Dynamics of the Indian Summer Monsoon Climate
Summary and Keywords
Lifeline for about one-sixth of the world’s population in the subcontinent, the Indian summer monsoon (ISM) is an integral part of the annual cycle of the winds (reversal of winds with seasons), coupled with a strong annual cycle of precipitation (wet summer and dry winter). For over a century, high socioeconomic impacts of ISM rainfall (ISMR) in the region have driven scientists to attempt to predict the year-to-year variations of ISM rainfall. A remarkably stable phenomenon, making its appearance every year without fail, the ISM climate exhibits a rather small year-to-year variation (the standard deviation of the seasonal mean being 10% of the long-term mean), but it has proven to be an extremely challenging system to predict. Even the most skillful, sophisticated models are barely useful with skill significantly below the potential limit on predictability. Understanding what drives the mean ISM climate and its variability on different timescales is, therefore, critical to advancing skills in predicting the monsoon. A conceptual ISM model helps explain what maintains not only the mean ISM but also its variability on interannual and longer timescales.
The annual ISM precipitation cycle can be described as a manifestation of the seasonal migration of the intertropical convergence zone (ITCZ) or the zonally oriented cloud (rain) band characterized by a sudden “onset.” The other important feature of ISM is the deep overturning meridional (regional Hadley circulation) that is associated with it, driven primarily by the latent heat release associated with the ISM (ITCZ) precipitation. The dynamics of the monsoon climate, therefore, is an extension of the dynamics of the ITCZ. The classical land–sea surface temperature gradient model of ISM may explain the seasonal reversal of the surface winds, but it fails to explain the onset and the deep vertical structure of the ISM circulation. While the surface temperature over land cools after the onset, reversing the north–south surface temperature gradient and making it inadequate to sustain the monsoon after onset, it is the tropospheric temperature gradient that becomes positive at the time of onset and remains strongly positive thereafter, maintaining the monsoon. The change in sign of the tropospheric temperature (TT) gradient is dynamically responsible for a symmetric instability, leading to the onset and subsequent northward progression of the ITCZ. The unified ISM model in terms of the TT gradient provides a platform to understand the drivers of ISM variability by identifying processes that affect TT in the north and the south and influence the gradient.
The predictability of the seasonal mean ISM is limited by interactions of the annual cycle and higher frequency monsoon variability within the season. The monsoon intraseasonal oscillation (MISO) has a seminal role in influencing the seasonal mean and its interannual variability. While ISM climate on long timescales (e.g., multimillennium) largely follows the solar forcing, on shorter timescales the ISM variability is governed by the internal dynamics arising from ocean–atmosphere–land interactions, regional as well as remote, together with teleconnections with other climate modes. Also important is the role of anthropogenic forcing, such as the greenhouse gases and aerosols versus the natural multidecadal variability in the context of the recent six-decade long decreasing trend of ISM rainfall.
The Indian Monsoon System
The Indian summer monsoon (ISM) or the South Asian monsoon, the strongest monsoon system in the world, shapes the health and well-being of one-sixth of the world’s population. The vagaries of the ISM not only determine the water and food security of the region but also significantly control the overall economy (Gadgil & Gadgil, 2006). With the seasonal contrast of rainfall being the hallmark of all monsoons, the ISM is characterized by a very wet summer and dry winter; therefore, the ISM is often defined by June–September rainfall (ISMR). The ISMR is part of the annual cycle of rainfall over the region, with an abrupt onset around the beginning of June and a slow withdrawal in late September and early October. Being the primary source of water for the region and hence the driver for well-being and prosperity, the monsoon season and monsoon rainfall are deeply etched in the socioeconomic fabric and literature of the region (Fein & Stephens, 1987). While the arrival of the monsoon is a source of joy for one and all, farmers dread the delay in its “onset,” and an ISM “drought” brings headaches for everyone. Therefore, major shifts in the ISM have shaped the rise and fall of civilizations in the region. The abrupt weakening of the Indian summer monsoon around 4100 yr ago (Dixit et al., 2014) and sustained weak monsoons for about 200 years may have been responsible for the extinction of the urban Harappan civilization in the Indus Valley (Staubwasser et al., 2003). On shorter timescales, floods (strong ISM) and droughts (weak ISM) resulting from year-to-year variations of the relatively stable ISMR cause death, destruction, and misery to the people of the region. The great famine in India during 1877–1878 was a result of the monsoon failure in 1876 and severe large-scale drought in 1877. Similarly, back-to-back droughts in more recent times, during 1965 and 1966, caused serious food grain deficiencies in India. The crisis was so serious that the country had to import large amounts of food grains from other countries, which in turn led to a balance-of-payments economic crisis in the country.
In addition to being the deciding factor for water resources and food security over the South Asian region, the ISM has left an indelible imprint on global climate and general circulation. The quantum of ISM rainfall represents a large heat source in the atmosphere and influences global circulation, introducing a significant zonal asymmetry in the Northern Hemisphere during boreal summer and influencing the global climate. For example, the descending motion associated with the Rossby wave response of the monsoon heat source leads to the Mediterranean-type climate to the west of the monsoon region (Rodwell & Hoskins, 1996, 2001). Also, the wind response of the monsoon heat source to the east of the heat source leads to significant influence on the evolution of the El Niño and Southern Oscillation (ENSO) phenomenon (Wu & Kirtman, 2003) and thereby impacts global climate through ENSO teleconnections.
Therefore, skillful prediction of ISMR at least one season in advance not only is a crucial imperative for policymakers and water resource managers in the Indian monsoon region but is also required for improving seasonal prediction globally. In the wake of the 1877–1878 famine, Sir Henry Blanford, then Director General of Meteorology, India Meteorology Department (DGM, IMD), came under pressure to make such forecasts. After intense study, he e produced the first seasonal forecast of Indian monsoon rainfall in 1886 based on its relationship with snow cover over the Himalayas, a relationship he had discovered earlier (Blanford, 1884). While trying to build on Blanford’s efforts in forecasting the ISMR, Sir Gilbert Walker, appointed DGM IMD in 1904, explored the relationship between ISM rainfall and various global climate signals in an attempt to discover other predictors (Walker, 1923). In this process, he discovered the famous seesaw in pressure between the east and the west Pacific, known as the Southern Oscillation (Walker, 1928), with a strong relationship with monsoon rainfall. Despite intense efforts by Indian scientists and the global climate community to improve ISM rainfall predictions, the seasonal prediction of ISMRs has remained one of the most challenging problems in climate science and has made only slow progress (Rajeevan et al., 2012). This failure is attributed to the gap in our understanding of the dynamics of the monsoon climate. For example, what maintains the climatological mean monsoon, what drives its variability, and what limits its predictability? Our current knowledge of these subjects and the gaps in our understanding have been highlighted in a number of monographs and reviews during the past three decades (Fein & Stephens, 1987, Chang & Krishnamurti, 1987, Webster et al., 1998, Wang, 2006).
Without repeating this knowledge base, a new perspective on ISM as being driven not by the surface temperature contrast between land and ocean but by tropospheric temperature contrasts over land and ocean is provided here. Space restrictions limit this discussion to the dynamics of the climatological Indian monsoon climate and its short term variations while refraining from addressing questions relevant to the dynamics of very long-term variations. For example, on geological timescales, the origin of the contemporary Indian summer monsoon and its link with the uplift of the Himalayan mountain range and the Tibetan Plateau has been a subject of considerable debate and will not be addressed here. However, a good review of issues and uncertainties can be found in Molnar et al. (2010). The ISM, though a regional phenomenon, is an integral part of the global climate, and a framework is provided here for understanding what maintains the climatological mean monsoon and its variability.
The Annual Cycle and Year-to-Year Variations of the ISM
As the name “monsoon” (meaning season) suggests, historically monsoons have been associated with the seasonal reversal of winds (Ramage, 1971), and the Indian monsoon is characterized by the largest seasonal reversal of winds (Figure 1,a,b). Before the advent of the mechanized ships with engines, sailors depended on these winds for their travel, and Arabian, Greek, and Portuguese traders made effective use of the monsoon winds to do business in South Asia for several centuries (Tripati & Raut, 2006). While the seasonal reversal of winds is a trademark of the Indian monsoon, the change in rainfall with the seasons (Figure 1c, d) is much more important for the people of the region, for the agricultural output, water resources, and economy of the region depend crucially on it. Therefore, for the people of the region, monsoon means rainfall. As Figure 1 (a,b,c,d) shows, the seasonal changes in winds are closely linked to the seasonal changes in precipitation in the region as well as globally. The change in winds is a result of seasonal migration of the intertropical convergence zone (ITCZ), and the precipitation changes with the season are a result of the seasonal migration of the rain band associated with the ITCZ. As Figure 1 again shows, the ITCZ and associated rain band are coupled phenomena with global extent, with the seasonal migration of the rain band over the Indian monsoon region being the largest in the world. The two phenomena are coupled because the heating of the atmosphere by the latent heat release during precipitation drives winds, while the convergence of moisture driven by these winds drives the precipitation. Thus, as a phenomenon, the Indian monsoon could be considered a convectively coupled phenomenon whereby the wet summer contrasted to a dry winter is intimately linked with the southwesterlies in summer and northeasterlies in winter. Therefore, the dynamics of the Indian monsoon and its variability is intimately linked with the dynamics of the ITCZ and its variability (Held & Hou, 1980, Goswami et al., 1984, Neelin & Held, 1987, Chou & Neelin, 2003, Schneider et al., 2014).
The rain band (ITCZ) makes the largest seasonal excursion in the Indian monsoon region from its climatological winter location at around 10oS (Figure 1d) to about 25oN, while maintaining a secondary rain band between the equator and 10oS in boreal summer. With the ITCZ moving only up to 15 oN over the South China Sea and western Pacific, the main ITCZ appears tilted in the north-northwest direction during boreal summer from its location over the equator at 150oE, extending to appear at about 25oN over India at 80oE (Figure 1c). Elsewhere in the world, the maximum northward excursion made by the ITCZ is only to about 10oN (Figure 1c). These features of the seasonal migration of the ITCZ are illustrated in the annual cycle of climatological daily precipitation over Indian longitudes (70 o–90 oE) (Figure 2a) and western Pacific (120 o–140 oE) (Figure 2b).
One important feature of the southwesterly winds during the summer monsoon season is the low level westerly jet over the central Arabian Sea. The southwesterlies flow across south and central India and the north Bay of Bengal and turn into southeasterlies in the sub-Himalayan North India, creating a large-scale low-level cyclonic vorticity (the monsoon trough; see Figure 1a). The low-level cyclonic vorticity is accompanied by a huge anticyclone in the upper atmosphere in the Indian monsoon region known as the Tibetan Anticyclone (Figure 1e), leading to the largest easterlies anywhere in the world during northern summer over the north equatorial Indian Ocean (Figure 1e). The reversal of the winds between the lower and upper atmosphere is associated with a deep north–south overturning circulation known as the regional Hadley circulation (Figure 2c). Another important and characteristic feature of the Indian monsoon is its sudden onset around June 1 (with an interannual standard deviation of ~ 7 days), when the ITCZ makes a rapid northward migration from about 5 oN to about 15 oN within a period of about one week; an indication of this movement can be seen even in the climatological annual cycle of precipitation over the region (Figure 2a). The annual cycle of daily precipitation averaged over central India (20o–25 oN, 75 o–80 oE) further illustrates the suddenness of the onset (Figure 2d). The most striking aspect of the onset, however, is the sudden and phenomenal increase in the strength of the winds over the Arabian Sea associated with it. As illustrated in Figure 2d, the annual cycle of kinetic energy of winds at 850 hPa averaged over (10 o–15 oN, 55o–60 oE) for 2008, indicating an 8- to 10-fold increase in the kinetic energy within less than a week at the time of onset!
The spatial distribution of the seasonalmean rainfall is as important as the quantum of rainfall for agriculture planning and local water resource management. The climatological June–to September (JJAS) rainfall from CPC merged analysis of precipitation (CMAP) reveals the characteristic high-rainfall regions west of the Western Ghat mountain and in northeast India, while low rainfall over the rain-shadow region of south India and the gradual increase in aridity from east to west over central and north India lead to desert-like conditions in Rajasthan and Gujarat (Figure 1a). An integral part of this spatial distribution of Indian monsoon rainfall is also the high rainfall over the north Bay of Bengal and Irrawaddy Delta (Figure 1a). This is why Goswami et al. (1999) produced an extended Indian monsoon rainfall (EIMR) index, as JJAS precipitation averaged over 10o–30 oN, 70 o–100 oE. While this index is a good representation of the ISM heat source, the length of such an index is restricted by the availability of satellite data since 1979. Hence, most Indian monsoon variability studies use JJAS rainfall averaged over continental India (also referred to as all India rainfall or AIR) as an index of the Indian monsoon; the data can go as far back as 1871. Quality-controlled rain gauge observations from a large number of stations maintained by the India Meteorology Department (IMD) have led to several rainfall datasets over India for monsoon variability studies at monthly (Parthasarathy et al., 1994), and daily (Pai et al., 2014) timescales.
How variable is the seasonal mean monsoon climate? The coefficient of variability on the interannual timescale (Figure 3a), based on such long datasets, shows that regions of low climatological mean rainfall are regions with high interannual variability, whereas regions of high climatological mean rainfall are areas of low interannual variability, indicating high vulnerability for the arid regions. The Indian monsoon as represented by ISMR is a reasonably stable system, with the interannual standard deviation of the seasonal mean being 10% of its long-term mean. At the same time, however, the extremes of Indian monsoon rainfall as defined by deviations of ISMR from mean by more than one standard deviation (Figure 3b) have large socioeconomic impact on the region. “Flood” and “drought” years are those years that depart significantly from the long-term mean record of ISMR and have normalized values greater than +1 or less than −1, respectively. This definition of floods and droughts is used here to refer to large-scale (countrywide) floods or droughts. On any smaller scale, a definition of “drought” based entirely on rainfall may be restrictive as it may depend on temperature as well as subsurface water availability. The interannual variability of the Indian monsoon rainfall within this period appears to be roughly distributed normally, with drought extremes (26) slightly outnumbering flood extremes (21). The worst drought in India’s history took place in 1877, with ISMR recording a –29% departure from the long-term mean, while the worst flood occurred in 1961, with ISMR recording a +19% departure from the long-term mean. While the monsoon extremes represent strong signals and hence may be predictable, it only represents 30% of the total number of years. For the remaining 70% of the years, the Indian monsoon is “normal” and represents a weak signal, with limited predictability. This makes the prediction of ISMR a challenging problem. Therefore, it is important to understand not only what drives the “extremes” of the monsoon but also what drives the “normal” monsoons. Also, ISMR is a spatially averaged quantity useful mainly for policymakers to plan for the country’s food and water security. For the farmers and local water managers, however, ISMR is not very useful, and information on the spatial distribution of the seasonal mean rainfall is more important. The spatial distribution of seasonal mean anomalies rainfall from its climatology (Figure 4a) for a flood year (Fig.4b), a drought year (Figure 4c), and a normal year (Figure 4d) indicates that the distribution of seasonal mean anomalies is quite homogeneous during the extreme monsoon years (Figure 4b,c), while it is highly nonhomogeneous during “normal” years (Figure 4d). In other words, during a major drought (flood), the seasonal mean rainfall is below normal (above normal) over most of the country. However, during a “normal” year, pockets of droughts or floods could be distributed randomly over the country. Hence, a prediction of a “normal” ISMR is not very useful to the farmers and local water managers.
The correlations between the Indian monsoon rainfall and the Southern Oscillation (SO) discovered by Gilbert Walker (1928) appear to be the result of teleconnections between variations of the sea-surface temperatures (SSTs) over the eastern tropical Pacific, namely, the El Niño and the ISM. The El Niño and La Niña are opposite phases of SST variations associated with the SO (Philander, 1990). El Niño and Southern Oscillation (ENSO) influence the ISM through an atmospheric bridge. To facilitate that discussion, Figure 3b marks the El Niño and La Niña years.
An 11-year running mean of ISMR anomalies indicates that the year-to-year variations are modulated by a multidecadal variability (see the bold black line in Figure 3b), with approximately three decades of above-normal conditions followed by three decades of below-normal conditions. It is intriguing, however, that the recent years (after 1950) seem to be an exception, with the ISMR decreasing monotonically for more than five decades. In the backdrop of global warming and increased moisture content in the atmosphere, this “long” decreasing trend of the ISM rainfall is rather counterintuitive.
Subseasonal Composition of the ISM Climate
In order to understand the dynamics of the Indian summer monsoon climate as represented by JJAS rainfall, it is essential to know what constitutes the ISM climate. The ISM climate is the result of precipitation from synoptic scale disturbances such as the lows and depressions (Sikka, 1977), as well as the amplitude and frequency of occurrence of subseasonal rain spells. For space reasons, the genesis or nature of the synoptic disturbances and their contribution to the seasonal mean is not discussed in detail, but the subseasonal oscillations are described briefly as they represent a very large signal and influence the seasonal mean significantly. The daily rainfall anywhere in central India (Figure 5a) during the summer monsoon season occurs in spells of “active” (above-normal rainfall) conditions interspersed with “breaks” of much drier conditions. Spectral analysis of such a daily rainfall time series indicates that these spells result from two modes of oscillations in the subseasonal timescales: a 10- to 20-day period (Krishnamurti & Bhalme, 1976, Chen & Chen, 1993) and a 30- to 60-day period (Yasunari, 1979, Sikka & Gadgil, 1980, Figure 5b). Together, the subseasonal variations represent a very large signal, with an amplitude as large as the annual cycle (Figure 5c). The 10- to 20-day mode, also known as the quasi-biweekly mode, has a spatial structure of an equatorial Rossby wave, with a horizontal scale of approximately 6000 kilometers driven by a convective feedback (Chatterjee & Goswami, 2004) that propagates westward with no significant northward propagation. The spatial structure of the 30- to 60-day mode is much larger and is characterized by a repeated northward progression of the anomalous ITCZ (Figure 6). This mode has been shown to be driven by a radiative–convective–circulation feedback (Goswami & Shukla, 1984), while the northward propagation has been shown to result from feedback between the anomalous circulation generated by the anomalous convective heating (Jiang et al., 2004).
The fact that the spatial structure (Figure 6) of the dominant monsoon intraseasonal oscillation (MISO) is similar to that of the seasonal mean (Figure 1a) indicates that, during an active spell, precipitation over continental India enhances and strengthens the large-scale monsoon circulation, while during a break phase both precipitation over India and the large-scale monsoon circulation weaken. Therefore, depending on the amplitude and frequency of occurrence of the active/break spells during a season, the seasonal mean monsoon itself could be stronger or weaker. Thus, the MISO statistics could lead to a significant component of “internally” driven interannual variability of the seasonal mean monsoon (Goswami & Xavier, 2005). As this component of the interannual variability of the seasonal mean monsoon is not predictable one season in advance, it limits the predictability of the seasonal mean monsoon (Goswami et al., 2006). By virtue of the significant modulation of the large-scale circulation and thermodynamic conditions by the active and break spells, the MISO also makes the genesis and intensification of synoptic systems (lows and depressions) much more conducive during the active phases than during the break phases. As a result, more than four times as many low-pressure systems (LPS, lows and depressions) are found to occur during active spells compared to break spells (Goswami et al., 2003). The modulation of the cloudiness and surface winds by the MISO leads to large changes in the net heat flux to the ocean, influencing the sea-surface temperature (SST; Sengupta et al., 2001), which in turn influences convective activity and precipitation on this timescale. While the MISOs may be initiated by atmospheric dynamic and thermodynamic processes, the observed MISOs are a result of modulation of the amplitude, frequency, and northward propagation characteristic by the air–sea interaction associated with them. More details about the subseasonal variability of Indian summer monsoons can be found in Goswami (2012).
For farmers and water managers, advance knowledge of active and break spells is very important and useful for agricultural practices (e.g., sowing and harvesting) and optimum use of water resources. How predictable are these spells? The large amplitude of the phenomenon and its coupling with the ocean indicate that the MISO should be predictable at an extended range (beyond the range of weather predictability). In fact, predictability studies using observations and models indicate a 25- to 30-day limit on the predictability of the MISO (Goswami & Xavier, 2003; Waliser et al., 2003). To facilitate real-time monitoring and for forecast verification of the MISO, a low-order description of MISO was recently proposed (Suhas et al., 2013) in terms of two indices called MISO1 and MISO2. Coupled ocean–atmosphere climate models required to simulate the MISO have improved over the years, and some models are now able to simulate the MISO with acceptable fidelity (Sharmila et al., 2013). Use of such a model demonstrates that MISO could be predicted with useful skill at a lead time of 15 to 20 days (Abhilash et al., 2014a, 2014b). These forecasts have already proved useful for farmers and water managers. With such extended range forecasts of active/break spells a reality, in addition to the seasonal forecast of ISMR, updating these forecasts of MISO every few days (say five days) has become an excellent option not only for policymakers but also for farmers and water managers.
A Mechanistic Model of the Indian Summer Monsoon
To build any model of the Indian monsoon, it is necessary to recognize the unique geographical setting in which the Indian monsoon occurs, with the Indian continent sandwiched between the tall and sharply rising Himalayan Mountains to the north, together with the warm Indian Ocean to the south. A model of the monsoon must not only explain the seasonal reversal of the winds and precipitation but also the deep vertical structure of the summer monsoon circulation. The classical model of the Indian monsoon does not involve precipitation and explains the seasonal reversal of winds due to the land–ocean surface temperature gradient arising because of the different heat capacity of land and water. In boreal summer, land heats up faster than the upper layer of ocean, creating a low pressure over land relative to a cooler ocean to the south, and winds converging to the low pressure produce the cross-equatorial flow and southwesterlies. In the beginning of the summer monsoon, May–June SST and the land-surface temperature (Figure 7a) indeed show that the land-surface temperatures over the Indian subcontinent are much warmer than the SST over the ocean to the south, consistent with the above picture. The opposite is true during boreal winter when the land-surface temperature over India is much cooler than the SST over the ocean to the south (Figure 7c), with northeasterlies over India and the Arabian Sea converging toward a low pressure over the warmer ocean. However, during the summer monsoon (July–August), the land-surface temperatures are cooler than the SST in the south (Figure 7b). What, then, maintains the strong southwesterlies at surface as well as at low levels during the monsoon? While the north–south surface temperature gradient theory could explain the onset of the Indian monsoon, it is inadequate to explain the sustenance of the monsoon during the whole season. In the absence of precipitation, if the surface heating gradient is the main driver of the Indian monsoon, it will lead to a north–south overturning circulation limited only to the boundary layer in vertical extent, as such heating would extend only to the top of the boundary layer. As a result, such a theory would also fail to explain the deep vertical structure of the Indian monsoon.
The surface winds could be driven not only by surface pressure gradients (SST gradients; Lindzen & Nigam, 1987) but also by deep atmospheric heating (Gill, 1980). The surface winds and the winds above the boundary layer during the monsoon season appear to be driven by a deep heat source where the monsoon precipitation is an integral component of driving and sustaining the Indian monsoon after onset. A hint of what drives the Indian monsoon can be seen when we note that, while the surface temperature gradient reverses after the onset, the tropospheric temperature (TT; averaged over 600 hPa and 200 hPa) during boreal summer (Figure 8a) is much larger over the continent than that over the ocean to the south, a situation that reverses itself in winter (Figure 8b). Indeed, the anomalously large TT over the Asian monsoon region during boreal summer indicates a large tropospheric, deep heat source (and associated pressure gradient) that sustains the Indian monsoon. What is responsible for maintaining this anomalously large TT over the large region during summer? What is the role of the Himalayan Mountains and the Tibetan Plateau (Figure 7d) in maintaining this TT? The spatial extent of the TT anomaly is much larger than that of the summer precipitation associated with the ISM (Figure 1c), indicating that the elevated heat source caused by surface heating of the Tibetan Plateau and Himalayan Mountains plays an important role in producing this TT anomaly. The onset of ISM is induced by higher TT over northern India, driven primarily by dry thermodynamics such as an orographically elevated heat source and a persistent heat anomaly arising from planetary scale stationary waves. After the onset, however, the ISM precipitation adds to this TT and helps maintain the ISM.
The evolution of the climatological mean daily TT averaged over (60–100oE) as a function of latitude during the year (Figure 9a) indicates something rather interesting. If we draw a line around 5 oN, it is noted that it transforms from maximum south of the line to maximum north of it around June 1, the climatological onset date for Indian monsoon! This is coincident with the sudden northward jump of the rain belt within less than a week around the same time (Figure 2a), an onset of ISM as defined from precipitation. Thus, the change in the sign of the north–south gradient of TT over the region seems to drive the onset of the Indian monsoon. As TT is related to the deep heat source, it makes good sense to use the change of sign of the TT gradient for defining onset as well as “withdrawal” of monsoon as it relates to winds turning southwesterly (or northeasterly) near the surface as well as above the boundary layer. To make it quantitative, we define two boxes as shown in Figure 8a, one representing the continental heat source between (5o–35 oN, 40 o–100oE) and another representing the oceanic heat source between (15oS–5oN, 40 o–100oE). The north–south TT gradient is represented by differences between TT averaged over the north box and TT averaged over the south box (ΔTT). It may be noted from the daily climatological ΔTT (Figure 9b) that the gradient changes sign from negative to positive around June 1, ushering in the onset of monsoon, and it goes back to negative around October 1, marking complete “withdrawal” of monsoon from the country. Therefore, the TT gradient provides an objective way of defining the onset as well as the “withdrawal” of the Indian monsoon and thereby a new definition of the length of rainy season (LRS). The onset defined as such is well correlated with the traditional definition of monsoon onset over Kerala (MoK), as used by the India Meteorology Department (Xavier et al., 2007). How does the change of sign of the TT gradient lead to the onset?
As the northern summer progresses, the Indian landmass heats up, compared to the ocean to the south. The leading surface temperature gradient leads to cross-equatorial flow, and ensuing southwesterlies near the surface bring in warm, moist air to the continent. By the second half of May, the land-surface temperature as well as the low-level moisture is very high over the continent, but there is no organized precipitation in sight, with the ITCZ still stuck around 5 oN. This is due to an inhibition (for convection) introduced by a large-scale subsidence over the continent, caused by persisting precipitation and convective activity over the maritime continent. Therefore, even though there are convergent southwesterlies at the surface, it is overlaid by a divergent southward flow above the boundary layer. By late May, even though the convective available potential energy (CAPE) is very high, it is unable to produce large-scale deep convection in the north without a large-scale trigger to overcome the inhibition. Such a trigger could come only from a large-scale dynamic instability. It has been shown that it comes in the form of a “moist symmetric inertial instability” (Tomas & Webster, 1997; Krishnakumar & Lau, 1998). Strong zonally elongated vertical velocity associated with such an instability overcomes the inhibition and facilitates explosive development of the ITCZ around 10 oN at the time of onset. The northward appearance of the ITCZ also increases the vertical easterly shear of winds and facilitates the rapid northward propagation of the ITCZ and progression of the Indian monsoon following the onset. For the “symmetric inertial instability” to occur, the zero of the absolute vorticity at 850 hPa needs to be at north of the equator (Tomas & Webster, 1997). This can happen only if the large-scale circulation at 850 hPa over the region in the northern hemisphere is cyclonic. As the TT gradient becomes positive, the heat source is positive to the north, and convergent circulation at 850 hPa produces a large-scale cyclonic vorticity to the north, creating an environment conducive for the moist symmetric instability to occur. Therefore, the change of the sign of the TT gradient is a fundamental driver of the onset through facilitating the trigger of the symmetric instability. Thus, the TT gradient provides a comprehensive model of monsoon, explaining not only the sustenance of monsoon after onset but also the onset itself. It also provides a framework to understand the dynamics of Indian monsoon variations at interannual and longer timescales by examining and quantifying how different drivers influence the TT gradient by influencing the TT in the north or in the south, or in both.
The Himalayan Mountains and the Tibetan Plateau are critical to the existence of the Indian monsoon (Molnar et al., 2010). For one thing, the exceptional extent of the ITCZ’s northward migration in the Indian monsoon region would be impossible without this orography. Some debate has arisen over whether the “mechanical” effect of this mountain range in preventing the incursion of cold and dry air weakening convective activity is more important than its “thermodynamic” effect in creating an elevated heat source, influencing the TT, and thereby driving the monsoon’s strength (Boos & Kunag, 2010). However, the change in the sign of the TT gradient is entirely due to the thermodynamic effect of the mountain range, as the convective heating does not start contributing to the heat source until after onset. In the absence of such mountains, the onset is expected to be delayed by more than a month (Chakraborty et al., 2002), thereby resulting in a much shorter LRS and a weakened ISM. Therefore, we consider the thermodynamic effect of the Himalayan Mountains and Tibetan Plateau to be as important, if not more so, than its mechanical effect in maintaining the ISM’s strength.
Ocean–Atmosphere–Land Interactions, Teleconnections, and ISM Variability
The strength of the ISM is intimately related to the amplitude of the characteristically large annual cycle, and as is the case for the global climate, the annual cycle is driven primarily by the solar forcing. Therefore, on very long timescales (e.g., orbital periods), ISM variations are known to follow solar forcing (Cheng et al., 2016). Climate variability on relatively shorter timescales (e.g., interannual to centennial), however, could arise from internal processes within the interacting ocean–land–atmosphere–cryosphere climate system. The focal point of these interactions and feedbacks within the climate subsystems could be regional, leading to regional climate oscillations such as the El Niño and Southern Oscillation (ENSO) arising from local ocean–atmosphere interactions over the tropical Pacific (Philander, 1990; Sarachik & Cane, 2010; Wang & Picaut, 2004).
Climate variability in a particular region (e.g., the ISM region) could also arise through “teleconnections” emanating from regional fluctuations of climate in faraway regions (possibly arising from local air–sea interactions). These teleconnections can arise either through atmospheric global circulations (the atmospheric bridge) or through oceanic heat transport via oceanic thermohaline circulation. The atmospheric teleconnections can be understood in terms of atmospheric overturning circulations, namely, the east–west overturning circulation around the equatorial-vertical plane called the Walker circulation (Bjerknes, 1969) and a north–south overturning circulation known as the Hadley circulation and their interactions. The ascending branches of these overturning circulations are always associated with the heating arising from the high-precipitation centers. These heat sources could also influence the climate of faraway places through the planetary scale wave trains which they produce in the atmosphere. Hence, regional climate variability such as that of the ISM arises partly from such local ocean–atmosphere–land interactions and partly from teleconnections from other drivers. Therefore, to understand the dynamics of ISM climate variability, it is necessary to identify and quantify contributions from such local interactions and feedbacks and from other potential teleconnections.
As mentioned earlier, the ISM can be seen as a manifestation of the northward seasonal migration of the ITCZ. The ITCZ movement is governed by the annual SST cycle in this region. The observed annual cycle of SST in the region, in turn, is maintained as a large-scale air–sea interaction involving the monsoon winds. During the boreal summer, the northern Indian Ocean receives heat from the atmosphere, while the southern Indian Ocean loses heat to the atmosphere, with the situation reversing itself during the boreal winter (Figure 10a,b). Maintaining the observed annual SST cycle requires a shallow meridional circulation (Schott et al., 2002) to transport heat south across the equator during summer and north during winter. Models indicate that the Ekman transport forced by monsoon winds could greatly facilitate this cross-equatorial, seasonally reversing meridional transport (Figure 10a,b). Note that the winds south of 10°S do not change direction with the seasons. While the southward transport in summer may extend to the deep southern tropics (south of 20°S), the northward transport in winter is expected to be confined to the region north of 10°S. This indicates that the seasonal cycle of the South Asian monsoon is essentially maintained by large-scale ocean–atmosphere interaction (Webster et al., 1998; Loschnigg & Webster, 2000; Webster, 2006). Accordingly, it is easy to see that air–sea interaction could lead to interannual variations in the South Asian monsoon. For example, weak winds in a weak monsoon year would lead to weaker southward transport of heat, leaving the northern Bay of Bengal with higher SSTs and heat content during the following summer. This pattern then leads to vigorous synoptic disturbances (atmospheric lows and depressions) and stronger monsoon winds the following year. The stronger southward transport of upper ocean heat in a strong monsoon year leads to a depletion of heat in the north Indian Ocean and cooler SST during the following year, preparing the condition for a weak monsoon. Ocean reanalysis indicates significant differences in mean meridional circulation during the northern summer between strong and weak monsoon years (Figure 10c,d). This air–sea interaction process could lead to a biennial oscillation of the South Asian monsoon (Loschnigg et al., 2003).
In order to reliably predict the seasonal mean ISM rainfall (the AIR), it is important to understand what drives the year-to-year variations of the ISM rainfall (Figure 3b). Empirical studies over the years have attempted to identify many global and regional climate parameters associated with the ISM, with the hope of finding such drivers. The strongest such signal has been found in the association with the ENSO (Pant & Parthasarathy, 1981, Rasmusson & Carpenter, 1983). A sliding correlation between AIR and an ENSO index (here we have taken Niño3.4 SST) with a 21-year sliding window indicates (Figure 11a) that the relatively strong correlation (CC ~ –0.65) found in the past between the two has lessened significantly during the last two to three decades. This decrease is evident in Figure 3, where we have marked the El Niño years as well as the La Niña years. It is clear that El Niño years tend to be associated with below-normal monsoon rainfall, with many El Niño years coinciding with monsoon droughts, while no El Niño years are associated with a flood. In contrast, La Niña years tend to be associated with above-normal rainfall, and no La Niña year is known to be associated with a drought. This is consistent with the moderate but not very strong correlation between the two. However, many monsoon droughts (as well as floods) have no relation to ENSO. Thus, many ISM extremes could occur without being driven by ENSO (ISM “internal” dynamics). Even at its peak, the ENSO explains only about 40% of ISM interannual variability, leaving a significant component of ISM variability as yet not identified with any predictable, slowly varying forcing.
Why the ENSO–monsoon relationship weakened in recent years has been a subject of debate for some time. It has been discovered that El Niño itself occurs in different shades (e.g., the central Pacific [CP] El Niño, or the eastern Pacific [EP] El Niño), and only a certain shade of El Niño (the CP El Niño) is more effective in influencing the Indian monsoon rainfall (Krishnakumar et al., 2006), with the frequency of occurrence of different shades of El Niño being responsible for the weakening relationship. This question is not well settled, and contrary views are also available. Another study (Xavier et al., 2007) claims that the “real” ENSO–monsoon relationship has actually not decreased and the weakening correlation between AIR and ENSO is a manifestation of increasing “climate noise” in AIR owing to the increase in extreme rain events in recent years.
What is the teleconnection through which the ENSO SST in the central and eastern equatorial Pacific SST weakens the ISM? Movement of the precipitation zone and associated atmospheric heating from its climatological western Pacific position to the central-eastern Pacific during an El Niño results in an eastward shift of the Walker circulation. The Walker circulation anomaly leads to increased low-level convergence and increased deep convection in the equatorial Indian Ocean, forcing decreased convection and rainfall over continental India. The amount of shift of the Walker circulation depends on the shade of El Niño (whether a CP El Niño or an EP El Niño), thereby modifying the influence on equatorial Indian Ocean convection and further on monsoon rainfall. A thermodynamic interpretation of this teleconnection was found recently (Goswami & Xavier, 2005, Xavier et al., 2007) when it was discovered that ENSO could influence ISM rainfall by influencing the TT gradient in the Asian monsoon region. The heat source associated with El Niño or La Niña also results in stationary Rossby waves in the subtropics and leads to a persistent negative (positive) TT anomaly over northern India and southern Eurasia (the northern box as marked in Figure 8a), weakening (strengthening) the TT gradient over the Indian monsoon region and thereby the ISM rainfall. This also leads to a delayed onset and early withdrawal of ISM in El Niño years and early onset and delayed withdrawal in La Niña years. The evolution of composite ΔTT during El Niño and La Niña years (Figure 11b) indicates that this process leads to a shortening (lengthening) of LRS during the El Niño (La Niña) years. The LRS is strongly correlated with ENSO, and one mechanism through which ENSO seems to control ISM rainfall is by controlling LRS. As shown in Figure 11b, in addition to shortening (lengthening) LRS, El Niño (La Niña) also suppresses (elevates) the annual cycle of ΔTT, adding further to the weakening (strengthening) of ISM rainfall. When both of these effects are combined, a new thermodynamic index of summer monsoon (TISM) is defined as the integral of positive ΔTT in a year. The strong correlation between TISM and LRS rainfall (0.75) indicates the wisdom of using TISM as an index of ISM rainfall. Unlike the correlation between AIR and Niño3.4 (Figure 11a), the correlation between TISM and Niño3.4 has remained quite high in recent years, without showing any major decrease (Figure 8 of Xavier et al., 2007). While any index of seasonal mean rainfall (e.g., AIR) could be influenced by the small-scale rainfall extremes, TISM is not affected by such small-scale high-frequency events and relates primarily to the predictable low-frequency component of ISM rainfall.
The possible role of the Indian Ocean Dipole Mode (IODM; Saji et al., 1999; Webster et al., 1999) in the ISM has also been a subject of considerable interest. A strong ISM with strong cross-equatorial winds (and strong longshore winds along the coasts of Borneo) could initiate a positive IODM that strengthens through local air–sea interactions, peaking in October (Saji et al., 1999; Li et al., 2003). A positive correlation between IODM and precipitation over India (Saji & Yamagata, 2003; Ashok et al., 2004) could therefore be simply related to the fact that the ISM plays a significant role in driving the IODM. Warmer SST in the Arabian Sea during a positive IODM could, in principle, lead to a larger moisture flux into the Indian region and a strengthening of the ISM. As significant warming of AS associated with the IODM takes place toward the end or after the ISM season, a strong influence of IODM-related SST on the Indian summer monsoon is unlikely. If during some epoch, however, the seasonality of the IODM changes in such a way as to peak within the monsoon season (JJAS), a much stronger positive impact on the ISM rainfall could be expected.
Even before Walker’s findings of ENSO as a major driver of the interannual variability of ISM, Banford’s discovery of the Himalayan snow cover and monsoon relationship (Blanford, 1884) had opened the possibility that Eurasian snow cover might be a major driver of ISM interannual variability. Walker started using Himalayan snow cover as one of the predictors of ISM rainfall (Walker, 1910), a practice that was continued by IMD for many years. A negative relationship between Eurasian snow cover and ISM intensity found in some early studies (Han & Shukla, 1976; Bamzai & Shukla, 1999) based on limited observations was, however, not replicated in later studies (e.g., Robock et al., 2003) with longer, more reliable observations. This is attributed to the fact that the snow cover variability over Eurasia is complex and that snow cover over different parts of Eurasia relate to ISM rainfall differently (Kriplani & Kulkarni, 1999). ENSO influence on the snow cover over Eurasia also complicates this relationship (Fasullo, 2004). Through the radiative effects, the Eurasian snow cover can influence both the TT temperature in northern India and southern Eurasia (the north box) and ISM rainfall by affecting the ΔTT. Eurasian snow cover far away from this box, therefore, would have weak influence on the TT over this box, explaining some of the inconsistency in the relationship found in some early studies. Blanford’s original contention regarding the influence of snow cover over the Himalayas and Tibet, however, is still valid when data are stratified from the ENSO effect (Fasullo, 2004); complications arise only when the area is expanded to a much larger Eurasia. Thus, Eurasian snow cover appears to be a rather minor driver for ISM rainfall variability. However, the precise fraction of the ISM interannual variability that is explained by the Eurasian snow cover remains an open question.
Predictability of the ISM climate
Weather predictability is governed by the chaotic nature of the flow involving nonlinearity and instability and is limited by the amplitude of the variability (the signal) and the growth rate of small errors. As climate is defined in terms of space and time averages (e.g., AIR representing ISM climate), it is not governed by the flow instability and nonlinearity but by external forcing (e.g., solar forcing) and dissipation. Therefore, the predictability of climate depends not so much on the initial conditions but on the nature of the variability (Charney & Shukla, 1980, Shukla, 1981)). Like the tides, if the climate were perfectly periodic, it would be infinitely predictable. Unfortunately, the ocean–atmosphere–land interactions give rise to “internal” climate variability at a number of time and space scales that interact with the annual cycle driven by solar forcing; among themselves, they lead to a broadening of the spectrum around each dominant frequency. While the periodic component is predictable, the limit on the predictability of any climate signal (e.g., ENSO, Pacific Decadal Oscillation [PDO], the Atlantic Multidecadal Oscillation [AMO]) depends on the amplitude of both the dominant periodicity and the unpredictable “climate noise” arising from higher frequency, less predictable signals or oscillations. The width of the spectrum around that periodicity may be considered a measure of the amplitude of the “climate noise.” ENSO is a large-amplitude, short-term climate oscillation arising from coupled ocean–atmosphere interactions and provides the basis for enhanced predictability of tropical climate (Charney & Shukla, 1980; Shukla, 1981). For example, the theoretical limit on the predictability of the ENSO phenomenon as estimated from a coupled ocean–atmosphere model is about 3 years (Goswami & Shukla, 1991). This is consistent with the broadband nature of the ENSO spectrum with its dominant periodicity varying from 2 to 7 years.
The seasonal mean ISM rainfall (AIR) does not have any strong short-term periodic signal. It does have a biennial tendency, a linkage with ENSO, a weak 11-year periodicity, and a 50- to 80-year multidecadal oscillation (Goswami et al., 2015). Therefore, on an interannual timescale, the linkage with ENSO provides the most dominant predictable signal for ISM. However, as mentioned earlier, ENSO explains only about 40% of the ISM variability. What is the limit on the ISM’s predictability? An estimate of the “climate noise” or “internal” variability is required to estimate this limit. The ISM variability from a long simulation of an atmospheric climate model forced at the lower boundary only by repeating climatological SST provides such an estimate of “internal” variability (Goswami, 1998; Goswami & Xavier, 2005). Spread of simulation of AIR in an ensemble of simulations by a climate model forced with observed SST could also provide such an estimate of “internal” variability, and comparison with the interannual variation of the ensemble mean AIR could provide an estimated potential predictability (e.g., Kang & Shukla, 2006). These studies indicate that about 50% of interannual ISM variability is predictable. What does it mean in terms of the potentially achievable skill of seasonally predicting ISM rainfall? Correlations between ensembles of predictions by a model with its own simulations of ISM rainfall may provide such a skill by the model, while a correlation between predictions and actual observations indicates the model’s actual skill. Such potential skill obtained from a large number of predictions by a number of models is currently approximately 0.7, while the actual skill is approximately 0.4 (Rajeevan et al., 2012). The potentially achievable skill from these studies is consistent with the early estimate of potential limit on monsoon predictability being 50% of ISM variability. It also shows a significant gap in the current skill of models and potentially achievable skill. As these estimates are based on climate models that still have some biases in simulating the mean climate, these estimates may change with improvements of the climate models. IA significant improvement of the skill of seasonal prediction of ISM rainfall has been noted in one of the latest coupled climate models (Ramu et al., 2016). Although these are highly encouraging developments, dynamical prediction of the seasonal mean monsoon is still in its early days and continued work on improving models and coupled data assimilation would be critical in pushing the skill of seasonal prediction of the ISM toward its limit.
Is all the ISM variability unrelated to ENSO (about ~60% of ISM variability) unpredictable? If some of the remaining variability is driven by some other predictable driver, overall monsoon predictability may be enhanced. Recently, it was shown that when the equatorial SST anomalies are weak and going through a transition from El Niño to La Niña, or vice versa, SST anomalies over the north Pacific and north Atlantic influence the ISM rainfall and lead to the ISM’s significant interannual variability (Chattopadhyay et al., 2015). The framework of the ISM driven by the TT gradient also helps us to understand how the extratropical SST influences the ISM. The slowly varying extratropical SST causes a shift in the jet stream and storm tracks, thereby generating stationary waves leading to a large-scale persistent TT anomaly over northern India and southern Eurasia, thereby enhancing or decreasing the ΔTT and strengthening or weakening the ISM. The variations of the extratropical SST (and hence the TT anomalies in the north box) occur on a multidecadal timescale. When the tropical SST anomalies are strong (like a mature El Niño or La Niña), the ISM variability is governed primarily by the strong tropical SST teleconnection. However, when the tropical SST anomalies are weak, the strength of the ISM is driven by the extratropical SST teleconnection. Therefore, coupled models that simulate not only the tropical SST (such as ENSO) but also the extratropical SST with fidelity (such as PDO and AMO) are expected to enhance skill in making seasonal monsoon predictions. Efforts to identify other predictable drivers of ISM need to continue to improve monsoon predictions.
In addition to interactions between climate modes, the “climate noise” for interannual variability can also be generated through interactions between the subseasonal oscillations with the annual cycle. Because the spatial structure of the dominant MISO has significant projection on the spatial structure of seasonal mean ISM, the frequency of occurrence and the duration of the active and break spells could indeed affect the seasonal mean and its interannual variability (Goswami et al., 2006). As predictability of MISO is limited to 25 to 30 days, the “climate noise” generated by MISO is likely to be unpredictable, limiting the predictability of the ISM rainfall. How much of the “climate noise” or “internal” variability is governed by the MISOs? Model experiments simulating ISM’s “internal” interannual variability resulting from fixed annual forcing indicates that the linear contribution of MISOs, namely, the seasonal residual of MISO anomalies, explain about 25% of the “internal” interannual variability (Goswami & Xavier, 2005). As the MISOs have large amplitude (Figure 5c) and hence are nonlinear, MISOs could make a much larger contribution to ISM’s “internal” interannual variability through nonlinear interaction with the annual cycle (Goswami et al., 2006). If, however, the statistics on MISOs are modulated by slowly varying forcing, then the “climate noise” generated by the MISOs may not be completely unpredictable, thereby enhancing the potential predictability of the ISM rainfall. A recent study (Dwivedi et al., 2015) shows that the statistics on the MISO spells are indeed modulated by ENSO, suggesting that the potential predictability of ISM could be higher than previously thought.
Dynamics of the Multidecadal Variability of ISM
The instrumental records of ISM rainfall (Figure 3) indicate that approximately three decades of above-normal rainfall, with floods being more frequent than droughts, often follow approximately three decades of below-normal rainfall, with droughts outnumbering floods. These records indicate the existence of the ISM’s multidecadal variability. As this multdecadal variability appears to modulate the interannual variability (floods/droughts), proper characterization of the variability and understanding of its dynamics are likely to lead to improved seasonal prediction capability. By shifting the background mean by 2.5% above the long-term mean and by shifting it to 2.5% below normal in another phase, the multidecadal mode provides significant modulation of the interannual variability. However, the instrumental rainfall record is rather short to answer questions such as, What is the dominant periodicity associated with this variability, if at all one exists, and how robust is it? Also, the instrumental records indicate a rather long decreasing trend for ISM rainfall from 1960 to 2015 (Figure 12a). This apparently prolonged decreasing ISM rainfall trend is intriguing in the backdrop of increasing moisture content in the atmosphere due to anthropogenic warming, when the “expected” trend should have been increasing. Could such a decreasing trend of ISM rainfall be part of a natural multidecadal oscillation? Again, this question cannot be answered with the available instrumental record shown in Figure 3. Paleoclimate reconstruction of ISM rainfall with annual resolution has recently extended the record back more than 500 years by using tree-ring width (Borgaonkar et al., 2010). Further, using δ18O of cave stalagmites to reconstruct ISM rainfall has been extended to about 900 years (Sinha et al., 2007) and about 2000 years (Sinha et al., 2015), producing approximately yearly resolution. Even though these reconstructions of ISM rainfall may still not be useful for a detailed study of interannual variability, they seem to represent decadal to multidecadal variability well. A detailed study of several such long-reconstructed time series of ISM rainfall brings out a statistically significant 50- to 80-year multidecadal mode of variability (Goswami et al., 2015). Wavelet spectrum analysis of these time series (Goswami et al., 2015) shows that the mode is nonstationary, with one period persisting for a couple of centuries while shifting to another period for another extended span of time. The nonstationarity also contributes to the broadband aperiodic nature of the mode and shows that extended weak periods of five to six decades are not uncommon. Thus, the current decreasing trend of ISM rainfall may still be within its natural variability.
What drives the multidecadal variability of the ISM rainfall? Prominent multidecadal climate variations elsewhere could drive multidecadal variations of ISM through teleconnections. For example, the AMO is correlated with the multidecadal oscillation of ISM rainfall (Goswami et al., 2006a) through a teleconnection via modification of the TT gradient over the Indian monsoon region. It has also been shown that the multidecadal oscillation of ISM is related to a multidecadal mode of variability of ENSO through a teleconnection that is similar to the one that operates for the ENSO–monsoon relationship on the interannual timescale (Krishnamurthy & Goswami, 2000). Thus, multidecadal climate variability with its epicenter elsewhere (AMO, PDO, multidecadal ENSO) could impart a multidecadal component of ISM variability. As demonstrated for AMO, the multidecadal variability represents “internal” oscillations of the climate system involving ocean–atmosphere–land interactions (Delworth & Mann, 2000). Could local ocean–atmosphere feedback give rise to the multidecadal variability of ISM rainfall?
Recent examination of what drives the increasing trend of equatorial SST over the Indian Ocean (Figure 12b) and how it influences the Indian monsoon has led to the conclusion that indeed large-scale ocean–atmosphere feedback could lead to the ISM’s multidecadal variability. The warming trend of the equatorial Indian Ocean leading to a north–south SST gradient has been reported to be responsible for the decreasing trend of the land precipitation over the Indian monsoon region (Chung & Ramanathan, 2007). The decreasing trend of ISM rainfall over the past 50 to 60 years is indeed driven by the increasing trend of equatorial Indian Ocean SST (Roxy et al., 2015) by making the TISM decrease as the Indian Ocean SST increases (Figure 12b). The equatorial SST does this by making the TT over the southern oceanic box increase at a much faster rate than that over the northern box (Figure 12c). Thus, the equatorial warm SST has the tendency to weaken the ISM by weakening the TT gradient. But what is driving the increasing trend of equatorial IO SST? It has been shown that the weakening surface winds associated with the weakening strength of the ISM change the heat flux and ocean circulation in such a way as to increase the SST further (Swapna et al., 2013). Therefore, it is clear that the large-scale ocean–atmosphere interaction is responsible largely, if not fully, for the current decreasing trend of the ISM rainfall, establishing the notion that such regional air–sea interactions contribute to ISM’s multidecadal variability. ISM’s observed multidecadal variability is a result of contributions from the local air–sea interactions, together with contributions from teleconnections with multidecadal ENSO and AMO. The intrinsic timescale associated with the multidecadal variability arising from air–sea interaction involving the Indian Ocean SST and monsoon is not yet well established. Analysis of the recent decreasing trend of ISM and the increasing trend of Indian Ocean SST indicates that this timescale would be longer than 60 years. The Pacific Decadal Oscillation and the multidecadal ENSO are related, with a timescale of 50 to 70 years. The AMO also has a timescale of about 50 to 70 years. As a result of interactions between the timescales involved with the three different drivers, the multidecadal ISM mode assumes a broadband 50- to 80-year periodicity.
As discussed earlier, the current decreasing trend of ISM and the increasing trend of equatorial Indian Ocean SST are the result of a positive ocean–atmosphere feedback. In addition, this feedback leads to the monsoon’s multidecadal oscillation. For that to happen, a negative feedback is required that would arrest this positive feedback and make the ISM oscillate on this timescale. A clue to such a possible negative feedback involving the whole of the Indo-Pacific Basin is found from an examination of the trend in the outgoing longwave radiation (OLR) and surface winds during the past three to four decades (Figure 13a,b). Associated with the weakening monsoon precipitation, there is increasing convective activity (and rainfall) over the eastern Indian Ocean and maritime continent (Figure 13a). The atmosphere’s response to this tendency of increased convection in this region leads to strengthening easterly trades over central and western Pacific (Figure 13b). The response of the equatorial Pacific to this increased background easterly wind leads to deepening the mean thermocline in the west, shoaling it in the east, and increasing the east–west gradient of the mean thermocline. This oceanic background is conducive to more frequent La Niña conditions on interannual timescales and a multidecadal La Niña phase. Analysis of the low-frequency component of the SST indeed shows that the multidecadal ENSO (Figure 13c) is getting established in a La Niña phase (Figure 13d). The multidecadal La Niña is expected to increase positive TT anomaly over the north box, increase TISM, and start strengthening the ISM, eventually taking it to an above-normal state. This negative feedback is the key to the oscillating ISM multidecadal mode.
The challenge of simulating teleconnections with “external” drivers of ISM like ENSO with fidelity and simulating a large contribution of “internal” variability to ISM’s interannual variability makes the seasonal prediction of ISM a grand challenge. Slow but steady improvement of simulation of mean climate by coupled climate models over the past three decades has now made it possible to develop a seasonal prediction system for AIR with useful skill (Ramu et al., 2016). Together with predicting a spatially averaged seasonal mean quantity like AIR, it is also important to be able to predict the subseasonal monsoon spells. The fact that some models are now able to simulate the MISOs with reasonable fidelity has led to development of extended range predictions of the active and break spells 15 to 20 days in advance (Abhilash et al., 2014b). Thus, we are in the midst of an exciting time for climate model development for ISM forecasts. However, the seasonal skill of AIR forecasts as well as extended range predictions of MISO by any of these prediction systems is significantly below the potential limit on their predictability. Further, improvement of forecasts could come only from improving the existing systematic biases of the models in simulating the mean ISM climate and teleconnection with other external drivers like the ENSO. Improved coupled data assimilation is also expected to help improve seasonal as well as MISO forecasts.
What the anthropogenic greenhouse gases have done to the ISM during the past 50 years and how it is expected to influence ISM in the next 50 years is a question that remains uncertain (Turner & Annamalai, 2012). As indicated earlier, ISM’s decreasing trend during the past 50 years is in contrast to an expected increasing trend under anthropogenic influence and is most likely driven by an unstable ocean–atmosphere feedback. Some model studies claim that net cooling of the continent due to natural and anthropogenic aerosols may contribute heavily to the decreasing ISM rainfall (Ramanathan et al., 2005; Bollasina et al., 2011). However, these results are still being debated as they depend on the complexity with which the direct and indirect effects of aerosols are modeled. As we have shown, the decreasing trend of ISM could be part of a negative phase of a natural multidecadal variability. While the aerosols may contribute to the current decreasing trend in ISM rainfall, the pronounced multidecadal natural variability indicates that the role of the aerosols may be secondary. However, this remains an open question as the contributions from anthropogenic forcing, the natural variability, or the aerosols are not well constrained. Therefore, what is needed is to establish the robustness of ISM’s multidecadal variability as indicated earlier, with quantification of the period, amplitude, and uncertainty in estimating them using high-resolution multiproxy records. In order to separate the natural variability from the ISM response to the anthropogenic forcing, the coupled climate models must faithfully simulate the natural multidecadal variability of ISM, AMO, PDO, and ENSO. The climate models must also faithfully simulate ISM’s response to increased greenhouse gases. Unfortunately, almost all current climate models have large uncertainty in simulating both of these processes. For example, the majority of the CMIP5 (Coupled Model Inter-comparison Project, phase 5) indicate that ISM rainfall should increase by about 10% of mean by the end of the 21st century, as a result of increased greenhouse gases. However, it has been recently shown that this result is likely to have been due to a large systematic bias of the climate models in simulating too much convective precipitation compared to stratiform precipitation unlike in the observed climate (Sabeerali et al., 2014). Thus, it is clear that climate models with minimum systematic bias in simulating the ISM precipitation and other physical processes are critical to a reliable estimate response of ISM rainfall to climate change. Development of climate models that would simulate the observed multidecadal variability of ISM rainfall in the absence of increased greenhouse gas forcing is also critically important in making progress in delineating the natural variability from the anthropogenic response.
The annual cycle of ISM precipitation can be described as a manifestation of the seasonal migration of the ITCZ or the zonally oriented cloud (rain) band characterized by a sudden “onset.” The other important feature of ISM is its deep overturning meridional (regional Hadley circulation) associated with it, driven primarily by latent heat release associated with ISM (ITCZ) precipitation. The dynamics of monsoon climate, therefore, is an extension of the dynamics of the ITCZ. The classical land and sea-surface temperature gradient model of ISM may be adequate to explain the seasonal reversal of the surface winds, but it fails to explain the onset and the deep vertical structure of the ISM circulation. While the surface temperature over land cools after the onset, reversing the north–south surface temperature gradient and making it inadequate to sustain the monsoon after onset, it is the tropospheric temperature gradient that becomes positive at onset and remains strongly positive thereafter and maintains the monsoon. The change in sign of the tropospheric temperature gradient is dynamically responsible for a symmetric instability, leading to the onset and subsequent northward progression of the ITCZ. The unified ISM model in terms of the TT gradient provides a platform for understanding drivers of ISM variability by identifying processes that affect TT in the north and the south and influence the gradient.
The predictability of the seasonal mean ISM is limited by interactions of the annual cycle and higher-frequency monsoon variability within the season. Monsoon intraseasonal oscillation has a seminal role in influencing the seasonal mean, and its interannual variability is highlighted. Paleoclimate reconstructions of ISM rainfall reveal the ISM’s robust 50- to 80-year natural multidecadal variability in making the quantification of the global change signal in ISM difficult. While ISM climate on long timescales (e.g., multimillennium) largely follows the solar forcing, on shorter time scales ISM variability is governed by internal dynamics arising from ocean–atmosphere–land interactions, regional as well as remote, together with teleconnections with other climate modes. A conceptual model of how local ocean–atmosphere interactions, remote air–sea interactions (monsoon and Pacific Ocean), and teleconnections with multidecadal ENSO and AMO result in the multidecadal monsoon mode is proposed. The role of anthropogenic forcing such as the greenhouse gases and aerosols versus the natural multidecadal variability in the context of the recent six-decade-long decreasing trend of ISM rainfall also merits examination.
Chang, C. P., & Krishnamurti, T. N. (1987). Monsoon meteorology (Oxford Monograph on Geology and Geophysics, No. 7). Oxford: Oxford University Press.Find this resource:
Fein, J. S., & Stephens, P. (1987). Monsoons. New York: Wiley.Find this resource:
Ramage, C. S. (1971). Monsoon meteorology (International Geophysics Series, Vol. 15). San Diego, CA: Academic Press.Find this resource:
Wang, Bin. (2006). The Asian monsoon. Heidelberg: Springer, Praxis.Find this resource:
Abhilash, S., Sahai, A. K., Borah, N., Chattopadhyay, R., Joseph, S., Sharmila, S., De, S., Goswami, B. N., & Kuma, A. (2014a). Prediction and monitoring of monsoon intraseasonal oscillations over Indian monsoon region in an ensemble prediction system using CFSv2. Climate Dynamics, 42, 2, 801–2815.Find this resource:
Abhilash, S., Sahai, A. K., Pattnaik, S., Goswami, B. N., & Arun Kumar. (2014b). Extended range prediction of active-break spells of Indian summer monsoon rainfall using an ensemble prediction system in NCEP Climate Forecast System. International Journal of Climatology, 34, 98–113Find this resource:
Ashok, K., Guan, Z., Saji, N. H., & Yamagata, T. (2004). Individual and combined Influences of ENSO and the Indian Ocean Dipole on the Indian summer monsoon. Journal of Climate, 17, 3141–3155.Find this resource:
Bamzai A. S., & Shukla, J. (1999) Relation between Eurasian snow cover, snow depth, and the Indian summer monsoon: An observational study. Journal of Climate, 12, 3117–3132.Find this resource:
Bjerknes J. (1969). Atmospheric teleconnections from the equatorial Pacific. Monthly Weather Review, 97, 163–172.Find this resource:
Blanford, H. F. (1884). On the connexion of the Himalaya snowfall with dry winds and seasons of drought in India. Proceedings of the Royal Society of London, 37, 322.Find this resource:
Bollasina, Massimo A., Yi Ming, & Ramaswamy, V. (2011). Anthropogenic aerosols and the weakening of the South Asian summer monsoon, Science, 334, 502–505.Find this resource:
Boos, W. R., & Kuang, Z. (2010). Dominant control of the South Asian monsoon by orographic insulation versus plateau heating. Nature, 463, 218–222.Find this resource:
Borgaonkar, H. P., Sikder, A. B., Somaru Ram, & Pant, G. P. (2010). El Niño and related monsoon drought signals in 523-year-long ring width records of teak (Tectona grandis L.F.) trees from south India. Palaeogeography, Palaeoclimatology, Palaeoecology, 285, 74–84.Find this resource:
Carton, J. A., & Giese, B. S. (2008). A reanalysis of ocean climate using simple ocean data assimilation (SODA). Monthly Weather Review, 136, 2999–3017.Find this resource:
Chakraborty, A., Nanjundiah, R., & Srinivasan, J. (2002). Role of Asian and African orography in Indian summer monsoon. Geophysical Research Letters, 29, 50–51.Find this resource:
Charney, J. G., & Shukla, J. (1980). Predictability of monsoons. In Sir J. Lighthill & R. P. Pearce (Eds.), Monsoon dynamics (pp. 99–110). Cambridge, U.K.: Cambridge University Press.Find this resource:
Chatterjee Piyali, & Goswami, B. N. (2004). Structure, genesis and scale selection of the tropical quasi-biweekly mode. Quarterly Journal of the Royal Meteorological Society, 130, 1171–1194.Find this resource:
Chattopadhyay, R., Phani, R., Sabeerali, C. T., Dhakate, A. R., Salunke, K., Mahapatra, S., Suryachandra Rao, A., & Goswami, B. N. (2015). Influence of extra-tropical sea surface temperature on the Indian summer monsoon: An unexplored source of seasonal predictability. Quarterly Journal of the Royal Meteorological Society, 141, 2760–2775.Find this resource:
Chen, T.-C., & Chen, J.-M. (1993). The 10.20 day mode of the 1979 Indian monsoon: Its relation with time variation of monsoon rainfall. Monthly Weather Reviews, 121, 2465–2482.Find this resource:
Cheng, H., Lawrence E. R., Sinha, A., Spötl, C., Liang Yi, Chen, S., et al. (2016) The Asian monsoon over the past 640,000 years and ice age terminations, Nature, 534, 640–646.Find this resource:
Chou, C., & Neelin, J. D. (2003). Mechanisms limiting the northward extent of the northern summer monsoons over North America, Asia, and Africa. Journal of Climate, 16, 406–425.Find this resource:
Chung C. E., & Ramanathan, V. (2007). Relationship between trends in land precipitation and tropical SST gradient. Geophysical Research Letters, 34, L16809.Find this resource:
Delworth, T., & Mann, M. (2000). Observed and simulated multi-decadal variability in the NH. Climate Dynamics, 16, 661–676.Find this resource:
Dixit, Yama, Hodell, David A., & Petrie, Cameron A. (2014). Abrupt weakening of the summer monsoon in northwest India ~4100 yr ago. Geology, 4 (2), 339–342.Find this resource:
Dwivedi, S., Goswami, B. N., & Kucharski, F. (2015). Unraveling the missing link of ENSO control over the Indian monsoon rainfall. Geophysical Research Letters, 42, 8201–8207.Find this resource:
Fasullo, J. (2004). A stratified diagnosis of the Indian monsoon—Eurasian snow cover relationship. Journal of Climate, 17, 1110–1122.Find this resource:
Gadgil, S. & Gadgil, S. (2006). ‘The Indian Monsoon, GDP and Agriculture’, Economic and Political Weekly, 25 November, 4887–95.Find this resource:
Gill, A. E. (1980). Some simple solutions for heat-induced tropical circulation. Quarterly Journal of the Royal Meteorological Society, 106, 447–462.Find this resource:
Goswami, B. N. (1998). Inter-annual variation of Indian summer monsoon in a GCM: External conditions versus internal feedbacks, Journal of Climate, 11, 501–522.Find this resource:
Goswami, B. N. (2012). South Asian monsoon. In W. K. M. Lau & D. Waliser (Eds.), Intraseasonal variability in the atmosphere-ocean climate system (2d ed., pp. 21–72). Chichester: Springer-Praxis.Find this resource:
Goswami, B. N., Ajaya Mohan, R. S., Xavier, Prince K., & Sengupta, D. (2003). Clustering of low pressure systems during the Indian summer monsoon by intraseasonal oscillations. Geophysical Research Letters, 30(8), 1431.Find this resource:
Goswami, B. N., Annamalai, H., & Krishnamurthy, V. (1999). A broad scale circulation index for inter-annual variability of the Indian summer monsoon. Quarterly Journal of the Royal Meteorological Society, 125, 611–633.Find this resource:
Goswami, B. N., Kriplani, R. H., Borgaonkar, H. P., & Preethi, B. (2015), Multi-decadal variability of Indian summer monsoon rainfall using proxy data. In C. P. Chang, Michael Ghil, Mojif Latif, & Mike Wallace (Eds.), Climate change: Multi-decadal and beyond (pp. 327–346). London: World Scientific.Find this resource:
Goswami, B. N., & Shukla, J. (1984). Quasi-periodic oscillations in a symmetric general circulation model. Journal of the Atmospheric Sciences, 41, 20–37.Find this resource:
Goswami B. N., & Shukla, J. (1991). Predictability of a coupled ocean atmosphere model. Journal of Climate, 4, 3–22.Find this resource:
Goswami, B. N., Shukla, J., Schneider, E. K., & Sud, Y. (1984). Study of the dynamics of the inter-tropical convergence zone with a symmetric version of the GLAS climatic model. Journal of the Atmospheric Sciences, 41, 5–19.Find this resource:
Goswami B. N., Wu, G., & Yasunari, T.(2006). Annual cycle, intraseasonal oscillations and roadblock to seasonal predictability of the Asian summer monsoon. Journal of Climate, 19, 5078–5099.Find this resource:
Goswami, B. N., & Xavier, Prince K. (2003). Potential predictability and extended range prediction of Indian summer monsoon breaks. Geophysical Research Letters, 30(18), 1966.Find this resource:
Goswami, B. N., & Xavier, P. K. (2005). Dynamics of “internal” inter-annual variability of Indian summer monsoon in a GCM. Journal of Geophysical Research, 110, D24104.Find this resource:
Goswami, B. N., M. S. Madhusoodanan, C. P. Neema, and D. Sengupta (2006a). A physical mechanism for North Atlantic SST influence on the Indian summer monsoon[http://dx.doi.org/10.1029/2005GL024803]. Geophysical Research Letters, 33, L02706.Find this resource:
Hai, Cheng, R., Edwards, L., Sinha, A., Spötl, C., Yi, L., Chen, S., Kelly, M. et al. (2016). The Asian monsoon over the past 640,000 years and ice age terminations. Nature, 534, 640–646.Find this resource:
Han, D. G. & Shukla, J. (1976). An apparent relationship between Eurasian snow cover and Indian monsoon rainfall. Journal of the Atmospheric Sciences, 33, 2461–2462.Find this resource:
Held, I. M., & Hou, A. Y. (1980). Nonlinear axially symmetric circulations in a nearly inviscid atmosphere. Journal of the Atmospheric Sciences, 37, 515–533.Find this resource:
Jiang, X., Li, N. T., & Wang, B. (2004). Structures and mechanism of northward propagating boreal summer intrasesaonal oscillations. Journal of Climate, 17, 1022–1039.Find this resource:
Kang, In-Sik, & Shukla, J. (2006). Dynamic seasonal prediction and predictability of the monsoon. In Bin Wang (Ed.), The Asian Monsoon (Chapter 15). Berlin: Praxis.Find this resource:
Kripanani, R. H., & Kulkarni, A. (1999). Climatology and variability of historical Soviet snow depth data: Some new perspectives in snow—Indian monsoon teleconnections. Climate Dynamics, 15, 475–489.Find this resource:
Krishnamurthy V., & Goswami, B. N. (2000). Indian monsoon–ENSO relationship on interdecadal timescale. Journal of Climate, 13, 579–595.Find this resource:
Krishnamurti, T. N., & Bhalme, H. N. (1976). Oscillations of a monsoon system, Part I: Observational aspects. Journal of the Atmospheric Sciences, 33, 1937–1954.Find this resource:
Krishnakumar, V., & K. M. Lau. (1998). Possible role of symmetric instability in the onset and abrupt transition of the Asian monsoon. Journal of the Meteorological Society of Japan, 76, 363–383.Find this resource:
Krishnakumar, K., Rajagopalan, B., Hoerling, M., Bates, G. & Cane, M. (2006). Unraveling the mystery of Indian monsoon failure during El Niño. Science, 314(5796), 115–119.Find this resource:
Li, T., Wang, B., Chang, C.-P., & Zhang, Z. (2003). A theory for the Indian Ocean dipole–zonal mode. Journal of the Atmospheric Sciences, 60, 2119–2135.Find this resource:
Lindzen, R., & Nigam, S. (1987). On the role of sea surface temperature gradients in forcing low-level winds and convergence in the tropics. Journal of the Atmospheric Sciences, 44, 2418–2436.Find this resource:
Loschnigg, J., Meehl, G. A., Webster, P. J., Arblaster, J. M., & Compo, G. A. (2003). The Asian monsoon, the tropospheric biennial oscillation and the Indian Ocean zonal mode in the NCAR GCM. Journal of Climate, 16, 1617–1642.Find this resource:
Loschnigg, J., & Webster, P. J. (2000). A coupled ocean-atmosphere system of SST regulation of the Indian monsoon. Journal of Climate, 13, 3342–3360.Find this resource:
Meyers, G., McIntosh, P., Pigot, L., & Pook, M. (2007). The years of El Niño, La Niña, and interactions with the tropical Indian Ocean. Journal of Climate, 20, 2872–2880.Find this resource:
Molnar, P., Boos, R. W., & Battisti, D. S. (2010). Orographic controls on climate and paleoclimate of Asia: Thermal and mechanical roles for the Tibetan Plateau. Annual Review of Earth and Planetary Sciences, 38, 77–102.Find this resource:
Neelin, J. D., & Held, I. M. (1987). Modeling tropical convergence based on the moist static energy budget. Monthly Weather Reviews, 115, 3–12.Find this resource:
Pai, D. S., Sridhar, L., M., Rajeevan, M., Sreejith, O. P., Satbhai, N. S., & Mukhopadhyaya, B. (2014, January). Development of a new high spatial resolution (0.25° × 0.25°) long period (1901–2010) daily gridded rainfall data set over India and its comparison with existing data sets over the region. MAUSAM, 65(1), 1–18.Find this resource:
Pant, G. B., & Parthasarathy, S. B. (1981). Some aspects of an association between the southern oscillation and Indian summer monsoon. Archives for Meteorology, Geophysics, and Bioclimatology, Series B. 29, 245.Find this resource:
Parthasarathy, B., Munot, A. A., & Kothawale, D. R. (1994). All-India monthly and seasonal rainfall series: 1871–1993. Theoretical and Applied Climatology, 49, 217.Find this resource:
Philander, S. G. H. (1990). El Niño, La Niña and the Southern Oscillation. London: Academic Press.Find this resource:
Rajeevan, M., Unnikrishnan, C. K., & Preethi, B. (2012). Evaluation of the ENSEMBLES multi-model seasonal forecasts of Indian summer monsoon variability. Climate Dynamics, 38, 2257–2274.Find this resource:
Ramanathan V., Chung, C., Kim, D., Bettge, T., Buja, L., Kiehl, J. T., Washington, W. M., et al. (2005). Atmospheric brown clouds: Impacts on South Asian climate and hydrological cycle. Proceedings of the National Academy of Sciences, 102, 5326–5333.Find this resource:
Ramu, D. A., Sabeerali, C. T., Chattopadhyay, R., Rao, D. N., George, G., Dhakate, A. R., et al. (2016). Indian summer monsoon rainfall simulation and prediction skill in the CFSv2 coupled model: Impact of atmospheric horizontal resolution. Journal of Geophysical Research: Atmospheres, 121, 5, 2205–2221.Find this resource:
Rasmusson, E. M., & Carpenter, T. H. (1983). The relationship between eastern equatorial Pacific seas surface temperature and rainfall over India and Sri Lanka. Monthly Weather Reviews, 111, 517–528.Find this resource:
Robock, A., Mu, M. Q., Vinnikov, K., & Robinson, D. (2003). Land surface conditions over Eurasia and Indian summer monsoon rainfall. Journal of Geophysical Research, 108, 4131.Find this resource:
Rodwell, M. J., & Hoskins, B. J. (1996). Monsoons and dynamics of deserts. Quarterly Journal of the Royal Meteorological Society, 122, 1385–1404.Find this resource:
Rodwell, M. J., & Hoskins, B. J. (2001). Subtropical anticyclone and summer monsoon. Journal of Climate, 14, 3192–3211.Find this resource:
Roxy, M., Ritika, K., Terray, P., Murtugudde, R., Ashok, K., & Goswami, B. N. (2015). Drying of Indian subcontinent by rapid Indian Ocean warming and a weakening land-sea thermal gradient. Nature Communications, 6, 7423.Find this resource:
Sabeerali, C. T., Suryachandra, A., Rao, R. S., Dhakate, A. R., Salunke, K., & Goswami, B. N. (2014). Why ensemble mean projection of south Asian monsoon rainfall by CMIP5 models is not reliable? Climate Dynamics, 45(1–2), 161–174.Find this resource:
Saji, N. H., Goswami, B. N., Vinayachandran, P. N., & Yamagata, T. (1999). A dipole mode in the tropical Indian Ocean. Nature, 401, 360–363.Find this resource:
Saji, N. H., & Yamagata, T. (2003). Possible impacts of Indian Ocean dipole mode events on global climate. Climate Research, 25, 151–169.Find this resource:
Sarachik, E. S., & Cane, M. A. (2010). The El Niño-Southern Oscillation phenomenon. Cambridge, U.K.: Cambridge University Press.Find this resource:
Schott, F. A., Dengler, M., & Schoenefeldt, R. (2002). The shallow overturning circulation of the Indian Ocean. Progress in Oceanography, 53, 57–103.Find this resource:
Schneider, T., Bischoff, T. and Haug, G. H. (2014) Migrations and dynamics of the intertropical convergence zone. Nature, 513(7516), 45–53.Find this resource:
Sengupta, D., & Ravichandran, M. (2001). Oscillations of Bay of Bengal sea surface temperature during the 1998 summer monsoon. Geophysical Research Letters, 28(10), 2033–2036.Find this resource:
Sengupta, D., Goswami, B. N. &, Senan, R. (2001). Coherent Intraseasonal Oscillations of Ocean and Atmosphere during the Asian Summer Monsoon. Geophysical Research Letters, 28(21), 4127–4130.Find this resource:
Schneider, Tapio, Bischoff, Tobias, & Haug, Gerald H. (2014). Migrations and dynamics of the intertropical convergence zone. Nature, 513, 45–53.Find this resource:
Sharmila, S., Pillai, P. A., Joseph, S., Roxy, M., Krishna, R. P. M., Chattopadhyay, R., Abhilash, S., et al. (2013). Role of ocean-atmosphere interaction on northward propagation of Indian summer monsoon intraseasonal oscillations (MISO). Climate Dynamics, 41(1), 651–669.Find this resource:
Suhas, E., Neena, J. M., & Goswami, B. N. (2013). An Indian monsoon intraseasonal oscillations (MISO) index for real time monitoring and forecast verification. Climate dynamics, 40(11–12), 2605–2616.Find this resource:
Shukla, J. (1981). Dynamical predictability of monthly means. Journal of the Atmospheric Sciences, 38(12), 2547–2572.Find this resource:
Sikka, D. R. (1977). Some aspects of the life history, structure and movements of monsoon depressions. Pure and Applied Geophysics, 115, 1501–1529.Find this resource:
Sikka, D. R., & Gadgil, S. (1980). On the maximum cloud zone and the ITCZ over Indian longitudes during southwest monsoon. Monthly Weather Reviews, 108, 1840–1853.Find this resource:
Sinha, A., Cannariato, K. G., Stott, L. D., Cheng, H., Edwards, R. L., Yadava, M. G., et al. (2007). A 900-year (600 to 1500 AD) record of Indian summer monsoon precipitation from the core monsoon zone of India. Geophysical Research Letters, 34 (16), L16707.Find this resource:
Sinha, A., Kathayat, G., Cheng, H., Breitenbach, S. F. M., Berkelhammer, M., Mudelsee, M., Biswas, J., & Edwards, R. L. (2015). Trends and oscillations in the Indian summer monsoon rainfall over the last two millennia. Nature Communications, 6, 6309.Find this resource:
Staubwasser, M., Sirocko, F., Grootes, P. M., & Segl, M. (2003). Climate change at the 4.2 ka BP termination of the Indus Valley civilization and Holocene south Asian monsoon variability[http://dx.doi.org/10.1029/2002GL016822]. Geophysical Research Letters, 30, 1425.Find this resource:
Swapna, P., Krishnan, R., & Wallace, J. M. (2013). Indian Ocean and monsoon coupled interactions in a warming environment. Climate Dynamics, 42(9–10), 2439–2454.Find this resource:
Tomas, R., & P. Webster. (1997). The role of inertial instability in determining the location and strength of near-equatorial convection. Quarterly Journal of the Royal Meteorological Society, 123, 1445–1482.Find this resource:
Tripati, S., & Raut, L. N. (2006). Monsoon wind and maritime trade: A case study of historical evidence from Orissa, India. Current Science, 90(6), 864–871.Find this resource:
Turner, A. G., & Annamalai, H. (2012). Climate change and the South Asian summer monsoon. Nature Climate Change, 2, 587–595.Find this resource:
Turner, A. G., & Slingo, J. M. (2011). Using idealized snow forcing to test teleconnections with the Indian summer monsoon in the Hadley Centre GCM. Climate Dynamics, 36, 1717.Find this resource:
Waliser, D. E., Stern, W., Schubert, S. and Lau, K. M. (2003). Dynamic predictability of intraseasonal variability associated with the Asian summer monsoon. Quarterly Journal of the Royal Meteorological Society, 129(594), 2897–2925.Find this resource:
Walker GR (1910) Correlations in seasonal variations of weather II. Mem Indian Meteorol Dept, 21, 22–45.Find this resource:
Walker, G. T. (1923). Correlations in seasonal variability of weather. VIII: A preliminary study of world weather. Memoirs of the Indian Meteorological Department, 24, 75–131.Find this resource:
Walker, G. T. (1928). World weather. Quarterly Journal of the Royal Meteorological Society, 54(226), 79–87.Find this resource:
Wang, B. (2006). The Asian monsoon. Heidelberg: Springer, Praxis.Find this resource:
Wang, C., & Picaut, J. (2004). Understanding Enso physics—A review. In C. Wang, S. P. Xie, & J. A. Carton (Eds.), Earth’s climate (pp. 21–48). Washington, DC: American Geophysical Union.Find this resource:
Webster, P. J. (2006). The coupled monsoon system. In Bin Wang (Ed.), The Asian monsoon (pp. 1–66). Heidelberg: Springer, Praxis.Find this resource:
Webster, P. J., Magana, V. O., Palmer, T. N., Shukla, J., Tomas, R. T., Yanai, M., & Yasunari, T. (1998). Monsoons: Processes, predictability and the prospects of prediction. Journal of Geophysical Research, 103(C7), 14451–14510.Find this resource:
Webster, P. J., Moore, A. M., Loschnigg, J. P., & Leben, R. R. (1999). Coupled ocean–atmosphere dynamics in the Indian Ocean during 1997–98. Nature, 401, 356–360.Find this resource:
Wu, R. & Kirtman, B. P. (2003). On the impacts of the Indian summer monsoon on ENSO in a coupled GCM. Quarterly Journal of the Royal Meteorological Society, 129(595), 3439–3468.Find this resource:
Xavier, Prince, K., Marzin, Charline, & Goswami, B. N. (2007). An objective definition of the Indian summer monsoon season and a new perspective on the ENSO–monsoon relationship.Quarterly Journal of the Royal Meteorological Society, 133, 749–764.Find this resource:
Yasunari, T. (1979). Cloudiness fluctuations associated with the Northern Hemisphere summer monsoon. Journal of the Meteorological Society of Japan, 57, 227–242.Find this resource: