G 3. AN ELECTRONIC JOURNAL OF THE EARTH SCIENCES Published by AGU and the Geochemical Society

Transcription

1 Geosystems G 3 AN ELECTRONIC JOURNAL OF THE EARTH SCIENCES Published by AGU and the Geochemical Society Article Volume 8, Number 8 16 August 2007 Q08008, doi: /2007gc ISSN: Temporal variation in Sr and 87 Sr/ 86 Sr of the Brahmaputra: Implications for annual fluxes and tracking flash floods through chemical and isotope composition Santosh K. Rai and Sunil K. Singh Physical Research Laboratory, Navrangpura, Ahmedabad, India (rksant@prl.res.in; sunil@prl.res.in) [1] Temporal variations in dissolved Sr and its 87 Sr/ 86 Sr in the Brahmaputra River were determined for the first time by analyzing biweekly samples collected over a period of about a year at Guwahati. Sr in the Brahmaputra range from to mm with 87 Sr/ 86 Sr of to Sr concentration decreases in monsoon (except for the 15 June 2000 sample, which had anomalously high concentrations of Sr and major ions) compared to the nonmonsoon period. The decrease in elemental abundances, a factor of 2, is not proportional to the increase in discharge, an order of magnitude, indicating enhanced weathering during monsoon. This can be a cumulative effect of an increase in drainage area and in physical weathering during monsoon. 87 Sr/ 86 Sr and other proxies of silicate weathering show a relatively lower contribution from silicate weathering to the major ion and Sr budget during monsoon with a concomitant increase in carbonate weathering contribution. Shorter interaction time between water and minerals during monsoon coupled with the slower weathering kinetics of silicate compared to carbonate can be contributing factors to the seasonal changes in their relative contributions to the major ion budget. This study shows that the annual fluxes of various ions, SCat sil, and Sr sil calculated on the basis of measured biweekly concentrations and monthly water discharge data are roughly within ±35% of those calculated using either the lowest or the highest concentrations and annual water discharge. This implies that a tropical region such as in India, where river discharge is governed mostly by monsoon rains, the annual elemental fluxes measured during monsoon would be a good approximation for their discharge weighted annual fluxes; however, an additional sampling during a drier period would be required to understand the weathering processes. The geochemistry of the 15 June 2000 sample is anomalous. The chemistry of this sample is similar to those from the Tibetan region, an inference consistent with reports of a flash flood in the Brahmaputra due to a natural dam burst in the Yigong River of Tibet. These results highlight the use of the chemical and isotopic composition of water samples to track and quantify the source and discharge of a flash flood. Components: 6514 words, 8 figures, 3 tables. Keywords: Brahmaputra; temporal variation; 87 Sr/ 86 Sr; weathering; flash flood. Index Terms: 0790 Cryosphere: Weathering (1625, 1886); 0330 Atmospheric Composition and Structure: Geochemical cycles (1030); 0454 Biogeosciences: Isotopic composition and chemistry (1041, 4870). Received 15 February 2007; Revised 31 May 2007; Accepted 13 June 2007; Published 16 August Rai, S. K., and S. Singh (2007), Temporal variation in Sr and 87 Sr/ 86 Sr of the Brahmaputra: Implications for annual fluxes and tracking flash floods through chemical and isotope composition, Geochem. Geophys. Geosyst., 8, Q08008, doi: / 2007GC Copyright 2007 by the American Geophysical Union 1 of 14

2 1. Introduction [2] During the past couple of decades there have been a number of studies on the chemical weathering of the Himalaya based on the chemistry and Sr isotope composition of rivers flowing through it [Sarin et al., 1989, 1992; Krishnaswami et al., 1992, 1999; Blum et al., 1998; Harris et al., 1998; Singh et al., 1998; Galy et al., 1999; Galy and France-Lanord, 1999; Dalai et al., 2002; Bickle et al., 2003]. Many of these studies, particularly the earlier ones, rely on the analyses of a suite of samples collected along the course of the river during a single or a few selected seasons; though it is well known that the discharge of many of these rivers show significant variation during the year, primarily due to monsoon rainfall. The sources of water to these rivers are rainfall, glaciers and groundwaters, the dominant being rainfall, 80% of the total discharge of these rivers is during June September, the period of the summer monsoon. The uneven spatial and temporal distribution of rainfall and hence the river discharge can contribute to temporal variation in erosion rates over the drainage basin. Further, as the lithology over the drainage basins in the Himalaya varies spatially, the distribution of rainfall can also influence the chemical and isotopic composition of the rivers which integrate large drainage areas. Heavy localized rainfall, such as cloud burst, a common feature in the Himalaya, can also cause flash floods in a particular tributary or drainage subbasin, which in turn can affect the chemical and isotopic composition of the mainstream. All these promote the study of temporal variations in elemental fluxes over an annual cycle and the various components controlling to it. [3] More recently, a few studies have been reported [Krishnaswami et al., 1999; Galy and France-Lanord, 2001; Bickle et al., 2003; Tipper et al., 2006] which address to the temporal variation in chemistry of selected rivers from the Himalaya and its implications to silicate-carbonate weathering in the basin. Krishnaswami et al. [1999] on the basis of a synthesis of available data inferred that there is selective increase in carbonate weathering contribution (with respect to silicates) in the headwaters of the Ganga in monsoon, due to increase in physical erosion. Bickle et al. [2003] and Tipper et al. [2006] from monthly sampling of the Ganga and weekly sampling of the Marsyandi rivers in the Himalaya observed that carbonate weathering contribution is more enhanced during monsoon due to the higher weatherability of carbonates compared to silicates. In contrast, during drier season silicate weathering contribution to river water chemistry is more relative to that in monsoon, a likely reason being increased discharge from groundwater to rivers. [4] The present study is on the role of temporal variation in discharge on the erosion pattern in the Brahmaputra Basin. During the course of this study a flash flood was reported in the Brahmaputra [International Centre for Integrated Mountain Development (ICIMOD), 2001; Tewari, 2004] resulting from a dam burst in the Yigong River, a tributary of Tsangpo in Tibet. This study also addresses the potential of using major ion and Sr abundances and Sr isotope composition to independently track the source of a flash flood and to determine its impact on the water chemistry of the Brahmaputra mainstream and ensuing major element fluxes to the Bay of Bengal. The study forms a part of a comprehensive investigation on the Brahmaputra River System [Singh and France-Lanord, 2002; Singh et al., 2003, 2005, 2006; Garzanti et al., 2004; Singh, 2006]. 2. Brahmaputra Basin [5] The Brahmaputra originates from Kailash Mountain and flows east in Tibet following the Indus Tsangpo Suture (Figure 1). Around the Namche Barwa, it takes a U-turn and enters India. Near Pasighat it enters the Assam Plain and flows in EEN to WWS till Indo-Bangladesh border, where it turns south and joins the Ganga to form Padma. Along its course many tributaries join the Brahmaputra from north, east and south. These include the Dibang, the Lohit, the Subansiri, the Burhi Dihing, the Jia Bhareli and the Manas. The Brahmaputra receives most of its discharge from SW monsoon during the months of May to October. The NE monsoon rains during October December [Krishnaswami and Singh, 2005] also contribute to its water discharge. Other sources of water to the Brahmaputra are glaciers and groundwater. The large temporal variation in the water discharge of the Brahmaputra is evident from its hydrograph at Guwahati (Figure 2). [6] The geology of the Brahmaputra drainage system is quite diverse. The Brahmaputra and its tributaries flow through Tibet, Transhimalaya, Higher and Lesser Himalaya. Its drainage basin comprise of granites and gneisses, carbonates, calcalkaline rocks and dispersed alkaline and saline salts (Figure 1) [Singh et al., 2003, 2005, 2006]. 2of14

3 Geosystems G 3 rai and singh: sr and 87 sr/ 86 sr in the brahmaputra /2007GC Figure 1. The Brahmaputra River System. The locations of the biweekly samples, Guwahati, and that of the flash flood are marked by orange circles and crosses, respectively. Figure 2. Monthly water discharge of the Brahmaputra at Guwahati (data from Global Runoff Data Centre; available at The plot is based on the discharge for the years 1956 to Water discharge of the Brahmaputra varies by an order of magnitude between monsoon and nonmonsoon periods. About 80% of water in the Brahmaputra flows during monsoon, May to October. 3of14

4 Table 1. Sr, 87 Sr/ 86 Sr, and Major Ion Composition of the Brahmaputra Mainstream at Guwahati a Ca 2+, Cl, Date Na + Na* K + Mg 2+ mmol L 1 mmol L 1 2 SO 4 HCO 3 SiO 2 Sr 87 Sr/ 86 Sr Spiked 87 Sr/ 86 Sr Unspiked TDS, mg L 1 24 Oct Nov Nov Dec Dec Jan Jan Feb March March April April April May May June July July a Sr and 87 Sr/ 86 Sr from present study; major ion data from Singh et al. [2005]. 4of14

5 Figure 3. Biweekly variation in (a) Sr concentration and 87 Sr/ 86 Sr and (b) Ca, SO 4, HCO 3, and TDS in the Brahmaputra at Guwahati. The major ion concentrations vary by a factor of about two over the sampling period despite one order magnitude variation in the discharge. Barring the sample of 15 June, monsoon samples are diluted in all major ions. 87 Sr/ 86 Sr of the monsoon samples are also lower, indicative of a higher contribution of Sr from less radiogenic lithologies, carbonate/evaporites. The 15 June sample has elevated concentrations of most of the ions and Sr. This sample has its origin from a flash flood in Tibet (see text). The shaded band drawn through each property is a rough measure of seasonal trend. Numerous hot and cold springs present in Tibet and in the Himalaya also contribute to the chemistry of the Brahmaputra [Singh et al., 2006]. 3. Materials and Methods 3.1. Sampling [7] Sr and 87 Sr/ 86 Sr analyses of the water samples were made in this study; the major ion data are from Singh et al. [2005]. Details of water sampling are given by Singh et al. [2005]. Briefly, the samples were collected biweekly for a duration of 10 months from October 1999 to July 2000 from the Brahmaputra mainstream at Guwahati, India. Approximately five liters of water samples were collected each time from the mid-channel. The samples were filtered in July 2000 using 0.2 mm nylon filters and stored in HDPP bottles. Anions, cations, alkalinity, Sr content, and 87 Sr/ 86 Sr were measured in the same filtered aliquots Methods [8] The cation and anion data of these waters are from Singh et al. [2005]. Sr isotope analyses of the samples were performed as a part of this study. The measurements were made on two aliquots, one was spiked with enriched 84 Sr and the other was un-spiked. The spiked aliquots provided both Sr 5of14

6 Figure 4. Scatterplots of major ion, TDS, Sr, and 87 Sr/ 86 Sr with discharge. The concentrations of major ions and Sr decrease with discharge. The anomalous nature of the 15 June sample (O) is clearly evident. concentration and their isotope composition, whereas the unspiked samples yielded only 87 Sr/ 86 Sr. For spiked measurements 12 ml filtered water samples were mixed with 84 Sr enriched Sr spike in a 15 ml Savillex 1 vial and kept tightly closed overnight at 85 C for equilibration. The samples were dried, redissolved in 3N Seastar 1 HNO 3 and Sr separated from the acid solution using Sr specific resin. In case of unspiked samples, 5 to 8 ml of filtered water was taken in the Savillex vial, appropriate volume of 16N HNO 3 was added to make this 3N in HNO 3. The acidified water was passed through Sr specific resin to separate Sr. Purified Sr extracts were loaded on Re filament with Sr activator and the Sr isotope composition measured using ISOPROBE-T Thermal Ionisation Mass Spectrometer in static multicollection mode. Measured 87 Sr/ 86 Sr ratios are corrected for instrumental mass fractionation using 86 Sr/ 88 Sr as Along with samples NBS 987 standard were measured several times which yielded 87 Sr/ 86 Sr of ± (2s). Internal precision of analysis was always better than ±10 ppm (1 s m ). The 87 Sr/ 86 Sr of spiked and unspiked samples show good agreement (Table 1). The entire procedural blanks vary between 200 and 650 pg with 87 Sr/ 86 Sr of The total Sr analyzed for these samples were always 1 mg, 6of14

7 Figure 5. The plot of temporal variation in ratios of Ca, Mg, HCO 3, and SO 4 normalized to Na* (Na* = Na r -Cl r ). The data seem to fall into two groups, a nearly uniform lower value during November March, which increases to a higher fluctuating value during May August. The data of the 15 June sample are not plotted. orders of magnitude higher than the blank and hence the data are not corrected for blank. 4. Results [9] Major ion data of the samples [Singh et al., 2005], their Sr content and 87 Sr/ 86 Sr are given in Table 1 and Figures 3a and 3b. Total dissolved solids (TDS) of these samples vary by a factor of two, from 102 to 203 mg L 1 (Table 1) with the highest during June and the lowest in October. These temporal variations can be attributed to dilution effect resulting from increased discharge during monsoon, though the decrease is not in 1:1 proportion to increase in water discharge. The sample collected on 15 June 2000 deviates from this general trend [Singh et al., 2005]. The anions and cations also show variations by a factor of about two. Sr concentrations of the samples also vary by a factor of 2, from 604 nm in October 1999 to 1392 nm in June The temporal variations in Sr are similar to those of other cations and anions with lower value in monsoon and higher concentration in nonmonsoon, except for the 15 June 2000 sample (Table 1). The 15 June sample has anomalously elevated concentrations of most of the dissolved constituents. For K, Ca, SO 4 and TDS this sample has the highest concentration compared to others analyzed. The most anomalous concentration is for SO 4, this sample is 4 times enriched compared to other monsoon samples. [10] 87 Sr/ 86 Sr of the samples show measurable temporal variation over the year, it ranges from in May to in October. 87 Sr/ 86 Sr of the Brahmaputra water at Guwahati decreases in monsoon compared to other seasons. 5. Discussion [11] The time variation in the concentrations of major ions and Sr of the Brahmaputra at Guwahati is shown in Figure 3. Cations, anions and Sr all decrease with discharge (Figure 4) and reach near uniform value at high discharge. It is, however, seen that the dilution is not in 1:1 proportion to increase in water discharge, the concentrations of various elements decrease by factor of only about two, compared to an increase in the water discharge by about an order of magnitude. This can be due to availability of more drainage area for weathering during monsoon, unlike in nonmonsoon when the river shrinks. During nonmonsoon period, a significant fraction of the exposed surface in river drainage may remain out of contact with river water and hence be unavailable for interactions. In contrast, during monsoon most of the drainage basin comes in contact with river water and thus provides more surface area for chemical weathering. Further, during monsoon, physical weathering also increases, which in turn can promote more chemical weathering. 87 Sr/ 86 Sr of these samples also show a decreasing trend with discharge (Figure 4). Lower 87 Sr/ 86 Sr of the Brah- 7of14

8 Figure Sr/ 86 Sr versus Ca/Sr in the biweekly samples and the various end-members contributing Sr to the Brahmaputra. The 87 Sr/ 86 Sr data cluster around the two end-members, one the nonsilicates (carbonates/evaporites) and the other the TPBSIL, suggesting that Sr in these waters is derived primarily from carbonates and evaporites and from silicates of the Transhimalayan Plutonic Belt. The inset shows that 87 Sr/ 86 Sr in the monsoon samples are lower than that of drier months and the 87 Sr/ 86 Sr is negatively correlated with Ca/Sr. The results are consistent with the hypothesis that carbonates /evaporites Sr contribution during monsoon is higher. maputra in monsoon can be a result of relatively higher contribution of Sr from carbonates during this season, an inference consistent with the major ion data which show a concomitant increase in Ca/ Na*, Mg/Na*, HCO 3 /Na* (Figure 5). 87 Sr/ 86 Sr of the river water shows a marginal increasing trend during monsoon after an initial dip. This trend can be explained by increase in silicate weathering due to increase in physical erosion during monsoon [Millot et al., 2002]. As silicate minerals weather more slowly compared to the carbonates, their impact on water chemistry becomes observable after a time lag. However, carbonate weathering contribution is relatively enhanced compared to silicate contribution in monsoon compared to nonmonsoon period. In contrast, the SO 4 /Na* shows only a marginal change with time implying that the weathering kinetics of the lithologies supplying SO 4 and Na* during monsoon and nonmonsoon seasons are similar. The nonmonsoon values of Ca/ Na*, Mg/Na*, HCO 3 /Na* are quite uniform, whereas those for the monsoon show significant fluctuations, possibly due to variation in intensity and spatial pattern of the rainfall during monsoon. The sample collected on 15 June 2000 is unique compared to others; this aspect will be discussed later Sources of Sr to the Brahmaputra [12] Singh et al. [2006], on the basis of Na and Sr data of rivers from the Brahmaputra system (mainstream and tributaries) and those of silicate endmembers, inferred that silicate weathering accounts for 50% of Sr budget in the Brahmaputra and that 40% is derived from the Tibetan drainage. The results of Sr concentration and its isotope composition on the biweekly samples from the Brahmaputra at Guwahati (Table 1) are shown in a Ca/Sr versus 87 Sr/ 86 Sr scatterplot (Figure 6). The figure also includes the 87 Sr/ 86 Sr of various end-members [Singh et al., 2003] supplying Sr to the Brahmaputra. These end-members [Singh et al., 2006] are silicates of Transhimalayan Plutonic Belt (TPBSIL) and of the Himalaya (HIMSIL), Carbonates of Tethyan Sedimentary Series (TSSCARB) and of the Lesser Himalaya (LHCARB) and evaporites (EVAP). The Ca/Sr of evaporites from the Tibetan drainage is not available; hence typical value reported for evaporites from other regions is used. Ca/Sr of evaporites from the Lesser Himalaya (our unpublished results) are similar to those of the carbonates from the Lesser Himalaya and hence these two members are plotted together in Figure 6. In addition to these sources, hot and cold springs present in the Tibetan drainage may also supply Sr to the Brahmaputra; however, in the absence of any data on hydrothermal discharges and their Sr and Ca from these regions, the significance of their contribution to Sr budget of the Brahmaputra cannot be assessed. The clustering of the data points just above the TPBSIL and TSSCARB 8of14

9 Figure 7. Temporal variation in (a) SCat sil and Sr sil and (b) SiO 2 /TDS. All these components decrease during monsoon, indicating a relatively lower contribution of silicate weathering during monsoon. (Figure 6) shows that these sources are dominating the supply of Sr to these waters, with additional contribution from evaporites and HIMSIL. Contribution of Sr from silicate decreases in monsoon. Inset of Figure 6 indicates an overall decrease in 87 Sr/ 86 Sr with increase in Ca/Sr. The Ca/Sr of monsoon samples are generally higher compared to those of nonmonsoon. Similar observations have been reported for the Marsyandi river in Nepal [Tipper et al., 2006] and for the Ganga at Rishikesh in India [Bickle et al., 2003]. These results can be interpreted in terms of increased contributions from lithologies with higher Ca/Sr. Carbonates and evaporites fulfill this requirement. Also their weatherability is higher than that of silicates, and hence they are prone to increased weathering during monsoon. Lower Ca/Sr in the Brahmaputra river water during nonmonsoon compared to monsoon period can also result from removal of Ca from the water if the water is supersaturated in calcite. It has, however, been shown [Singh et al., 2005] that the Brahmaputra water at Guwahati is undersaturated or is only marginally saturated in calcite during monsoon and nonmonsoon periods ruling out calcite precipitation as a cause to decrease Ca/Sr. Enhanced physical weathering and availability of larger drainage area for weathering during monsoon, both could promote carbonate/ evaporite weathering. Further, the rapid runoff during monsoon could also contribute to decrease in silicate weathering; as the time of contact between the silicate mineral phases and river water would be less. The inference that carbonate contribution is enhanced during monsoon, suggests that long term variability in monsoon intensity could influence the relative silicate and carbonate erosion. It could be inferred from the above finding that warm and humid monsoon climate 9of14

10 Figure 8. Scatterplot of SCat sil and 87 Sr/ 86 Sr. The data lie almost on a line connecting nonsilicates and TPBSIL, supporting the view that these end-members regulate the Sr isotopic composition of the water samples. This conclusion is consistent with that derived from Figure 6. Samples having higher SCat sil have higher 87 Sr/ 86 Sr, particularly during nonmonsoon. would favor enhanced contribution from carbonate weathering, whereas silicate weathering contribution will be increased during drier periods Temporal Variation in Silicate Derived Cations and Sr [13] One of the key issues in the weathering studies of the Brahmaputra drainage is the assessment of temporal variation in weathering, particularly the role of monsoon on silicate weathering rates. In this study an attempt has been made for the first time to quantify the temporal variations of the silicate derived 87 Sr/ 86 Sr and Sr to the Brahmaputra water. These fractions are calculated from the water data using the forward model [Krishnaswami et al., 1999; Singh et al., 2005, 2006]. The SCat sil and Sr sil are derived using following relations: %S Cat sil ¼ where h n 100 Na* þ K r þ CaþMg oi Na Na* sil ½Na* þ K r þ Ca r þ Mg r Š % Srsil ¼ n o 100 Sr Na Na* sil Sr r Na* ¼ Na r Cl r Subscript r represents the measured concentration of dissolved ions in river water. ((Ca + Mg)/ ð1þ ð2þ Na) sil and (Sr/Na) sil are the ratios in silicate endmembers in the Brahmaputra watershed, the values for these ratios are 2.25 ± 1.1 and 4.5 ± 2.3, respectively [Singh et al., 2005, 2006]. [14] The temporal variations in SCat sil and Sr sil are shown in Figure 7a. SCat sil and Sr sil over a ten month time period Show a wide range, 35 52% 38 62%, respectively. These estimates are based on the ((Ca + Mg)/Na) sil and (Sr/Na) sil of silicate end-members which are subject to 50% uncertainties [Singh et al., 2005, 2006]. Considering these uncertainties, estimated SCat sil and Sr sil are prone to 12 and 50% uncertainties, respectively. [15] Barring the sample collected on 15 June 2000, the SCat sil and Sr sil seem to show two groups. One group are samples collected during May to October, with SCat sil and Sr sil of 38 ± 3 and 43 ± 4%, and the other group are the samples from November to April, they have 47 ± 3 and 54 ± 3%, respectively. These results indicate that during monsoon, the SCat sil and Sr sil are lower, i.e., silicate weathering contribution to cations budget is relatively less than those during the drier months. This result is also attested by the decrease in SiO 2 /TDS during monsoon (Figure 7b). The anomalously lower SCat sil and Sr sil in the 15 June sample will be discussed in a later section. These seasonal trends in Sr sil are consistent with the inference derived on the basis of 87 Sr/ 86 Sr of the Brahmaputra at Guwahati that the silicate contribution to Sr budget decreases during monsoon. The 87 Sr/ 86 Sr of the samples do 10 of 14

11 Table 2. Comparison of Chemical and Isotopic Composition of Waters of the Tibetan and the Himalayan Drainage With Those of the 15 June Brahmaputra Sample a Tibetan Himalaya 15 June Sample TDS/SiO 2, mg/mmol 1.38 ( ) 0.53 ( ) 1.36 Sr, mm 2.09 ( ) ( ) Sr/ 86 Sr ( ) ( ) HCO 3 /total anions 0.7 ( ) 0.84 ( ) 0.63 SO 4 /SiO ( ) 0.45 ( ) 3.06 a Tibetan and Himalayan values are from Singh et al. [2005, 2006, and references therein]. Values in parentheses show the range. not seem to have any systematic trend with SCat sil ; however, in samples with SCat sil 40% 50% the 87 Sr/ 86 Sr clusters at a higher value of The 87 Sr/ 86 Sr of samples seem to determined predominantly by two end-members, the silicate member, TPBSIL and the nonsilicate end-member (Figure 8) Anomalous Sample (15 June 2000) of the Brahmaputra at Guwahati [16] The sample collected on 15 June 2000, as mentioned earlier, has anomalously higher concentrations of Ca, Mg, K, HCO 3,SO 4 and Sr relative to samples collected about two weeks before and after it. These cations and the anions are enriched in this sample by factor of 2 with respect to other monsoon samples, except SO 4, for which the enrichment is about a factor of four. The cause for this enrichment is discussed below. [17] A severe flash flood in the Tibetan drainage was reported on 10 June 2000 [ICIMOD, 2001; Tewari, 2004] which suddenly raised the water level in the Siang River at Pasighat on 11 June 2000 by 5 meters. The discharge of the Siang also showed a concomitant rapid increase, from m 3 /sec on the previous day to a value of m 3 /sec on 11 June [ICIMOD, 2001; Tewari, 2004]. This flood is attributed to [ICI- MOD, 2001; Tewari, 2004] a massive landslide that occurred in the Yigong River on the 9 April Enhanced melting of snow and ice led to a very rapid damming the Yigong River with landslide debris, soil, and ice, the dimension of the dam being 130 m thick, 1.5 km long and 2.6 km wide. In April 2000, the discharge of the streamflow into this naturally created lake was 100 m 3 /s; the water was rising at the rate of one meter per day. Due to the continuous accumulation of water and loose nature of the dam material, the dam burst on 10 June 2000, resulting in an extensive flash flood downstream. This flash flood reached Pasighat on the morning of 11 June [18] The discharge measurements of the Siang River at Pasighat reported for 11 June show a pulse, with a peak flux of m 3 sec 1. The pulse seems to have lasted for about one and half days and the total volume of water discharged during this period being L[Tewari, 2004]. From local news agencies and other reports ( however, it seems that this flood continued even on 15 June in Arunachal Pradesh and Assam. On 14 June, 1000s of villages were inundated, whereas the Brahmaputra was flowing one foot above the danger mark at Dhubri on 15 June. As discharge data is unavailable after 12 June it is unclear if there were multiple pulses of flood discharge. Thus the water sample collected at Guwahati on 15 June is expected to have a significant component of the flood water of Tibetan origin. The geochemical data are consistent with this. Comparison of the concentrations of Ca, Sr and HCO 3 and TDS/SiO 2, HCO 3 /total anions, SO 4 /SiO 2 ratios (Table 2) in the 15 June sample with very sparse data the Tibetan waters [Singh et al., 2006, Harris et al., 1998] show that they are similar. The overall similarity in the concentration of major ions in the two waters and their 87 Sr/ 86 Sr indicate that the source of flood waters in the Brahmaputra at Guwahati on 15 June 2000 was dominated by contribution from the Tibetan drainage. [19] The 87 Sr/ 86 Sr and major ion composition of the water provide an independent approach to estimate Tibetan drainage contribution to the flood water discharge. The calculation assumes that the chemical and isotopic composition of the Brahmaputra river water at Guwahati is a binary mixture of Tibetan water and water from the Himalayan drainage. The water flux is calculated on the basis of the mass balance approach using Sr concentra- 11 of 14

12 Table 3. Fluxes of Various Ions and TDS of the Brahmaputra Calculated on the Basis of Three Different Approaches a Flux Based on Na K Ca, Sr, TDS, 10 9 mol yr 1 SO 4 HCO mol yr ton yr 1 Brahmaputra@Guwahati b Biweekly data Lowest conc Highest conc Ganga@Rishikesh c Monthly data Lowest conc Highest conc a Biweekly data, flux based on data in Table 1 and monthly water discharge; Lowest conc., flux based on lowest concentration and annual water discharge; Highest conc., flux based on highest concentration and annual water discharge. The 15 June data are not included. July data have been used for June, August, and September. b From this study. c Calculated from Bickle et al. [2003]. tion and 87 Sr/ 86 Sr (Table 2) and the following relation: where f T ¼ R B R H R T R H F T ¼ F BC Sr B f T C Sr T and F B ¼ F H þ F T f T denotes the fraction of dissolved Sr from Tibetan drainage calculated on the basis of 87 Sr/ 86 Sr of different reservoirs. R is 87 Sr/ 86 Sr, F is water flux and C is Sr concentration. The subscripts T, H, B stand for water from the Tibet, the Himalaya and the Brahmaputra, respectively. Product of F and C represent the flux of Sr from a given reservoir. On the basis of the flux and fraction of Sr from different reservoirs, flux of water of Tibetan drainage is calculated using the above equations. The calculation using the data of the May 31 sample shows that the contribution of water from the Tibetan drainage to the Brahmaputra at Guwahati is 25%. In a normal year (i.e., without flash flood) the proportion of Tibet:Himalaya for the month of June is expected to be same as that of May. Using an average water flux of Lsec 1 for June (Figure 2), the contribution from the Himalayan drainage is calculated to be Lsec 1. Using the Sr values for the June 15 sample and equation (3), the Tibet flux is derived to be Lsec 1. The average discharge of the Brahmaputra at Guwahati reported for the month of May is Lsec 1, whereas during the flash flood on 15 June 2000, it was calculated to be Lsec 1, out of which ð3þ 55% of water was from Tibet. Tibetan fraction during the peak of the flash flood on 11 June could have been much higher compared to this estimate for 15 June. The flux of water calculated using geochemical signatures compares well with that of reported for the flash flood at Pasighat [Tewari, 2004]. Most of the uncertainties in deriving these discharges will arise due to uncertainties in endmember values of Sr concentrations of the Himalayan and Tibetan rivers, which are 38 and 12%, respectively. The water discharge estimates for the flash flood will have an uncertainty of 40% due to uncertainties Sr concentrations of Himalayan and Tibetan end-members. 6. Conclusions [20] Sr concentration and 87 Sr/ 86 Sr of biweekly water samples of the Brahmaputra River at Guwahati measured in this study along with their major ion data show considerable variations over a year and have brought out significant information on the temporal variations in major ion, Sr isotope abundance and the factors contributing to them. The concentration of all major ions and Sr gets diluted by a factor of about two during monsoon compared to the nonmonsoon period; however, the extent of dilution is far less than the magnitude of increase in water discharge during monsoon, resulting in higher erosion rates during this period. These results further suggest that runoff regulates erosion rate in this basin. 87 Sr/ 86 Sr of the Brahmaputra River also varies on a seasonal scale, with more radiogenic values during the nonmonsoon. 87 Sr/ 86 Sr along with Sr sil and SCat sil show that, contribution of Sr and other cations from carbonate weathering increases during monsoon attributable 12 of 14

13 to combined effect of their higher weatherability, increase in available drainage area and physical weathering in the basin during monsoon. This study shows that in the Brahmaputra contributions of silicate weathering to total cations and to Sr vary by a factor of 1.5 over a year and the concentrations of various ions varies by a factor two. The variation in SCat sil and Sr sil estimated on the basis of the forward model is ±30%. The average SCat sil and Sr sil for the sampling period of 10 months (excluding 15 June) are 43 ± 5 and 48 ± 7% compared to extreme values of 35 to 52 and 38 to 62%, respectively. Similarly, the fluxes of major ions based on biweekly data are within ±35% of the values calculated on the basis of annual water discharge and assuming that lowest and highest measured concentrations are average of the year (Table 3). These results suggest that elemental fluxes are dominated by monsoon, and accounts 80% annual flux of TDS and other ions. Reported data for the Ganga at Rishikesh also show similar trend. If such trend is applicable to all monsoon-fed rivers, then their average concentration based on a two season sampling, one during monsoon and another during drier period should provide reliable estimates of annual fluxes and associated silicate weathering contribution to the dissolved elemental budget and CO 2 consumption rates. [21] Another important finding of this work is the tracing of source of a flash flood based on geochemical signatures. It is shown that the anomalously high concentrations of various ions in the Brahmaputra water sample collected on 15 June 2000 is sourced primarily from the Tibetan drainage. The flood was a result of a dam burst in one of the tributaries of the Tsangpo. Using Sr concentration and 87 Sr/ 86 Sr of the flood water and those of the Himalayan and Tibetan end-members, the flux of water during the flash flood was estimated. Further, the flux of cations and TDS supplied during this flash flood was shown to vary from 4 7% of the annual flux from the Brahmaputra [Singh et al., 2005], whereas for Sr it is about one tenth [Singh et al., 2006]. This study underscores the use of Sr and major ion concentration and 87 Sr/ 86 Sr of the river water in tracking the source of flash flood and also in estimating the water discharge during such events. Acknowledgments [22] Discussions with S. Krishnaswami have helped significantly to improve this manuscript. We thank C. France-Lanord, CRPG, Nancy, France, for his support and encouragement. This manuscript has been improved considerably by the comments of two reviewers. References Bickle, M. J., J. M. Bunbury, H. J. Chapman, N. B. W. Harris, I. J. Fairchild, and T. Ahmad (2003), Fluxes of Sr into the headwaters of the Ganges, Geochim. Cosmochim. Acta, 67, Blum, J. D., C. A. Gazis, A. D. Jacobson, and C. P. Chamberlain (1998), Carbonate versus silicate weathering in Raikhot watershed within the High Himalayan crystalline series, Geology, 26, Dalai, T. K., S. Krishnaswami, and M. M. Sarin (2002), Major ion chemistry in the headwaters of the Yamuna river system: Chemical weathering, its temperature dependence and CO 2 consumption in the Himalaya, Geochim. Cosmochim. Acta, 66, Galy, A., and C. France-Lanord (1999), Weathering processes in the Ganges-Brahmaputra basin and the riverine alkalinity budget, Chem. Geol., 159, Galy, A., and C. France-Lanord (2001), Higher erosion rates in the Himalaya: Geochemical constraints on riverine fluxes, Geology, 29, Galy, A., C. France-Lanord, and L. A. Derry (1999), The strontium isotopic budget of Himalayan rivers in Nepal and Bangladesh, Geochim. Cosmochim. Acta, 63, Garzanti, E., G. Vezzoli, S. Andò, C. France-Lanord, S. K. Singh, and G. Foster (2004), Sand petrology and focused erosion in collision orogens: The Brahmaputra, Earth Planet. Sci. Lett., 220, Harris, N., M. J. Bickle, H. Chapman, I. Fairchild, and J. Bunbury (1998), The significance of the Himalayan rivers for silicate weathering rates Evidence from the Bhote Kosi tributary, Chem. Geol., 144, International Centre for Integrated Mountain Development (ICIMOD) (2001), Mountain flash floods, Newsl. 38, Kathmandu, Nepal. (Available at icimod/publications/newsletter/new38/n38toc.htm) Krishnaswami, S., and S. K. Singh (2005), Chemical weathering in the river basins of the Himalaya, India, Curr. Sci., 89, Krishnaswami, S., J. R. Trivedi, M. M. Sarin, R. Ramesh, and K. K. Sharma (1992), Strontium isotopes and rubidium in the Ganga-Brahmaputra river system: Weathering in the Himalaya, fluxes to the Bay of Bengal and contributions to the evolution of oceanic 87 Sr/ 86 Sr, Earth Planet. Sci. Lett., 109, Krishnaswami, S., S. K. Singh, and T. Dalai (1999), Silicate weathering in the Himalaya: Role in contributing to major ions and radiogenic Sr to the Bay of Bengal, in Ocean Science, Trends and Future Directions, edited by B. L. K. Somalyajulu, pp , Indian Natl. Sci. Acad. and Akad. Int., New Delhi. Millot, R., J. Gaillardet, B. Dupré, and C. J. Allègre (2002), The global control of silicate weathering rates and the coupling with physical erosion: New insights from rivers of the Canadian Shield, Earth Planet. Sci. Lett., 196, Sarin, M. M., S. Krishnaswami, K. Dilli, B. L. K. Somayajulu, and W. S. Moore (1989), Major ion chemistry of the Ganga- Brahmaputra river system: Weathering processes and fluxes to the Bay of Bengal, Geochim. Cosmochim. Acta, 53, of 14

14 Sarin, M. M., S. Krishnaswami, J. R. Trivedi, and K. K. Sharma (1992), Major ion chemistry of the Ganga source waters: Weathering in the high altitude Himalaya, Proc. Indian Acad. Sci. (Earth Planet. Sci.), 101, Singh, S. K. (2006), Spatial variability in erosion in the Brahmaputra basin: Causes and Impacts, Curr. Sci., 90, Singh, S. K., and C. France-Lanord (2002), Tracing the distribution of erosion in the Brahmaputra watershed from isotopic compositions of stream sediments, Earth Planet. Sci. Lett., 202, Singh, S. K., J. R. Trivedi, K. Pande, R. Ramesh, and S. Krishnaswami (1998), Chemical and Sr, O, C, isotopic compositions of carbonates from the Lesser Himalaya: Implications to the Sr isotope composition of the source waters of the Ganga, Ghaghara and the Indus Rivers, Geochim. Cosmochim. Acta, 62, Singh, S. K., L. Reisberg, and C. France-Lanord (2003), Re-Os isotope systematics of sediments of the Brahmaputra River system, Geochim. Cosmochim. Acta, 67, Singh, S. K., M. M. Sarin, and C. France-Lanord (2005), Chemical erosion in the eastern Himalaya: Major ion composition of the Brahmaputra and d 13 C of dissolved inorganic carbon, Geochim. Cosmochim. Acta, 69, Singh, S. K., A. Kumar, and C. France-Lanord (2006), Sr and 87 Sr/ 86 Sr in waters and sediments of the Brahmaputra River System: Silicate weathering, CO 2 consumption and Sr flux, Chem. Geol., 234, Tewari, A. (2004), Study on soil erosion in Pasighat Town (Arunachal Pradesh) India, Nat. Hazards, 32, Tipper, E. T., M. J. Bickle, A. Galy, A. J. West, C. Pomies, and H. J. Chapman (2006), The short term climatic sensitivity of carbonate and silicate weathering fluxes: Insight from seasonal variations in river chemistry, Geochim. Cosmochim. Acta, 70, of 14