Accelerated melting of Himalayan snow and ice triggers pronounced changes in a valley peatland from northern India



[1] The Himalayan region of northern India depends on monsoon rains, together with snow and glacial melt, to supply life-sustaining water to one of the world's most densely populated areas. Here we provide high-resolution pollen and diatom evidence from a peat deposit in the Pinder Valley that shows a synchronous and abrupt ecosystem turnover toward a wetter state in the last two centuries that exceeded changes recorded over the last three millennia. Contrary to expectations, there was no relationship between recent proxy changes and summer monsoon precipitation. Strong relationships, however, were found with winter climate data. We link this recent unprecedented wetness to marked warming at higher elevations resulting in increased seasonal runoff and associated climatic feedbacks in this snow and ice-melt dominated region. In contrast to the expected desiccation and decomposition of most peat systems with warming, this site has instead become the wettest in its ca. 3500-year history.

1. Introduction

[2] The Asian Monsoon region, including the Indian Ocean Monsoon (IOM) region, has experienced many natural wet and dry periods throughout the Holocene epoch [Phadtare, 2000; Gupta et al., 2003]. Unlike longer Holocene records from the Asian monsoon region, high-resolution paleoclimate studies of the late Holocene suggest a high degree of spatial variability at these short time scales [Morrill et al., 2003]. Recent warming has accelerated the melting of Himalayan glaciers, which in turn has escalated the frequency of catastrophic floods and landslides. Given that this region is densely populated, the socio-economic repercussions of climatically-driven environmental changes are of great consequence. Although monsoon rains undoubtedly have direct impacts on the regional environment, the contributions made through various climatic feedback mechanisms to the hydrological flux of this complex mountain environment should not be underestimated. Here we provide diatom- and pollen-based evidence from a peat record showing that the last three millennia alternated between wet and dry periods consistent with changes in monsoon intensity. However, over the last ca. 200 years, the magnitude of change exceeded all previous intervals, with an ecosystem turnover to a much wetter state. We link this unprecedented increase in wetness to exceptional warming at higher elevations that has accelerated seasonal melt-water flux and has both directly and indirectly affected the moisture regime of the Pinder Valley.

2. Field Area

[3] The study site is located at an altitude of 2650 m a.s.l. in the Pinder Valley (30.05°N, 79.93°E), Kumaon Higher Himalayas, northwest India (Figure 1). The site is relatively isolated and has not been directly disturbed by humans, as no evidence exists either historically or in recent times. Agricultural activities have been restricted to lower elevations (up to ca. 1200 m a.s.l.). The local climate is strongly influenced by glaciers at higher elevations (>3800 m a.s.l.). In the spring (March to May), the peat site receives run-off from the melting of winter snow along the catchment of the valley slope, at which time it behaves like a hydrologically open, small pond. This area receives ∼1100 mm of annual precipitation mainly from the summer monsoon rains (July to September) at which time it is flooded. This fen peat, which developed in a small depression on the north face, receives inflow from the surrounding catchment through hydrological and erosional processes during the wet seasons. Sediment accumulation is derived predominantly from allogenic detrital organic materials washed in from the surrounding oak-dominated mixed forest.

Figure 1.

Map of northwestern India showing location of the peat site at Dhakuri, the nearest glaciers, the Mukteshwar climate station, and the watershed boundary.

3. Methods

3.1. Chronology

[4] The 125 cm-long peat profile was trenched and sampled continuously at 2.0 cm intervals. Accelerator Mass Spectrometry (AMS) techniques were used to radiocarbon date the bulk peat samples from selected intervals and the more recent sediments were analyzed for 210Pb activity through gamma spectrometry (Table S1 in the auxiliary material).

3.2. Proxy Analysis

[5] For each peat interval, pollen, diatoms, grass phytoliths, % organic matter (estimated through loss-on-ignition (LOI) at 525°C), and magnetic susceptibility were analysed. Pollen was extracted by conventional laboratory techniques [Moore et al., 1991]. A minimum of 800 pollen grains were counted and identified for each sample. Pine pollen was taxonomically identified to the level of species (Pinus wallichiana) as this is the only pine species that grows at this altitude in the study region. Diatom samples were prepared using standard techniques [Battarbee et al., 2001] and for each sample a minimum of 400 diatom valves were identified and enumerated. Pollen and diatom species were expressed as percent relative abundances. Phytoliths were enumerated and expressed as a percentage relative to all other siliceous microfossils.

[6] Detrended correspondence analysis (DCA) [Hill and Gauch, 1980] was used to summarize the major trends in the diatom and pollen data. Detrended canonical correspondence analysis (DCCA) [ter Braak, 1986], constrained to age/depth for the last ca. 250 years, was used to quantify the degree of diatom compositional turnover or units of beta-diversity, following the methods used by Smol et al. [2005]. Sample scores for DCA axes 1 and 2 (DC1, DC2) and DCCA are scaled in standard deviation units (SD) of compositional turnover [Birks, 1998] thereby making comparisons of the magnitude and direction of change among the different proxies within the stratigraphical sequence possible. To determine how the diatom changes we report in this study compare objectively to undisturbed reference sites, we used the Smol et al. [2005] suite of reference sites (i.e., 14 sites from Canada, Ireland and Scotland), as comparable reference data do not exist for the Himalayan region. Beta-diversity measures were not calculated for pollen as reference data for pollen do not exist at present.

[7] Diatom and pollen diversity were calculated for each sample using the Hill's N2 index of diversity [Hill, 1973]. Magnetic susceptibility analyses included low frequency magnetic susceptibility (χlf) and anhysteretic remanent magnetization (χARM), which was grown by a constant biasing field of 0.1 mT simultaneously with peak alternating field (a.f.) demagnetisation of 100 mT at the decay rate of 0.01 mT per cycle using a Molspin a.f. demagnetizer with an ARM facility. Changes in magnetic susceptibility were used as an aid for interpreting past changes in the influx of detrital magnetic material.

3.3. Climate Records

[8] Temperature and precipitation records, available since AD 1871 from the nearby weather station at Mukteshwar (2310 m a.s.l.; 29.47°N, 79.65°E), together with Indian Ocean sea surface temperature (SST) anomalies available since AD 1860, were used with a 15-year running mean to facilitate comparisons to our proxy records.

4. Results and Discussion

4.1. Multiple Indicators: Local and Regional Scales

[9] Our independent environmental proxies register a striking, abrupt, and simultaneous shift toward a wetter state in the last ca. 200 years following three millennia of asynchronous diatom and pollen alternations between wet and dry periods (Figure 2). Diatoms have fast turnover rates and live within the peat environment, whereas pollen reflects regional changes in slower growing terrestrial vegetation. Collectively these differences often result in the asynchronous nature of these two environmental indicators at the Pinder Valley site.

Figure 2.

Stratigraphic profiles of selected peat proxy indicators and corresponding climatic interpretations. The most common diatom and pollen taxa are expressed in percent relative abundances. Phytolith relative abundances were calculated relative to the sum of all siliceous microfossils. Lines depict diatom and pollen zones based on cluster analysis using constrained incremental sum of squares. Radiocarbon dates were calibrated (see Table S1 for details) and are presented in years AD [Anno Domini] in brackets and italicized and as years BP (before present) outside the brackets. 210Pb are on the left of the figure and are given in years AD. Dates with an asterisk were interpolated or extrapolated. The scales of Alnus and grasses have been exaggerated 3X along the x-axis for purposes of clarity.

4.2. Last Three Millennia

[10] From ca. 3500 cal. yr BP to approximately the turn of the 19th century, our peat proxies tracked fluctuations between wet and dry periods (Figure 2), largely consistent with paleoclimatic evidence for changes in monsoon strength. For example, the driest period recorded in the last three millennia occurred at the base of the sequence (ca. 3500 to 2300 cal. yr BP) where diatoms were absent and grass, and oak pollen predominated, corroborating evidence for a particularly weak monsoon phase widely reported throughout the IOM region [Van Campo et al., 1996; Gupta et al., 2003].

[11] An increase in Abies pindrow and Picea smithiana ca. 2300 cal yr BP and a further increase ca. 1570 cal. yr BP indicate that regional climate became wetter and warmer. Diatoms appear at ca. 2300 cal. yr BP (Figure 2) and together with an increase in LOI (Figure 3), provide further evidence for the end of a very dry phase and signal the development of a more hydrologically open aquatic system. Meanwhile, the abundance of grass phytoliths, which are typically deposited in situ, indicates that the surrounding area remained dry, although wetter than previous intervals. The grass phytolith trend matches the grass pollen trend, indicating that the grass pollen is of local origin.

Figure 3.

Data summarizing the main directions of change in our environmental proxies. Shaded areas correspond to climate interpretations from Figure 2. DC1 and DC2 refer to trends in the detrended correspondence analysis (DCA) for axes 1 and 2, respectively. Changes in the DCA plots are given in standard deviation (SD) units and are indicative of the magnitude of species turnover through time. N2 is a measurement of species diversity (Hill's N2) that uses the number of very abundant species and is measured in units of species numbers. LOI (loss on ignition at 525 °C) estimates the % organic matter content (OMC) in the sediment. Magnetic susceptibility measurements include low frequency magnetic susceptibility (χlf) and the anhysteretic remanent magnetic (χARM). 210Pb and 14C dates are given in years AD with calibrated radiocarbon dates in italics. For further details on dating, see Table S1.

[12] Following ca. 1570 cal. yr BP, the shallow water environment developed into a waterlogged fen as indicated by the steady increase in LOI in tandem with a progressive shift to a diatom assemblage characterized by higher relative abundances of epiphytic taxa consisting of species often associated with mosses and fluctuating moisture levels (Figures 2 and 3). An increase in grass and Quercus semecarpifolia pollen following ca. 760 cal. yr BP signal a shift to cooler and drier climatic conditions on a regional scale.

[13] An abrupt shift in the diatoms (ca. AD 1300) and pollen (ca. 1355 cal AD), clearly summarized by DCAs (Figure 3), indicates a switch to regionally wetter and warmer conditions and increased moisture in the peat environment. Pollen indicative of dry conditions decreased, while pollen indicative of wetter conditions increased (Abies pindrow, Picea smithiana). An unidentified benthic diatom species, referred to in this study as Staurosira species ‘pinder’, decreased to trace abundances and was replaced by a more complex diatom assemblage that prominently included Navicula minima, an epiphytic taxon whose increase here is consistent with warmer conditions. An increasing trend in χlf (i.e., inorganic detrital influx) and a distinct peak in χARM, together with a decrease in LOI (Figure 3), suggest increased delivery of authigenic magnetic and clastic material into the site. Collectively, the magnitude and timing of these indicator shifts suggest a pronounced environmental change consistent with increased precipitation and wind speed (i.e., strengthened monsoon), evidence of which is widespread throughout the Asian Monsoon region at this time [Thompson et al., 2000; Morrill et al., 2003].

[14] Following ca. AD 1525 until ca. AD 1563, epiphytic diatoms returned to dominance (Figure 2) and together with the decrease to trace abundances of Navicula minima suggests that the environment had become cooler and characteristic peat conditions developed. Tracking this return to a more typical peat environment (i.e., accumulation of organic matter), LOI once again increased (Figure 3). Lagging the diatoms by ca. 20 years, pollen assemblages depict cooler conditions in the regional environment with an increase in Abiespindrow as well as a decrease in wetness indicated by the decrease in Picea smithiana and Pinus wallichiana abundances (Figure 2). Inorganic detrital influx (χlf), although still relatively high, was decreasing as was χARM during this interval, suggesting that the delivery of authigenic magnetic materials into the site had decreased (Figure 3).

[15] Circa AD 1600 to AD 1730, pollen assemblages continued to record a cold and moist environment. Diatom diversity sharply decreased (Figure 3) with the assemblage now dominated by Staurosira species that, for this period only, included Staurosira construens var. venter (Figure 2), a diatom commonly found in high abundances in colder environments [Douglas and Smol, 1999]. This cold period recorded in both the pollen and diatom data is consistent with western Himalayan tree-ring-based spring temperature reconstructions [Yadav and Singh, 2002], indicating that the first half of the 1600s was the coolest period of the last ca. 400 years.

[16] The extended cool and moist period expressed by our proxy indicators starting ca. 400 years ago and ending near the turn of the 19th century (Figure 2) corresponds with the period that is commonly referred to as the “Little Ice Age” (LIA) [e.g., Barlow, 2001]. Although this period of cooling has been documented in parts of Europe and to some degree in North America, there is no clear consensus for a protracted and widespread cooling period during this time in the Asian Monsoon region, with variations recorded both in its occurrence and in the timing [Borgaonkar et al., 2002; Esper et al., 2002; Yadav and Singh, 2002; Cooke et al., 2003]. Our proxy evidence for a protracted cool and moist event starting ca. 400 years ago potentially extends the geographic range of the LIA to this area of the Himalayas. Further studies are needed to substantiate this cooling trend in the IOM region.

4.3. Last ∼200 Years

[17] In the last ca. 200 years, both regional and local within-peat environmental proxies registered a striking, abrupt and synchronous shift toward a wetter state (Figure 2). This recent synchronicity suggests that, for the first time in this record, a major ecological threshold had been crossed, where conditions had changed to such a degree that its impacts were simultaneously and clearly registered in both the regional vegetation (pollen data) and within the local peat environment (diatom data).

[18] The magnitude of the abrupt post-Little Ice Age change is particularly evident by the degree of diatom compositional turnover (DCCA, DC1 and DC2) and likewise in the diatom diversity profile (Figure 3). Beta-diversity changes, estimated by DCCA analysis in the Himalayan diatoms, exceeded the mean reference value (1.0 SD unit) calculated by Smol et al. [2005], with a compositional turnover of 1.74 SD units. According to the criteria used by Smol et al. [2005], a compositional turnover exceeding 1.5 SD units would be classified within the highest category of ecological change and be indicative of a marked community shift.

[19] In the context of past diatom changes, the recent wet period is the most prominent, with an almost complete shift from an assemblage of benthic and epiphytic species commonly associated with mosses and indicative of fluctuating moisture regimes, to one dominated by tychoplanktonic Aulacoseira alpigena, a mostly free-floating diatom typically found in deeper waters. Synchronous with the diatom shifts, changes in the pollen relative abundances are abrupt, with a clear decrease in Abies pindrow and an increase in Picea smithiana, also indicating warmer and wetter conditions (Figure 2). Organic matter content (LOI) increased to 80%, which was likely the result of a substantial increase in organic influx from the surrounding catchment (Figure 3). For the first time in the peat record, grass phytoliths decrease to trace levels. Collectively, these data provide strong evidence that the peat environment has undergone an abrupt and unparalleled shift to wetter conditions within the last ca. 200 years.

[20] The Pinder Valley peat record disagrees with tree-ring based spring temperature [Yadav and Singh, 2002] and precipitation [Singh et al., 2005] reconstructions from nearby sites at similar altitudes. Undoubtedly, summer monsoon precipitation plays a major role in the hydrology of the study area, but other seasons and, similar to lacustrine systems in the Tibetan Himalayas [Morrill, 2004], other sources of water may also be important contributors to the moisture regime of the Pinder Valley.

[21] The last ca. 100 years of our proxy data were compared to available instrumental records from the Mukteshwar climate station as well as IO SST anomalies (Figure 4). The shift toward wetter conditions expressed in the Pinder Valley peat record was not clearly related to an increase in Indian summer monsoon rainfall, as might be intuitively expected. For example, strong negative relationships were found between summer monsoon temperature anomalies and IO SST anomalies (r = −0.72) (Figure 4a), as well as between summer monsoon precipitation and our wetness indicator, Aulacoseira alpigena (r = −0.75) (data not shown). An increase in IO SST would enhance surface evaporation, increase moisture fluxes to the region, and generally result in increases in monsoon rainfall [Li et al., 2001]. As such, we would expect a positive relationship between IO SST anomalies and summer monsoon precipitation data. However, unlike temperature, no relationship was found between monsoon precipitation and IO SST (r = 0.05) nor between monsoon precipitation and pollen and diatom compositional changes (DC1 and DC2). On the other hand, winter (NDJF) climate station data were clearly related to diatom compositional changes (DC1) with a strong negative relationship to winter temperature anomalies (r = −0.88) and a clear positive relationship with winter precipitation anomalies (r = +0.54). These relationships suggest that winter climate in this region plays a key role in the hydrological flux to the peat environment and that warmer winter temperatures, like those recorded at the Mukteshwar station following ca. AD 1940 (Figure 4b) were particularly important.

Figure 4.

Seasonal climate anomalies (relative to the 1876 to 1995 monthly climatology mean) from the Mukteshwar station (29.47°N, 79.65°E) and Indian Ocean Sea Surface Temperatures (SSTs) over the last ca. 100 years. A 15-year running mean was applied to all climate data. (a) A comparison of Indian Ocean SSTs (1846 to 2001) trends with summer monsoon (JAS) temperature anomalies. SSTs anomalies are in hundredths of a degree Celsius. (b) Winter (NDJF) temperature anomalies. (c) Summer monsoon precipitation anomalies. (d) Winter precipitation anomalies.

[22] The Pinder Valley is surrounded by glaciers at higher elevations and the valley slopes are covered with winter snow. Although melt-waters originating from the glaciers have no direct contact with this peat site, the recent acceleration of glacial melting, together with winter snow melt, may have increased humidity in the valley and resulted in localized cooling. Warmer temperatures have been known to profoundly impact seasonal runoff patterns in snow- and ice-melt-dominated regions [Barnett et al., 2005] such as the Pinder Valley, with an increase in moisture and localized cooling. It is therefore conceivable that increasing temperatures at higher elevation sites play a substantial role in contributing to the hydrological flux of lower elevation sites.

[23] Pronounced warming at higher elevation sites, which commenced ca. AD 1800 and became more prominent after ca. AD 1950 [Anderson et al., 2002; Thompson et al., 2003], has resulted in substantial glacier retreat throughout the Himalayas [Liu and Chen, 2000; Dobhal et al., 2004; Sharma, 2004], similar to melting observed in other mountainous regions [e.g., Meier and Dyurgerov, 2002; Thomspon et al., 2005]. Many of these Himalayan glaciers currently have a negative mass balance and face terminal retreat due to the lack of new ice formation [Kulkarni et al., 2004]. For example, dated terminal glacial moraines depicted in satellite images from NASA ( = 16584) indicate that the Gangotri Glacier, one of the largest in the Himalayas, has receded substantially since AD 1780. Similar to many other glaciers in the region [e.g., Sharma, 2004], this retreat was slower prior to the mid-1900s, with rapid wasting thereafter.

[24] Substantial post-AD 1950 changes in our peat record (Figure 2) are concurrent with marked changes in temperature and precipitation data recorded throughout many regions of the Indian Himalayas [e.g., Kumar et al., 1994; Yadav et al., 2004] including those at the Mukteshwar climate station (Figure 4). Since ca. AD 1950 climatic warming over the Tibetan Plateau has been unmistakable at sites located at elevations greater than 4000 m a.s.l., with the degree of warming increasing with elevation [Liu and Chen, 2000; Thompson et al., 2003]. As well, Indian Ocean SSTs have recorded a sharp increase from the late 1950s to the present (Figure 4a). At the Mukteshwar station, however, these above average temperatures were only evident in the winter season, and like many lower elevation sites [e.g., Shrestha et al., 1999; Singh and Sontakke, 2002], cooler temperatures were recorded both annually and during the summer monsoon period (Figures 4a and 4b) suggesting local negative feedback mechanisms. In addition to the hydrologically-linked processes previously outlined, there are a number of possible inter-related mechanisms believed responsible for this more recent (post ca. 1950) cooling trend stemming from a wide-ranging rapidly growing population, increased atmospheric inputs, increased detrimental agricultural practices, large-scale deforestation and consequential soil degradation throughout India and parts of Asia that have collectively resulted in negative climatic feedback effects upon the surrounding environment [Menon et al., 2002; Yadav et al., 2004]. Whatever the mechanism behind this recent cooling trend, it has undoubtedly played an important role in maintaining wet conditions in the Pinder Valley.

5. Conclusions

[25] High-resolution, multi-proxy, paleoecological data from a Central Himalayan peat core archive a series of climatically-induced environmental changes over the last three millennia. However, these changes were dwarfed in comparison to the striking ecosystem change to a wetter state recorded over the last two centuries. The magnitude and abruptness of this recent shift suggests an ecological threshold had been crossed, as both the regional vegetation and the local peat environment recorded a synchronous change for the first time in the last three millennia. The magnitude of this change is particularly evident in a nearly complete compositional turnover to a diatom assemblage more suited to an open water environment. Ultimately, these changes were triggered by recent global warming and associated negative feedbacks on the climate system; contrary to expectations, we found no relationship between our recent proxy changes and summer monsoon precipitation. One well-documented repercussion of warming is that many peat deposits become drier, to the extent that some stop accumulating and begin to decompose. Our study provides an example of the opposite effect, namely that with warming, this site has become the wettest in its ca. 3500-year history. Clearly, however, once mountain glaciers melt completely, as some are predicted to do in the coming decades [Sharma, 2004], and climate warming continues, the supply of water from snow and glaciers to downslope sites will cease.


[26] We are grateful to the two anonymous reviewers whose insightful comments greatly strengthened our inferences and our manuscript. We thank members of PEARL for improving the quality and clarity of this manuscript, to Piyoosh Rautela for his help in peat trenching and sample collection, and to B.R. Arora for providing working facilities at WIHG, India. This study was supported by NSERC Canada, and the Department of Science and Technology, Government of India.