Agricultural production and the availability of fresh water in Indian subcontinent critically depend on the monsoon rains. Therefore it is vital to understand the causal mechanisms underlying the observed changes in the Indian monsoon in the past. Paleomonsoon reconstructions show that the water discharge from the Ganges-Brahmaputra River system to the Bay of Bengal was maximum in the early to mid-Holocene; data from the Western Arabian Sea and Omanian speleothems indicate declining monsoon winds during the Holocene, whereas records from the South West Monsoon (SWM) precipitation dominated eastern Arabian Sea show higher runoff from the Western Ghats indicating gradually increasing monsoon precipitation during the Holocene. Thus there exists considerable spatial variability in the monsoon in addition to the temporal variability that needs to be assessed systematically. Here we discuss the available high resolution marine and terrestrial paleomonsoon records such as speleothems and pollen records of the SWM from important climatic regimes such as Western Arabian Sea, Eastern Arabian Sea, Bay of Bengal to assess what we have learnt from the past and what can be said about the future of water resources of the subcontinent in the context of the observed changes.

1. Introduction

The Indian economy is based on agriculture, which mostly depends on the monsoon rain and to some extent on river flow and ground water resources. In the absence of monsoon that brings adequate rain, crop yield is reduced and due to recurrent droughts there may even be severe shortage of drinking water. The water resources of India comprise rivers, lakes, and ground water aquifers and the amount of water they hold is linked to the rainfall on the one hand and human exploitation on the other. Thus it is important to have a correct long-term forecast of the monsoon that can help in the proper management of our water resources [1]. Monsoon prediction is seriously hampered by the nonavailability of past data, which is limited to about hundred years [2]. It is very difficult to predict the monsoon without understanding its full variability. Generating quantitative paleomonsoon data using available, dateable, natural archives, such as deep sea and lake sediments, varved sediments, and speleothems is a starting point towards this end [36].

Monsoon is a term derived from an Arabic word “Mausim” meaning weather. It is technically applied to the seasonal reversal of winds in the Indian subcontinent and Africa, especially in the Arabian Sea, due to land-sea thermal and pressure contrast. It is mainly due to coupled heating and cooling of Himalaya (Tibetan plateau) and the southern Indian Ocean and the consequent movement of the ITCZ [7]. The Asian monsoon system is a dynamic component of the modern climate system and changes in this convectively active region can result in severe droughts or floods over large, densely populated regions [8]. The inherent seasonality of monsoon circulation leads to cool, dry winters and warm, wet summers over the Asian landmass. These seasonal changes in atmospheric circulation and precipitation also affect the ocean, leading to a strong seasonality in the strength and direction of ocean currents, sea-surface temperature (SST), and salinity patterns, as is observed in the Indian Ocean and the South China Sea (SCS). In specific regions, such as the Northwestern Arabian Sea, these dynamics lead to well-defined seasonal upwelling regimes in the open-ocean and near-shore environments [9]. The south Asian monsoon has been known to be stronger during warm climate (interglacial/interstadial) and weaker during cooler periods (glacial/stadial) ([10] and references therein), while the winter monsoon behaves the other way [1113]. Thus monsoons are components of the global climate that play an important role in water resources of the Indian subcontinent.

The Southwest monsoon (SWM) occurs during June to September and the Northeast Monsoon (NEM) affects the southern parts of the Indian subcontinent during October to December [14]. AISRTS (All India Summer Monsoon Rainfall Time Series) is available from 1871 onwards [15], which has documented the last ~140 years of rainfall. If rainfall exceeds by more than 10% from the long-term average, it is called as excess rainfall year, while when it is lower by 10% or more, it is a deficient rainfall year. But to understand full variability of monsoon, which assumes added importance in view of the presently experienced global warming, we require records of monsoon during differing climatic conditions extending back to thousands of years.

2. Multiproxy Comparison of Studies from Different Regions

Evolution and variability of the Asian monsoon system are believed to respond to at least five types of large-scale climate forcing or changes in boundary conditions [49], including (i) the tectonic development of the Himalayan-Tibetan orography, in million-year time scales, (ii) changes in the atmospheric CO2 concentration, in time scales of tens of thousand of years, (iii) changes in the Earth's orbit that result in periodic variations of seasonal solar radiation, in time scales of tens of thousands of years, (iv) changes in the extent of ice sheets (thousand years time scales), and (v) internal feedbacks within the climate system (multiple-time scales). These factors act simultaneously and over different time scales to amplify or lessen the seasonal development of continental heating/cooling, land-sea pressure gradients, latent heat transport, and moisture convergence, all of which control the strength of the monsoon circulation. We present below a comparative analysis of multiproxy studies from diverse terrestrial (speleothems—stalactites and stalagmites from Indian and Oman caves) and marine (western, northern and eastern Arabian Sea along with Bay of Bengal) realms divided into different time periods, which would help us to understand the spatiotemporal variability and complexity of the south Asian monsoon. The focus of this paper is the high-resolution records with accurate, absolute chronology since Last Glacial Maximum (~21,000 years before present)—a period which covers extensive glaciation, deglacial period witnessing rapid climatic fluctuations, and finally the Holocene (past ~11,700 years, [4]), which is a period of relatively unvarying warmth.

2.1. Monsoon and Associated Oceanographic Effects from Marine Proxies

During the summer and winter monsoons the surface oceanic circulation in the Northern Indian Ocean (Arabian Sea and Bay of Bengal) experiences changes in direction in consonance with the changing wind patterns [50, 51]. Intense upwelling occurs along the Somalian and Oman coasts with a transport of 1.5–2 Sv in the upper 50 m [52]. The typical temperature of the upwelled water is 19–24°C [53]. The reason attributed for such intense coastal upwelling is the Ekman divergence due to the flow of strong winds parallel to the coast. The central Arabian Sea exhibits a bowl-shaped mixed layer deepening under the effect of Findlater Jet wind-stress forcing and Ekman pumping [54, 55]. The cold and dry Northeast monsoon winds accompanied by the Ekman pumping cause subduction of the high salinity surface waters in the northern Arabian Sea [56, 57].

The upwelling zones along the Somalian and Oman coasts cause intense biological and geochemical changes in this region with SST falling by ~4°C as nutrient-rich deeper water surfaces that enhance the sea surface biological productivity considerably [9, 50, 58]. Weak upwelling also occurs along coastal southwest India [50, 59]. During the Northeast monsoon, minor upwelling is observed in the northeastern Arabian Sea [50]. The cold and dry NE monsoon winds cause the deepening of the mixed layer to a depth of 100–125 m due to convective mixing in the northern Arabian Sea, which leads to nutrient injection and hence high productivity during winter monsoon in this region [60, 61]. The typical productivity values for the western Arabian Sea are 2.0, 1.0, and 0.5 g C/m2/day for the SW monsoon, NE monsoon, and the intermonsoon periods, respectively [62, 63]. Similarly for the eastern Arabian Sea the typical productivity values are 0.6, 0.3, and 0.2 g C/m2/day for the SW monsoon, NE monsoon, and the intermonsoon periods, respectively [64]. As the moisture laden SW monsoon winds approach the Western Ghats they are forced to ascend resulting in copious precipitation and runoff into the coastal Arabian Sea, reducing the sea surface salinity considerably [21]. Denitrification takes place due to the very low concentration of oxygen in the entire Arabian Sea from 250 m to 1250 m water depths [65, 66]. This oxygen minimum zone (OMZ) is due to the high-oxygen consumption below the thermocline for the oxidation of organic matter supplied by the high overhead surface productivity. Furthermore the sluggish flow of the oxygen poor intermediate water [66, 67] along with a strong tropical thermocline (due to relatively high SST that prevents mixing of the oxygen-rich surface waters with the deeper waters) maintains the OMZ [68, 69]. Thus OMZ and denitrification are the interplay of monsoon winds and the ensuing productivity along with other climatically controlled factors such as ocean ventilation rate [40, 7073].

Such pronounced changes in the seawater characteristics make the Arabian Sea ideal for deciphering the past changes in monsoon intensity. The surface productivity that manifests itself in many forms such as organic, calcareous, and siliceous productivity, also affects the carbon isotopic composition of the seawater, which is preserved in the calcitic shells of various foraminifera. Similarly the SST and sea surface salinity alter the oxygen isotopic composition of these shells and they get recorded in the sea sediments. The nitrogen isotopic composition of sedimentary organic matter can indicate the denitrification intensity relatable to productivity variations. Thus the downcore variations of such proxies could help document the past variations in monsoon intensity and the related climatic changes.

3. Discussion

Paleomonsoon studies in the Indian region were initiated around 30 years ago by Prell et al. [74] and Bryson and Swain [75]. Since then, a large number of workers ([2, 1627, 2932, 4144, 7678] and references therein) have carried out high-resolution monsoon studies in archives from various locations in and around the subcontinent that are influenced by the monsoon winds/precipitation. The data thus generated has helped document the fluctuations in the past monsoon strength both in space and time. Different proxies such as planktic/benthic foraminiferal abundances, their stable oxygen and carbon isotope ratios [18, 19, 21, 23, 24, 76, 77], varved sediments [41, 42], speleothems [2, 2932], tree-rings [79], δ13C and δ15N of lacustrine and marine organic matter [10, 73, 8083], and pollen records [84, 85] have been used to obtain paleomonsoon records with different resolutions. A few of the important proxies and their significance are shown in Table 2.

Recent marine studies have the advantage of accurate AMS (Accelerator Mass Spectrometry) 14C dating, and in some case, have provided time resolutions as low as ~50 years (e.g., [23]) that can go up to subdecadal scale in extreme cases such as varved sediments [41]. The past strength of SWM was elucidated by using the above proxies from different monsoon-sensitive geographical regime (Figure 1) such as (i) the western Arabian Sea—experiences high productivity due to SWM wind induced upwelling [1619, 76, 77]—thus record strength of monsoon winds; (ii) the northern Arabian Sea—affected by amount of Indus river discharge [43, 44] and associated varve thickness [41, 42] relatable to SWM precipitation intensity; (iii) India and Oman—the growth rates and stable oxygen and carbon records of speleothems were used to quantitatively reconstruct SWM precipitation intensity [2, 29, 30, 32, 86]; (iv) the south-eastern Arabian Sea sediment cores which are influenced by the surface runoff due to SWM precipitation from the western Ghats of India [21, 23, 25]—thus records strength of SWM precipitation; (v) water discharge from the Ganga-Brahmaputra (G-B) river system into the Bay of Bengal (BOB) [3537]. The major inferences drawn from these regions have been presented in Table 1. For Indian populace, the more important aspect of monsoon is precipitation variability that may result in severe droughts or devastating floods. Therefore it is more important to decipher this aspect for which eastern Arabian Sea is better suited than other regions of Indian Ocean. From the comparative analysis, as represented in Figure 2, among different regions, it is clear that at short-time scales, monsoon exhibits a high-spatial variability—different regions experience different trends in monsoon intensity. But when we look at multi-millennial time scales then we find that monsoon records from varied realms exhibit similar trends.

Fleitmann et al. [32] deciphered declining SWM precipitation during the Holocene based on speleothems from Oman. But this region is near the edge of the monsoon precipitation and receives very little rain as compared to the Indian subcontinent. Moreover, such arid/semiarid regions with dynamic karstic terrains have been reported to have long residence time of water up to decades [88]. Also, the strong evaporation in such regions could greatly alter the δ18O of precipitation during infiltration and in the upper portion of the vadose zone [89], making such reconstructions somewhat ambiguous. In the western Arabian Sea during the Holocene, a few of the SWM wind intensity, based proxies (e.g., content of G. bulloides—a calcareous micro-organism flourishing in the cooler, upwelled waters during SWM season) showed that monsoon winds have been declining, following the reduction in insolation. On the other hand, studies from the eastern Arabian Sea, which record SWM precipitation, have indicated otherwise. The eastern Arabian Sea receives abundant fresh water as either direct precipitation or runoff from the adjacent Western Ghats (Sahyadri Hills) that induces intense orographic precipitation during the SW monsoon. This reduces the sea surface salinity (SSS) that is reflected in the oxygen isotopic composition (δ18O) of planktic foraminifera. Such reconstructions have indicated increasing strength of SWM precipitation during the Holocene [21, 22, 27]. Similarly, Agnihotri et al. [28] have found increasing productivity during the Holocene based on denitrification intensity (δ15N), which is relatable to SWM strength. To resolve this apparent contradiction between the studies from the western and the eastern margins of the Arabian Sea, Tiwari et al. [10] studied productivity proxies from the western Arabian Sea. They found out that the reason for declining trend observed in the planktic carbonate production (relative abundance (%) of G. bulloides—[19] and declining %CaCO3—[10]) is the preference to silicate productivity over calcareous productivity during periods of enhanced monsoon winds. During initial stages of SWM (weaker monsoon winds), upwelling takes place from shallower regions bringing nitrate and phosphate to the surface that supports calcareous microorganisms. But as the monsoon progresses and the winds became stronger, upwelling takes place from the deeper waters, injecting silicate to the photic zone, which enhances siliceous productivity [90, 91]. This has been observed in sediment records of past climate as well [45, 92]. Other productivity indicators (organic carbon, δ15N, δ13C of three species of foraminifera) analyzed by Tiwari et al. [10] unambiguously indicate strengthening monsoon during Holocene, unlike insolation that declines during the same period. This multiproxy result indicates that, on sub-Milankovitch, multi-millennial timescales, monsoon is predominantly governed by internal feedback mechanisms and lagged summer insolation maxima by several thousand years, as noted earlier by Clemens et al. [49].

4. Conclusions

In essence, the above discussion shows that there exists a large spatial variability in monsoon records, which becomes more pronounced on shorter timescales. But on longer time scales, a more coherent picture emerges. On glacial-interglacial time scales, SWM was stronger (weaker) during the warmer (colder) periods, with a minimum during the Last Glacial Maximum. During deglaciation, monsoon fluctuated widely with weaker monsoons during colder episodes such as Younger Dryas and stronger monsoons during warm episodes such as Bølling-Allerød. During the early Holocene, a widely reported SWM maximum is followed by a decline. Thereafter, during the Holocene, the SWM either stayed uniform or showed decline only after the mid-Holocene, which needs to be verified further with accurately dated records from both the western and eastern Arabian Sea margins. During the Holocene, monsoon did not decline following the reducing insolation, which highlights the importance of internal feedback mechanisms. On short time scales (millennial to subcentennial), a period of widespread aridity is reported at ~2000 yr BP followed by arid periods at ~1500 yr BP, ~1100 yr BP, ~850 yr BP, and ~500 yr BP [23]. On such short time scales (centennial to subcentennial), monsoon has been reported to follow insolation [93]. This highlights complex dynamics of SWM at different timescales. This intercomparison of monsoon records from different regions show that despite the considerable spatial variability in monsoon strength, it increases during warmer periods in general. This indicates that monsoon may strengthen in the future scenario of global warming that corroborates the model results represented in IPCC AR4. An understanding of this requires systematic studies covering various regions under the SWM realm using multiple proxies at different spatiotemporal scales.


The authors thank ISRO-GBP for funding. M. Tiwari thanks the Director-NCAOR for encouragement. A. K. Singh thanks Delhi University for support. This is NCAOR Contribution no. 06/2011.