the onset of the Correspondence: Dr. P. V. S. Raju, Asian Disaster Preparedness Center, Asian Institute of Technology, Klong Luang, Pathumthani 12120, Thailand, e-mail: [email protected]

break is a gradual process, whereas the cessation of the break is much more rapid. Prior to the occurrence of the break over India, the surface meteorological fields and derived surface heat fluxes clearly indicate significant changes over the South China Sea and South-East Asia. The precursor signals over the South China Sea can be used as potential predictors for the forecasting of the break period in the summer monsoon over India.

1. Introduction The Indian summer monsoon (ISM) exhibits variability on a wide range of spatial and temporal scales, which is of critical importance to the vast population dependent on the monsoon. The intraseasonal variability of the monsoon that manifests itself as a series of rainy and dry spells, which may extend from several days to weeks within the monsoon season, not only influences the day-to-day life and economic activities across India but also has a cumulative effect at the seasonal scale, leading to widespread drought and floods. A number of studies have been carried out which identify breaks based on different criteria over regions at differing spatial scales (Krishnan et al. 2000; Annamalai and Slingo 2001; De and Mukopadhyaya 2002). In this study, we considered the break monsoon days

136

P. V. S. Raju et al.

based on a classical method defined by the India Meteorological Department (IMD), as the period when the monsoon trough is shifted northward from its normal position over the Indo-Gangetic plains to the foothills of the Himalayas (Ramamurthy 1969; Rao 1976) and the absence of easterly winds over the northern parts of India at 1.5 km above sea level (De et al. 1998). During this period, precipitation activity over India becomes considerably subdued but increases over some parts of the country such as the foothills of the Himalayas, north-eastern states of India, and southern India. There is considerable overlap between the traditional method used by the IMD and the rain breaks identified by Gadgil and Joseph (2003). In addition, there is some overlap with the breaks defined by Krishnan et al. (2000) based on days with positive outgoing long-wave (OLR) radiation anomalies over north-west and central India (i.e., western part of monsoon zone). It is clear that the OLR anomaly pattern of the breaks exhibits large negative anomalies (Krishan et al. 2000), and negative rainfall anomalies (Ramamurthy 1969) over the eastern zone. Due to these differences in definition, some breaks are not coincident (e.g., 25th June to 9th July, 1992) with the break days identified using the traditional IMD method. Thus, the criteria adopted by Krishnan et al. (2000) do not always lead to the break monsoon as defined by the traditional IMD method. A large number of studies have identified the morphology of breaks during the Indian summer monsoon. Ramaswamy (1962) found that break periods in the ISM are influenced by the intrusion of mid-latitude troughs into the Indo-Pakistan region in the middle and upper troposphere. Unninayar and Murakami (1978) discovered that a bifurcation of the Tibetan high under the influence of a mid-latitude trough near the IndoPakistan region resulted in weak monsoon periods. Raman and Rao (1981) found that changes associated with a weak monsoon include the occurrence of a stagnant blocking ridge in the upper troposphere between 90
E and 115
E over East Asia. Using surface pressure data spanning 40 years (1933–1972), Krishnamurti and Ardanuy (1980) concluded that break monsoon spells are associated with the westward propagation of trough-ridge systems with a periodicity of about 10–20 days. Rodwell (1997) found that breaks in

the Indian summer monsoon could be triggered by extra-tropical weather systems in the Southern Hemisphere. Based on in diagnostic analyses and a simple numerical model, Krishnan et al. (2000) conducted a spatial and temporal evaluation of convective anomalies associated with monsoon breaks by using 17 years (1979–1995) of satellite derived OLR and 850 hPa wind data from NCEP reanalysis. They demonstrated that low latitude Rossby wave dynamics, in the presence of a monsoon basic flow, driven by steady northsouth differential heating, is the primary physical mechanism controlling monsoon breaks. A comprehensive analysis of daily rainfall departures, wind anomalies and satellite derived OLR associated with the commencement=cessation of the break monsoon condition have been studied by De and Mukhopadhyay (2002). Annamalai and Slingo (2001) stated that during the active phase of the monsoon, convection is enhanced significantly over the Indian sub-continent, extending to the Bay of Bengal and equatorial west Pacific. Furthermore, Joseph and Siji Kumar (2003) found that during the break monsoon period the strong cross-equatorial low level jet stream (LLJ) at 850 hPa is oriented south-eastwards over the eastern Arabian Sea and flows east between SriLanka and the equator. In contrast, in the active monsoon period the LLJ axis moves from the central Arabian Sea, eastward through peninsular India, and provides moisture for increased convection and the formation of the monsoon depression over the Bay of Bengal. Gadgil and Joseph (2003) found that OLR and circulation patterns are associated with rain breaks and active spells, and identified that a quadrapole over the Asia-west Pacific region is a basic feature of weak spells in the intra-seasonal variation over the Asia-west Pacific region. A number of studies (Pisharoty 1965; Das 1983; Mohanty et al. 1983; Mohanty and Mohan Kumar 1990) have demonstrated that the air-sea interaction processes over the Indian seas play a considerable role in the initiation and maintenance of convective activity over the monsoon region. The large-scale evaporation and transportion of moisture from the oceanic region to the Indian land mass is considered to be the main source of convection over land, producing monsoonal rainfall. Wang et al. (2005) suggested that the antecedents of the active=break monsoon

The evolution of mean conditions of surface meteorological fields

137

Fig. 1. Geographical map showing different region

emerge in the western equatorial Indian Ocean. However, a question arises regarding how these convective anomalies that lead to the active and break periods of the monsoon occur. In general, during the break monsoon, a considerable reduction in precipitation, cloudiness and low-level winds over most parts of the country are observed. These changes can be related to changes in air-sea interaction processes and large-scale flow fields, both at the surface as well as in the overlying atmosphere. The purpose of the present study is to examine the changes in mean surface meteorological fields (basic fields and surface fluxes of heat and moisture) over India and the adjoining seas (Fig. 1) that lead to the break situation in the summer monsoon over India. 2. Data and methodology The availability of the reliable and homogeneous global meteorological reanalysis dataset from the National Center for Environmental Prediction= National Center for Atmospheric Research (NCEP=NCAR) has been the main impetus in carrying out a composite study of monsoon break cases over a 29 year period (1968–1996) in the context of the observed changes in surface meteorological fields associated with evolution= cessation of the monsoon break condition over India. The dataset is a result of the co-operative effort of NCEP to produce global reanalysis of atmospheric fields to support the needs of the

research and climate monitoring community. The effort involves the recovery of land surface, ship, rawinsonde, aircraft, satellite data and delayed mode GTS data, and quality control and assimilation with a four dimensional data assimilation system that is kept unchanged over the entire period (Kalnay et al. 1996). The data have little influence on the model and are close to the actual station data fed into the model to derive the gridded dataset (Kalnay et al. 1996). Various studies (e.g., Annamalai et al. 1999; Krishnan et al. 2000; Sperber et al. 2000) which have utilised of the NCEP reanalysis confirm the usefulness of this dataset. However, due to low model resolution and the lack of reliable observations, Barros and Lang (2003) found that moisture is under-estimated by the NCEP reanalysis over central Nepal. Break monsoon situations have been taken from an IMD scientific report (De et al. 1998) for the period 1968–1996 and are presented in Table 1. The breaks which extend over a period of three days or more have been considered and examined. Furthermore the break situations that occur after the establishment of the monsoon over the entire country, but before the commencement of the withdrawal epoch of the monsoon, are only considered. From the global reanalysis dataset at a 2.5
 2.5
resolution, data for the duration of all break situations have been extracted over the region 30
S–30
N; 30
E– 150
E for the period 1968–1996. In addition,

138

P. V. S. Raju et al.

Table 1. Break periods of Indian summer monsoon during 1968–1996 Year

Break period

No. of days

1968 1969 1970 1971 1972

25–29 August 17–20 August 12–25 July 17–20 August 17 July–3 August, 1–6 September 23 July–1 August 30 August–3 September 24–28 July 15–18 August 16–21 July 17–23 July, 15 August–3 September 17–20 July 26–30 July, 23–27 August 22–25 August 20–24 July 22–25 August 29 August–3 September 28 July–1 August 5–8 July, 13–15 August 10–12 July, 29–31 July 8–10 July, 27–31 July 3–9 September 19–21 July 12–15 August 1–5 July

5 4 14 4 21

1973 1974 1975 1977 1978 1979 1980 1981 1983 1984 1985 1986 1987 1988 1989 1990 1991 1993 1995 1996

9 5 5 4 6 27 4 10 4 5 4 6 5 7 6 8 6 3 4 5

two pentad datasets covering a ten day priod before and after the break period, are considered. The daily averaged reanalysis datasets are used to composite the surface meteorological parameters, commencing two pentads before the onset of break situation (pre-break2, pre-break1) and two pentads after the cessation of break monsoon, (post-break1 and post-break2). The differences in the surface meteorological fields between the two adjacent phases of the break monsoon situations (pre-break2 and pre-break1, pre-break1 and break, break and post-break1, post-break1 and post-break2) are examined in order to study changes in the evolution of the surface fields. The pre- and post-break pentads are the periods of active phase of the monsoon. The tendency in the change of the surface meteorological fields with respect to the turning of the sequence of the events is considered in order to illustrate the variations which take place that lead to a shift from active to break monsoon and break to active monsoon. i.e., the difference between pre-break2

and pre-break1, pre-break1 and break, break and post-break1, and post-break1 and post-break2. Furthermore, the Student’s t-test has been applied to the basic surface parameters to identify the variations of the significant regions at the 95% confidence level. The Student’s t-test is computed using the formula: X nþ1  X n t ¼ rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 1 1
þ Nnþ1 Nn where
is the standard deviation, Xn þ 1 and Xn are the parameters, and Nn þ 1 and Nn are the number of cases (the subscript n þ 1 and n indicate the prior and present break pentad) respectively. The t-test was used to test the significance of the variations at the 95% level and was applied under the assumption that the data are normally distributed. Furthermore, to compare the net balance of the radiative and turbulent fluxes of heat and moisture at the air-land interface, the net heat budget equation (W m2 ) has computed as follows: QN ¼ QR  QB  QH  QE where QN is the net heat flux (W m2 ), QR is the incoming shortwave radiation flux (W m2 ), QB is the effective outgoing long-wave radiation flux (W m2 ), QH is the sensible heat flux (W m2 ), and QE is the latent heat flux (W m2 ). In the above budget equation, QR contributes significantly to the heat gain in the tropics, whereas the other terms contribute mainly towards heat loss. 3. Results Variations in the basic surface meteorological fields and derived surface heat fluxes that lead to a break situation in the summer monsoon are examined over India and the adjoining seas. All computations are made using daily averaged NCEP reanalysis data to estimate the variance for each case separately and then averaged over 31 cases to arrive at a composite. The surface fields with a reasonable degree of variability during the pre-break, break and post-break periods are presented in the following sections. In Figs. 2–11, we present a) mean composite climatology of break days and the difference between b) prebreak2 and pre-break1, c) pre-break1 and break,

The evolution of mean conditions of surface meteorological fields

139

Fig. 2. Geographical distributions of precipitation rate (mm day1 ) for (a) break and the difference of (b) pre break2–pre break1 (c) pre break1–break (d) break–post break1 (e) post break1–post break2. Shaded regions are statistical significance at 95% confidence level

and d) break and post-break1. The statistically significant regions at the 95% confidence level are shaded in Figs. 2–11. 3.1 Variations in the basic meteorological fields The geographical distribution of precipitation rate (mm day1 ) is depicted in Fig. 2. During the break period (Fig. 2a) a higher precipitation rate is found closer to the foothills of the Himalayas, sub-Himalayan West Bengal, northeastern states of India and coastal regions of the Bay of Bengal, coastal Andhra Pradesh, Tamilnadu, Myanmar, South China Sea and south equatorial Indian Ocean. The rainfall rate is lower over central and western=north-western India during the break phase. The pattern of rainfall distribution during the break period is consistent with the work of Ramamuthy (1969). In the pre-break2 to pre-break1 phase, there is significant reduction of 2–4 mm day1 in the precipitation rate extending from the South China Sea, east equatorial Indian Ocean, adjoining

Vietnam, to the north Bay of Bengal (Fig. 2b). There is an increase in precipitation rate by 1 mm day1 over the south Indian Ocean. During prebreak1 to break (Fig. 2c), the area of reduced precipitation rate observed in the earlier phase extends westwards and occupies the region of the monsoon trough zone, north and central Bay and north-east Arabian Sea. The precipitation rate has reduced by 2–4 mm day1 in the monsoon trough zone. There is an increase in the precipitation rate by 1 mm day1 in the south equatorial Indian Ocean adjoining Indonesia, south-west Bay, coastal Tamilnadu and Andhra Pradesh (Fig. 2c). After the cessation of the break, the process is reversed i.e., rainfall activity increases significantly over the Bay of Bengal, Arabian Sea and monsoon trough zone (Fig. 2d). During post-break1 to post-break2, precipitation continues to increase along the monsoon trough zone, central India, and north and central Bay of Bengal, but the increase is not found to be significant (Fig. 2e). It should be noted that the process of break commencement is gradual, but

140

P. V. S. Raju et al.

Fig. 3. Geographical distributions of mean sea-level pressure (hPa) for (a) break and the difference of (b) pre break2–pre break1 (c) pre break1–break (d) break–post break1 (e) post break1–post break2. Shaded regions are statistical significance at 95% confidence level

that of break cessation is faster as is evident from Fig. 2d and e when compared with Fig. 2b and c. It is important to note that the large change in precipitation rate over the South China Sea during pre-break2 to pre-break1 (Fig. 2b) could be a precursor to the commencement of the break situation over the Indian region. Following the changes over the South China Sea, large changes are found in the monsoon trough zone extending from the north-west Bay of Bengal across central India to the north-east Arabian Sea during the pre-break and break to post-break periods. The geographical distribution of mean sea level pressure (MSLP) is illustrated in Fig. 3. During the break phase of the monsoon (Fig. 3a), the pressure gradient is weak over the equatorial Indian Ocean (5
N–5
S) and south Bay of Bengal. It is rather steep over the north Arabian Sea, and north-west India. Figure 3b shows that during pre-break2 to pre-break1, MSLP increases extensively over the entire monsoon trough zone and most parts of the country. In addition, over the Bay of Bengal, Indo-China and South China

Sea there are considerable increases in MSLP by 1–1.5 hPa. During pre-break to break, the pressure further rises over the Indian landmass, Arabian Sea and parts of Arabia. The maximum rise of 1.5–2 hPa is found in the seasonal low pressure zone over north-west and central India (Fig. 3c). The area of increased pressure shifts west-northwestwards as the break approaches (notice the changes in Fig. 3b and c). The increased pressure over India reduces the northsouth pressure gradient between the seasonal heat low over India and high pressure area over the Mascarene high in the Indian Ocean, resulting in a weak cross-equatorial flow and weakening of the monsoon current (Malurkar 1950). During the post-break period, a reversal in the pattern of pressure change can be seen. However, noteworthy changes are observed from the break to post-break1 period only. It is once again observed that significant pressure changes leading to the break occur over a comparatively longer period, and extend over two pentads, while the cessation of the break is faster and momentous

The evolution of mean conditions of surface meteorological fields

141

Fig. 4. Geographical distributions of zonal wind (m sec1 ) for (a) break and the difference of (b) pre break2–pre break1 (c) pre break1–break (d) break–post break1. Shaded regions are statistical significance at 95% confidence level

changes are observed in the first post-break pentad period only. Important changes in meteorological fields are not observed during post-break1 to post-break2. Hence, in the subsequent discussion, this aspect is not presented. The distributions of zonal and meridional wind (at 10 m) are presented in Figs. 4 and 5, respectively. Figure 4a shows that during the break phase of the monsoon, weak westerlies prevail north of the equator and easterlies to the south. From pre-break2 to pre-break1 there is a substantial decrease in the zonal wind (0.5–1.5 m sec1 ) along a belt centred along 15
N. The belt extends from the central Arabian Sea to the South China Sea across the central Bay of Bengal and Mayanmar. There is a slight increase in wind speed over the eastern end of the monsoon trough and west Pacific Ocean. There is also a considerable decrease in the easterlies of 1 m sec1 in the central equatorial Indian Ocean (Fig. 4b). During the pre-break1 to break period, the zonal wind along the belt at 15
N continues to decrease and extend westwards, while in the monsoon trough zone, over the Indo-Gangetic plains, it continues to increase (Fig. 4c). The commencement of the break may be anticipated based on a decrease in the zonal wind speed over the South China Sea and an increase over the west Pacific Ocean during pre-break2 to pre-break1. The decrease in the zonal wind along 15
N is the result of a weakening of the monsoon westerlies and

the low level Somali jet. This is due to the cross-equatorial low level jet (at 850 hPa) being oriented south-eastward over the eastern Arabian Sea and flowing between SriLanka and the equator (Joseph and Sijikumar 2003). Furthermore, the weak low level westerlies which prevail over the Arabian Sea and peninsular India have been linked to years with deficit rainfall (Annamalai et al. 1999; Raju et al. 2002). In the Gangetic plains, the dry westerlies increase in strength and the moist easterlies weaken as a result of a shift in the monsoon trough to the north. In the Indian Ocean, particularly close to equator between 75
E and 105
E, there is a decrease in the strength of the zonal winds on either side of the equator indicating a weakening of the monsoonal current. During the revival of the monsoon i.e., break to post-break1 (Fig. 4d), the process is reversed. There is a decrease in the strength of the westerlies and an establishment of easterlies over the Indo-Gangetic plains, and an increase in the westerlies over the Indian peninsula and adjoining Arabian Sea and Bay of Bengal. In the equatorial south Indian Ocean, stronger trade winds are observed. Figure 5a shows that during the break, the whole of the monsoon domain is dominated by a southerly flow with maximum strength over the Arabian Sea along the Somalian coast. From pre-break2 to pre-break1, there is a reduction in the southerly component of the wind by 1.5 m sec1

142

P. V. S. Raju et al.

Fig. 5. Geographical distributions of meridional wind (m sec1 ) for (a) break and the difference of (b) pre break2–pre break1 (c) pre break1–break (d) break–post break1. Shaded regions are statistical significance at 95% confidence level

in the region between the equator to 15
S and 80
E to 105
E (Fig. 5b). The maximum significant decrease in the meridional wind is observed over the east equatorial Indian Ocean and South China Sea, and can serve as a precursor to the approaching break when one examines changes during the pre-break1 to break phase. There is a marked reduction in the southerly component during the pre-break1 to break phase (Fig. 5c), over the south-east Arabian Sea off the Kerala coast as well as the south-east Bay. Significant weakening of the southerly component of the wind is also observed in the north-east Bay of Bengal, south and south-east Arabian Sea, and adjoining Indian coast. In the Madagascar Channel, a decrease in the southerlies is observed, but in the east Indian Ocean there is an increase in the southerlies. The weakening in the east equatorial Indian Ocean from pre-break2 to pre-break1, followed by the weakening in the Madagascar and north-east Arabian Sea and Bay of Bengal, may be due to a weakening of the cross-equatorial flow and its gradual movement north and north-westward. In the monsoon trough zone, extending from northwest India to the head of the Bay of Bengal, a striking increase in the meridional wind component is observed during the transition from prebreak1 to break (Fig. 5c). The magnitude of the changes over north-west India is observed to be larger than for the rest of the trough zone. The decrease in meridional wind in the Madagascar

Channel occurs from pre-break1 to break followed by an increase during the break to postbreak1 period (Fig. 5c and d). The decrease in the southerly component results in a decreased cross-equatorial flow and a weakening of the monsoon current. Important changes are found over the monsoon trough zone during the break to post-break1 period (Fig. 5d). The distribution of precipitable water content (kg m2 ) is shown in Fig. 6. Total precipitable water content of the atmospheric column indicates the availability of moisture. A higher moisture content can possibly provide higher precipitation under suitable conditions. Fig. 6a shows that the amount of precipitable water content is at a maximum over the head of the Bay of Bengal and the South China Sea during the break phase. There is a gradual decrease in precipitable water content by 1–2 kg m2 from the pre-break2 to pre-break1 period in the monsoon trough zone extending across Myanmar to the South China Sea. There is a significant increase in precipitable water content of 2–4 kg m2 in the south Indian Ocean between the equator to 15
S, and 75
E to 105
E (Fig. 6b). During the pre-break1 to break phase, the precipitable water content over the monsoon trough zone decreases rapidly by 3– 5 kg m2 . The decrease is also observed over the Indian landmass and the Bay of Bengal (Fig. 6c). Comparing Fig. 6b and c the region of decreased precipitable water content has shifted

The evolution of mean conditions of surface meteorological fields

143

Fig. 6. Geographical distributions of vertically integrated precipitable water content (kg m2 ) for (a) break and the difference of (b) pre break2–pre break1 (c) pre break1–break (d) break–post break1. Shaded regions are statistical significance at 95% confidence level

west-north westward with time. The change in the South China Sea from pre-break2 to prebreak1 and pre-break1 to break can serve as a precursor of the commencement of the break over the Indian region. An increase of 1–2 kg m2 is observed in the south Indian Ocean but the decrease in the south-east Indian Ocean close to Indonesia is more marked. The reduction in the total precipitable water content in the monsoon trough zone during the break phase of the monsoon could be due to reduced transport as indicated by the zonal and meridional wind (Figs. 4 and 5) resulting in a lower magnitude flux convergence of moisture over the Indian landmass and a reduction in rainfall. The South China Sea, monsoon trough zone and equatorial Indian Ocean between the equator to 15
S, and 75
E to 105
E show considerable changes from pre-break2 to pre-break1 and break, and can serve as a useful signal for the occurrence of the break epoch. The reversal in the pattern for the revival of the monsoon i.e., from break to post-break1, is rather slow and water content decreases over the north-eastern states of India, increases over the central Bay, and progresses in the monsoon trough zone (Fig. 6d). The distribution of total cloud amount (in octa) indicates that there is a higher amout over the north-east Bay of Bengal, adjoining north-eastern states of India and Myanmar during the break period of the summer monsoon (Fig. 7a). A sig-

nificant reduction in total cloudiness area by 0.4 octa first appears over the South China Sea and adjoining Indo-China peninsula during the prebreak2 to pre-break1 period. In the southern Indian Ocean, cloudiness increases between the equator to 15
S, and 75
E to 105
E (Fig. 7b). During the pre-break1 to break period, cloud cover increases over the South China Sea and east equatorial Indian Ocean, and reduces significantly over the Indian landmass and adjoining western Bay of Bengal. Thus, changes in the South China Sea in pre-break2 to pre-break1 indicate the likely initiation of the break monsoon situation over the Indian region. The reduction is more prominent over the monsoon trough zone and north-western parts of the country where a remarkable reduction in rainfall is observed during the break spell in the summer monsoon. In the southern Indian Ocean, the region between 15
S to 30
S, and 60
E to 100
E shows a gradual increase in cloudiness from pre-break1 to break. In the post-break period (Fig. 7d), the process is nearly reversed. During the cessation of the break, and proceeding from break to postbreak1, the increase in cloudiness beguns over the southern tip of India and the equatorial region, and proceeds northwards where there is a decrease in cloudiness over north-east India. These results agree with earlier studies which suggested that the active=break monsoons characterise a northward propagation of the maximum cloud zone

144

P. V. S. Raju et al.

Fig. 7. Geographical distributions of total cloud amount (octa) for (a) break and the difference of (b) pre break2–pre break1 (c) pre break1– break (d) break–post break1. Shaded regions are statistical significance at 95% confidence level

from the equatorial region towards the continental land mass (Yasunari 1979; Sikka and Gadgil 1980). Outgoing longwave radiation (OLR) is inversely related to the total cloud amount, particularly the convective clouds that result in precipitation. Low OLR values correspond to emissions at higher levels and hence higher cloud tops. Therefore, large negative (positive) OLR anomalies are associated with large positive (neg-

ative) anomalies in rainfall over the region. The geographical distribution of OLR is presented in Fig. 8. During the break period (Fig. 8a), lower values of OLR are found over the north-east Bay of Bengal, north-eastern states of India, adjoining Mayanmar, South-East Asia and equatorial western Pacific indicating widespread convection over these regions (Krishnan et al. 2000). In consistent with other meteorological fields, there is an increase in OLR by 10–15 W m2 from the

Fig. 8. Geographical distributions of outgoing longwave radiation (W m2 ) for (a) break and the difference of (b) pre break2–pre break1 (c) pre break1–break (d) break–post break1. Shaded regions are statistical significance at 95% confidence level

The evolution of mean conditions of surface meteorological fields

pre-break2 to pre-break1 period in the latitudinal belt around 15 N, and extending from 90 E to 120 E indicating a decrease in convective activity and hence rainfall. The increase is significantly higher over the South China Sea (Fig. 8b). However, during the pre-break to break period there is an increase in OLR by 5–10 W m 2 over the monsoon trough region from north-west India to the eastern Bay of Bengal. To the south of this increased OLR region, there is a zone of decreased OLR oriented in an east-west direction centred on 5 N–10 N and extending from 90 E–120 E. At the same time, in the southern Indian Ocean, between 15 S to 30 S, and 60 E to 100 E, there is a decrease in OLR from prebreak1 to break. This decreased zone of OLR implies an increase in convective activity and hence rainfall. This also indicates some kind of trough-ridge system in OLR that moves in a eastsoutheast to west-northwest direction from the South China Sea and southern region to the monsoon trough zone (Krishnamurti and Ardanuy 1980). During post-break, the pattern of OLR is reversed i.e., there is a decrease of OLR in the monsoon trough zone and an increase in the southern Indian Ocean (Fig. 8d). The OLR pattern closely follows the total cloud amount pattern from pre break2 to post-break1 (Fig. 7). Again, remarkable changes are noticed over the South China Sea during pre-break2 to pre-

145

break1, with some indication of the initiation of the break situation over the Indian landmass. 3.2 Variations of the surface energy fluxes About 99.9% of energy in the Earth’s atmosphere orginates from radiation from the Sun, which is in the form of the shortwave radiation. A useful way of considering the transfer of shortwave radiation through the atmosphere is to assess the magnitude of losses due to reflection and the gains as a result of absorption both within the atmosphere and at the Earth’s surface. In this section, we discussed various net heat budget components at the surface and associated changes during the break spell in the summer monsoon. The distribution of incoming short wave flux (SWF) at the surface is shown in Fig. 9. SWF increases from south to north in the Northern Hemisphere with a maximum over the Arabian Peninsula during the break phase of the monsoon (Fig. 9a). The Arabian Peninsula (mostly desert) and north Arabian Sea region are usually cloud free. There is an increase in the net shortwave radiation flux from pre-break2 to pre-break1 over the South China Sea, the head of the Bay of Bengal and the southern Indian Ocean between 15 S to 30 S, and 60 E to 90 E (Fig. 9b). However, from pre-break1 to break an increased area of SWF is found over the monsoon trough

Fig. 9. Geographical distributions of net shortwave radiation flux (W m 2 ) for (a) break and the difference of (b) pre break2–pre break1 (c) pre break1–break (d) break–post break1. Shaded regions are statistical significance at 95% confidence level

146

P. V. S. Raju et al.

Fig. 10. Geographical distributions of latent heat flux (W m2 ) for (a) break and the difference of (b) pre break2–pre break1 (c) pre break1– break (d) break-post break1. Shaded regions are statistical significance at 95% confidence level

zone and the Bay of Bengal region, the Indian landmass and the Arabian Sea (Fig. 9c), and a decrease in SWF is founded over the South China Sea. In the southern Indian Ocean the area of decreased SWF shifts towards the east and covers the region between the equator to 15
S, and 90
E to 140
E. These changes in SWF can be brought about by changes in the magnitude of cloud cover. The reduction in incoming SWF is due to the increase in cloudiness. The process is reversed during the break to post-break1 phase (Fig. 9d). There is a decrease in SWF over the central Bay and adjoining Indo-China. The changes in SWF are in good agreement with the total cloud amount as shown in Fig. 7. Changes in the South China Sea, monsoon trough zone and equatorial southern Indian Ocean are statistically significant at the 95% confidence level. Latent heat flux (LHF) from the surface of the Earth is due to evaporation. The geographical distribution of LHF is depicted in Fig. 10. During the break, LHF is found to be higher over the south-west Bay of Bengal and southern equatorial Indian Ocean. Proceeding from pre-break2 to pre-break1, there is a decrease in latent heat flux by 20–30 W m2 over the South China Sea, Bay of Bengal and eastern Arabian Sea. There is also a decrease in the southern Indian Ocean by 30 W m2 between the equator to 15
S, and 90
E to 105
E (Fig. 10b). From pre-break1 to

break (Fig. 10c), it further decreases by 20– 30 W m2 . The reduction is more marked in the Arabian Sea and the Bay of Bengal. However, the reduction is prominent over the Arabian Sea. The reduction was at its maximum over the South China Sea and southern Indian Ocean during the pre-break2 to pre-break1 phase, and it shifted westwards into the western Bay of Bengal and Arabian Sea during the pre-break1 to break phase. In the post-break period (Fig. 10d), there is an increase in the latent heat flux both over the Arabian Sea and Bay of Bengal, particularly over the southern parts and adjoining northern equatorial Indian Ocean. In the Indonesian region and southern Indian Ocean, there is an increase in LHF during pre-break1 to break (Fig. 10c) Which reverses during break to postbreak1 (Fig. 10d). The reduction in latent heat flux during the process of commencement of the break is due mainly to the reduction in wind speed, as is evident from Figs. 4 and 5. The net heat flux (NHF) is calculated as the sum of incoming shortwave radiation flux, effective outgoing longwave radiation flux, sensible heat flux and latent heat flux. The distribution of NHF at the surface is illustrated in Fig. 11. During the break phase (Fig. 11a), NHF is at a minimum over the southern Indian Ocean and a maximum over the Arabian Peninsula. It is rather low over the southern Arabian Sea and the south-

The evolution of mean conditions of surface meteorological fields

147

Fig. 11. Geographical distributions of net heat flux (W m2 ) for (a) break and the difference of (b) pre break2–pre break1 (c) pre break1–break (d) break–post break1. Shaded regions are statistical significance at 95% confidence level

west Bay of Bengal, indicating that considerable evaporation takes place over the southern Indian Ocean and adjoining parts of the Bay of Bengal and Arabian Sea. During pre-break2 to prebreak1 (Fig. 11b), there is a net heat gain of 20–50 W m2 over the South China Sea, central Bay of Bengal, central Arabian Sea and southeastern Indian Ocean close to Indonesia. The increase in the net heat gain implies decreased cloudiness, particularly convective clouds, and decreased evaporation from the ocean surface due to weak surface winds. Considerable heat gain continues in the Bay of Bengal, Arabian Sea and the monsoon trough zone during prebreak1 to break, but the pattern of net heat flux is reversed in the South China Sea (Fig. 11c) i.e., heat loss over this region due to an increase in convective activity. The South China Sea shows major changes during pre-break2 to pre-break1. There is significant net heat gain from pre-break2 to pre-break1, but net heat loss is observed from pre-break1 to break in the South China Sea and south-eastern Indian Ocean, off Indonesia. This variation in NHF in the South China Sea and south-eastern Indian Ocean from pre-break2 to pre-break1 and pre-break1 to break can indicate the possible commencement of the break monsoon situation over India. During break to postbreak1, the process is reversed, as evident from the changes observed in the central Bay of Bengal and Arabian Sea (Fig. 11d).

4. Discussion and conclusions The variations in mean surface meteorological fields (basic fields and derived surface heat fluxes), total precipitable water content, cloud cover and OLR that lead to the break situation in the summer monsoon over India are analysed over India and the adjoining seas. The temporal variation in all the parameters considered in this study are consistent with are another, and provide an indication of the initiation of the break monsoon situation over India which is associated with the suppression of convective activity over the Indian landmass and adjoining Bay of Bengal and Arabian Sea. Furthermore, the present study demonstrates clearly the close association between the changes in surface fields with convective activity over the South China Sea, east equatorial Indian Ocean and the initiation of break=revival or active cycles. The results agree with the earlier findings of Krishnan et al. (2000) based on a comprehensive diagnostic analysis of daily satellite derived OLR and NCEP reanalysis wind data for 17 years. It is noted that approximately ten days before the initiation of the break monsoon over India, convective activity reduces over South China Sea and the east equatorial Indian Ocean. However, with the commencement of the break monsoon situation over the Indian landmass, an increase in convective activity occurs. These changes leading to the break sit-

148

P. V. S. Raju et al.

uation over the Indian landmass are associated with the trough-ridge system (Krishnamurti and Ardanuy 1980). Thus, it is summarised that the main regions of activity associated with the initiation of convection and the revival of the monsoon are namely the South China Sea, the east equatorial Indian Ocean, and the Indian landmass with adjoining coastal seas. An enhancement of convective activity is associated with an increase in precipitation, cloud cover, precipitable water and surface wind, and a decrease in mean sea level pressure, OLR and net oceanic heat gain. About 10 days prior to the commencement of the break situation in India, a sharp decrease in rainfall, zonal wind, precipitable water content, could cover, and an increase in surface pressure, OLR and net heat gain are observed over the South China Sea. All these variations are a clear indication of a decrease in convective activity over the South China Sea in the pre-break2 period. During the approach of the commencement of the break condition over India, rainfall activity increases over the South China Sea and east equatorial Indian ocean region as well as an increase in zonal wind, precipitable water, total cloud amount, and a decrease in mean sea level pressure, OLR and oceanic net heat loss. These variations support one another and clearly demonstrate an increase in convective activity over the South China Sea and eastern Indian Ocean. At the same time over the Indian landmass, the reverse variation in these fields is statistically significant, and indicate suppressed convective activity. During the reversal of the active monsoon, variation in all these meteorological fields indicates an increase in convective activity over the Indian landmass, and a decrease in convective activity over the South China Sea and east equatorial Indian Ocean. The analysis of mean surface meteorological parameters and surface heat fluxes during the active=break phase of the summer monsoon over India reveals the following broad conclusions. During the break phase in the summer monsoon, large amounts of precipitation are observed over the north-east Bay of Bengal, north-east India and adjoining Myanmar where low values of OLR are found. In addition, significantly higher LHF is observed during the break monsoon period over the south-east Indian Ocean and South

China Sea, and weak LHF over the Arabian Sea during the break phase. The pattern is reversed in NHF i.e., a lower magnitude over the south-east Indian Ocean and South China Sea, and a higher magnitude over the Arabian Sea. It is noted that large and statistically significant changes in the surface meteorological parameters and energy flux are observed over south and South-East Asia, South China Sea and equatorial Indian Ocean two pentads before the break. In the pre-break2 to pre-break1 phase, considerable changes are observed in the precipitation rate, zonal wind, meridional wind, precipitable water content, total cloud amount, pressure, LHF, OLR, SWF and NHF over the South China Sea and adjacent region extending to the north of the Bay of Bengal. This significant area extends westwards or west-northwestwards, becomes more pronounced and occupies the region of the monsoon trough zone, central India and adjoining coastal seas as the break approaches. In the east equatorial Indian Ocean, particularly close to the Indonesian region, the variation in surface fields extends from west-southwest to east-northeast and sometimes exhibits a reverse in sign from the pre-break2 to pre-break1 and to break. Approximately ten days prior to the initiation of the break situation in the monsoon over the Indian landmass, most of the surface meteorological fields and surface heat fluxes clearly indicate significant changes over the South China Sea and SouthEast Asia. A reduction in monsoon strength and in convective activity follows. The changes in surface meteorological fields from break to postbreak are faster than the for pre-break to break. Thus, forecasting the onset of the break monsoon may be comparatively easier than the cessation. Acknowledgements We wish to express sincere thanks to NCEP=NCAR, USA for providing the reanalysis datasets. We are also thankful to the anonymous reviewers for useful comments and suggestions. References Annamalai H, Slingo JM (2001) Active=break cycles: diagnosis of the intraseasonal variability of the Asian summer monsoon. Clim Dynam 18: 85–102 Annamalai H, Slingo JM, Sperber KR, Hodges K (1999) The mean evolution and variability of the Asian summer

The evolution of mean conditions of surface meteorological fields monsoon: comparison of ECMWF and NCEP-NCAR reanalysis. Mon Wea Rev 127: 1157–1186 Barros AP, Lang TG (2003) Monitoring the monsoon in the Himalayas: observations in Central Nepal, June 21. Mon Wea Rev 131(7): 1408–1427 Das PK (1983) IMO monograph on monsoon. Fifth IMO Lecture Series, World Meteorological Organization, Geneva, 155 pp De US, Mukhopadhyay RK (2002) Breaks in monsoon and related precursors. Mausam 53: 309–318 De US, Lele RR, Natu JC (1998) Breaks in southwest monsoon. Pre-Published Scientific Reaport No. 1998=3 Gadgil S, Joseph PV (2003) On breaks of Indian monsoon. Proc Indian Acad Sci (Earth Planet Sci) 112: 529–558 Joseph PV, Sijikumar S (2003) Intraseasonal bariability of the low level jetstream of the Asian summer monsoon. J Climate 17: 1449–1458 Kalnay E, et al (1996) The NCEP=NCAR 40 years re analysis project. Bull Amer Meteor Soc 77: 437–471 Krishnamurti TN, Ardanuy P (1980) The 10 to 20-day westward propagating mode and ‘‘Braeks in the Monsoons’’. Tellus 32: 15–26 Krishnan B, Zhang C, Sugi M (2000) Dynamics of breaks in the Indian summer monsoon. J Atmos Sci 37: 1354–1372 Mohanty UC, Mohan Kumar N (1990) A study of surface marine boundary layer fluxes over the Indian seas during different epochs Asian summer monsoon. Atmos Environ 24A: 823–828 Mohanty UC, Dube SK, Singh MP (1983) A study of heat and moisture budget over the Arabian Sea and their role in the onset and maintenance of summer monsoon. J Meteorol Soc Jpn 61: 208–221 Pisharoty PR (1965) Evaporation from the Arabian Sea and Indian south west monsoon. In: Pisharoty PR (ed) Proceedings of International Indian Ocean Expedition, pp 43–54

149

Raju PVS, Mohanty UC, Rao PLS, Bhatla R (2002) The Contrasting features of the Asian summer monsoon during surplus and deficit rainfall years over India. Int J Climatol 22(15): 1897–1914 Ramamurthy K (1969) Monsoon of India: some aspects of ‘‘break’’ in the Indian southwest monsoon during July and August. Forecasting Manual. Indian Met. Dept. Poona, India, No. 4 18.3 pp 1–57 Raman CRV, Rao YP (1981) Blocking highs over Asia and monsoon droughts over India. Nature 289: 221–223 Ramaswamy C (1962) Breaks in the India summer monsoon as a phenomenon of interaction between easterly and the subtropical westerly jet streams. Tellus 14: 337–349 Rao YP (1976) Southwest monsoon. Meteor. Monogr., No. 1, India Meteorological Department, pp 1–367 Rodwell MJ (1997) Breaks in the Asian monsoon: the influence of southern hemisphere weather system. J Atmos Sci 54: 2597–2611 Sikka DR, Gadgil S (1980) On the maximum cloud zone and the ITCZ over Indian longitudes during the southwest monsoon. Mon Wea Rev 108: 1840–1853 Sperber KR, Slingo JM, Annamalai H (2000) Predictability and the relationship between the subseasonal and interannual variability during the Asian summer monsoon. Quart J Roy Meteor Soc 126: 2545–2574 Unninayar MS, Murakami T (1978) Temporol variations in the northern hemispheric summer monsoon circulations. Indian J Meteor Hydrol Geophys 29: 170–186 Wang B, Webster P, Haiyan Teng (2005) Antecedents and self-induction of active-break south Asian monsoon unraveled by satellites. Geophys Res Lett 32(4): L04704, doi: 10.1029=2004GL020996 Yasunari T (1979) Cloudiness fluctuations associated with the Northern Hemisphere summer monsoon. J Meteorol Soc Jpn 57: 227–242