- Email: [email protected]
Late Pleistocene–Holocene hydrologic changes in the interfluve areas of the central Ganga Plain, India
Geomorphology 54 (2003) 279 – 292 www.elsevier.com/locate/geomorph
Late Pleistocene–Holocene hydrologic changes in the interfluve areas of the central Ganga Plain, India Pradeep Srivastava a,1, Indra Bir Singh b, Shikha Sharma b, Uma Kant Shukla c, Ashok Kumar Singhvi a,* a
Planetary and Geoscience Division, Physical Research Laboratory, Navrangpura, Ahmedabad-380 009, India b Department of Geology, Lucknow University, Lucknow-226007, India c Department of Geology, Kumaun University, Nainital-263 002, India Received 11 October 2000; received in revised form 4 November 2002; accepted 5 December 2002
Abstract Abandoned channel belts, ponds and point bar deposits of palaeochannels in the interfluve regions of the central Ganga Plain suggest changes in the morphohydrologic conditions during the Latest Pleistocene – Holocene. Stratigraphy of these ponds comprises channel sand at the base overlain by shell-bearing clayey silt. The contact of the two facies marks the phase when channels converted into standing water bodies. Point bar deposits of some palaeochannels are overlain by oxidised aeolian sand, indicating that the channel abandonment possibly occurred due to the desiccation and aridity in the region. Optically stimulated luminescence (OSL) chronometry of the pond sediments suggests that the deposition of the basal channel sand started before 13 ka and continued up to f 8 ka. The ponds formed around 8 – 6 ka when the channel activity ceased. Evidence from the point bar deposits also indicates that the fluvial activity in the region ended sometime during 7 – 5 ka. This was followed by aeolian aggradation. The present study thus suggests that the hydrologic conditions in the Gangetic plains, i.e. initiation of channels and their abandonment, formation of microgeomorphologic features such as ponds and their eventual siltation, were controlled largely by climatic changes (i.e. monsoon changes) supported by tectonic activity. For the past 2 ka, increasing human and related agricultural activity has substantially accentuated the natural siltation rate of ponds. D 2003 Elsevier Science B.V. All rights reserved. Keywords: Ganga Plain; Interfluve; Ponds; Alluvial ridges; Luminescence dating; Holocene climate change
1. Introduction Abandoned channel belts, ponds and palaeo point bar deposits have been useful in palaeohydrologic * Corresponding author. Tel.: +91-79-6302129x4366; fax: +9179-6301502. E-mail address: [email protected] (A.K. Singhvi). 1 Present address: Department of Geography, University of Georgia, Athens, GA 30602, USA.
reconstructions. Palaeo river changes in response to climate changes have been investigated in detail in South East Asia, Australia, Africa and South America (Bishop and Godley, 1984; Rotnicki, 1991; Thomas, 2000). However, in the Indian context, such studies have been minimal. The Ganga Plain in northern India is a major component of the Himalayan foreland basin and is one of the largest, fluvially controlled depositional systems of the world. This region has three distinc-
0169-555X/03/$ - see front matter D 2003 Elsevier Science B.V. All rights reserved. doi:10.1016/S0169-555X(02)00361-6
280
P. Srivastava et al. / Geomorphology 54 (2003) 279–292
tive geomorphic surfaces of regional extent. It has been suggested that these surfaces formed as a consequence of climatic change and associated base level conditions during the Late Pleistocene –Holocene (Singh, 1996). The Upland Interfluve Surface is the most prominent and widespread geomorphic surface of the Ganga Plain. In the present study, geomorphic and sedimentological analysis of the sediments of abandoned channel belts, associated point bar deposits (alluvial ridges), ponds and lakes on this surface along with their luminescence dating were used to reconstruct palaeohydrologic changes through time.
quantities of sediment load. Locally referred to as Tals, ponds are small, closed water bodies that act as sediment sinks for the rest of the area. Alluvial ridges occur as sandy mounds within abandoned channel belts and often preserve the point bar deposits of the paleochannels (Pc). In the present study (spanning f 4000 km2 area), three representative ponds and two alluvial ridges from Gomati – Sai and Sai –Ganga River interfluves were examined and based on their sedimentary successions, luminescence chronology and the regional/ global climatic record, palaeohydrologic changes in the region are inferred.
2. Geomorphic setting
3. Morphology and sedimentology of ponds
The study area spans part of rivers Ghaghara – Gomati, Gomati – Sai and Sai – Ganga interfluves (Figs. 1 and 2). Satellite data (Fig. 1a) show that, (i) the river Ganga (SW corner of the imagery) is strictly braided and has f 12 km broad flood plain with meander scars on its river valley terrace (T1 terrace) and, (ii) the Sai and the Gomati Rivers (extreme NE corner) are meandering streams with narrower ( f 1 km) flood plains. Regional slope of the area is NW –SE. A cross-section along A – AV shows that River Ganga is incised by f 25 m into the upland terrace (T2 terrace), the Gomati up to 10 m and the Sai up to f 7 m (Fig. 1b). The region between the two river valleys occurs as a raised 20 – 50 km broad interfluve area on which smaller subwatersheds (up to fourth order), abandoned channels belts and alluvial ridges occur. The inference that interfluves (T2 terrrace) are the oldest geomorphic surfaces is based on the fact that all the other surfaces are either superimposed or incised into it. Earlier luminescence ages from near the top of this surface in the region suggest that its aggradation continued at least until 7 ka (Srivastava et al., 2003a). The sedimentary sequence of this surface comprises fine sand and silt with calcrete layers, suggesting its formation due to the activation and abandonment of smaller river channels, lake activity and sheet flows on flat surfaces (Singh et al., 1999). Presently, f 30% of the interfluve area is being drained by small, highly sinuous, underfit channels that are incapable of transporting any appreciable
The dimensions of the ponds are variable and their depth ranges between 1 and 5 m. They are local sinks, where sediment – water transfer into the pond occurs by sheet flow and minor rills. No delta building takes place and no outflow channels that could drain out the pond water, exist. The water budget of the ponds shows seasonality such that during the summer months some of the ponds dry out. Based on an analysis of shapes and sizes of 1388 ponds, three basic pond types were identified (Figs. 2 and 3): 1. Type 1 ponds have an area of >0.1 – 2.7 km2. These ponds are perennial and occur as chains of arcuate and linear lakes associated with linear belts of abandoned channels and their cutoff meanders. Occasionally, circular or subcircular shapes also occur. About 14% of the ponds accounting for f 50% of the total ponded area belong to this category. 2. Type 2 ponds occur as small and arcuate, linear or oval-shaped water bodies and, have an area of f 0.05 –0.1 km2. About 7% of the ponds occupying >20% of the total ponded area belong to this category. 3. Type 3 ponds occur in clusters, at times scattered throughout the area and are generally located in linear depressions. Such ponds have an area of < 0.05 km2 and f 79% of the ponds covering f 30% of total ponded area belong to this group.
P. Srivastava et al. / Geomorphology 54 (2003) 279–292
281
Fig. 1. (A) Satellite data showing the general geomorphology of the area. Note the occurrence of ponds as parts of paleochannels (Pc) in interfluve areas (T2). T1 is a river valley terrace showing a meander scroll plain (SP). T0 is an active flood plain. (B) Cross-section along A – AV. Note the channel incision and raised interfluve areas and position of ponds and alluvial ridges.
Pond density, i.e. (total number of ponds/square kilometre) varies from 0.73 to 0.1 and higher pond density is accompanied by clustered occurrences of
ponds of Type 3 Stratigraphy of the larger ponds comprises a sandy unit at the base overlain by lensoid shell layer (marl deposit), followed by f 2 – 3 m
282
P. Srivastava et al. / Geomorphology 54 (2003) 279–292
Fig. 2. Geomorphology of the study area. Water bodies viz, Chunauti Tal, Uganna Tal and Misa Tal, marked A, B and C, respectively, are studied in detail for the stratigraphy and chronology.
thick clayey silty sediment. The shell layer marks the initiation of pond activity and has been radiocarbon dated to 9 – 8 ka BP (Agarwal et al., 1992; Singh, 1996). In the present study, the ponds with clear geomorphic evidences of an association with abandoned channel belt (sinuous and linear geometry)
were selected for sedimentological study and are discussed below (Fig. 1). Chunauti Tal: It is a linear pond f 125 m wide and f 1000 m long. The basal unit comprises micaceous fine sand grading into a 1.7-m-thick greyish yellow silty fine sand unit. The sandy unit is capped
P. Srivastava et al. / Geomorphology 54 (2003) 279–292
283
Fig. 3. Pond types in the region. The dotted line represents the depressed areas connecting two smaller ponds.
by 0.75-m-thick black ferrugenised clayey silt with sporadic gastropod and bivalve shells. Uganna Tal: It is an arcuate pond inscribing an island, has a length of f 1200 m and width of f 150 m. In a 1-m deep trench, the basal 30 cm is micaceous fine sand (3– 4 B), overlain by f 25 cm bioturbated, ferrugenised silty clay. This is capped by 45-cm- thick black clayey silt. Misa Tal: It is an arcuate pond with a width of f 150 m and a length of f 1500 m with a stratigraphic sequence comprising basal micaceous fine sand overlain by f 60 cm unit of shell-bearing clayey silt. This silt is overlain by 1.8-m-thick black clayey silt with gastropod shells. The clay percentage varies up to 30%. Overall, the pond sequences comprise grey coloured micaceous sands at the base followed by lenticular units of shell and clayey sediments. The facies association of sand overlain by shell-bearing clayey silt suggests that an active channel carrying fine sand was abandoned and converted into ponds in which sedimentation of a clay-rich shell-bearing unit occurred. The sediment type, the depositional environment of pond-fill deposits and their geomorphic pattern (arcuate, linear) indicates their ori-
gin as a consequence of channel abandonment in the region. The presence of bioturbated ferruginized silty clayey units indicates sedimentation under stagnant water conditions. Table 1 Lithofacies of alluvial ridges Lithofacies
Thickness (m)
Textural attributes
Cross-bedded sand
0.5 – 1.5
Parallel laminated sand Rippled fine sand
0.5 – 2.5
grey, large-scale trough and planar cross-bedded, moderate mottling, occasional clay balls grey to yellowish grey, individual laminae 0.5 – 1.0 cm thick, moderately mottled grey to yellowish grey, present in form of lenses in different facies; ripple wavelength V 20.0 cm and height up to 5.0 cm grey-buff, grey, extensive mottling; both animal and plant burrows are visible yellow reddish yellow, highly ferruginized, no physical structures but occasional burrows are visible
0.2 – 0.3
Bioturbated fine sand
1.0 – 3.0
Well-sorted ferruginized sand
1.0 – 3.0
284
P. Srivastava et al. / Geomorphology 54 (2003) 279–292
4. Alluvial ridges Alluvial ridges are another important features associated with the abandoned channel belts. Facies, palaeohydrology and luminescence chronology of these features have been reported by Singh et al. (1998) and Srivastava et al. (2000). Morphology, sedimentology and chronology of these features are described below. The ridges exist on the upland interfluve surface and are of oval, arcuate and linear shaped with a positive relief of 2 –5 m and can be traced up to 0.25 –
2.5 km. Table 1 and Fig. 4 summarise the lithological details on the alluvial ridges. Overall, the stratigraphy shows a fining upwards point bar sequence at the base overlain by well-sorted, oxidised aeolian sand. Based on the geometry of the sand body and the grain-size changes in a similar point bar deposit (Gangaganj ridge, Fig. 4), Singh et al. (1998) inferred that a gradual decline of water discharge (from 447 to 7 m3/s) occurred with eventual abandonment of the channel. Presence of wind blown sand over point bar deposits suggests a period of enhanced aridity. A similar inference has been made in the western Ganga
Fig. 4. Lithological succession and chronology of alluvial ridges (a) at Gangaganj and (b) at Gahira bypass.
P. Srivastava et al. / Geomorphology 54 (2003) 279–292
Fig. 5. Palaeowatershed. Probable water divide demarcated on the basis of pattern of slope marked with thick arrows.
285
286
P. Srivastava et al. / Geomorphology 54 (2003) 279–292
Plain. Slope and geomorphic analysis of the Survey of India toposheets (1:50,000 scale) indicates a palaeowatershed with 55 alluvial ridges within the palaeo water-divide (Fig. 5).
5. Luminescence dating Luminescence dating of sediments relies on the fact that, during their predepositional weathering and transport, daylight exposure of minerals constituting the sediments bleaches their geological luminescence to a near zero residual level (I0). Thus, at the time of their sedimentation, the grains have zero or near zero luminescence. On burial, further daylight exposure ceases and reaccumulation of the luminescence signal starts due to radiation exposure arising from the decay of natural radioactivity viz. 232Th, 238U and 40K in the sediment. A small contribution additionally comes from the cosmic rays. Thus, reaccumulation of the luminescence signal continues until the excavation and sample collection. The total luminescence signal (Inat) comprises the initial residual level I0 and the luminescence (Id) acquired since burial (i.e. Inat = Id + I0). Id can be related to the age of burial via the annual rate of luminescence induction by the relation, Age ¼ Paleodose=Annual dose ¼ P=DT Where P is the equivalent dose or the paleodose representing the laboratory radiation dose that induces luminescence intensity Id. DT represents the total annual radiation dose and is computed using the elemental concentrations of natural radioactivity. The event dated by the luminescence technique is the most recent transport and burial event (see e.g. Aitken, 1985, 1998). Aeolian sediments are transported by saltation and/ or suspension over long distances and consequently experience direct daylight exposure for an extended duration. This enables a reasonable assumption that their geological luminescence is bleached to an optimum residual level. Contrastingly, fluvially transported sediments experience a wide variety of daylight exposure before deposition. This is due to attenuation and modification of daylight and its spectrum, by the water column, turbulence, the grain size and the sediment load. Thus, in general, the daylight
bleaching of fluvially transported sediments is heterogeneous and partial. The advent of Optically Stimulated Luminescence (OSL) dating technique with rapid bleachability of OSL signal helped in minimising some of these problems. The feasibility of using infrared stimulated luminescence dating (IRSL) in Ganga Plain was first demonstrated by Someshwarrao et al. (1997) and Srivastava et al. (2003b) have applied these methods in the region. In the present study, samples from the basal channel sand and overlying clayey units were dated to develop a chronology of the events of channel activity, their abandonment and the development of the ponds. Sample nos. CH 1 – 3 were collected from Chunauti Tal sequence, GP 20 – 21 from Uganna Tal and ME 1 – 2 from Misa Tal. Fig. 1 provides the sample locations and stratigraphic logs. All the samples were pretreated with 10% HCl and 30% H2O2 to remove carbonates and organic matter. The samples were then dry sieved to obtain the 105– 150 Am grain size fraction. Density separation using Na-polytungstate (q = 2.58 g/cm3) was carried out to separate Quartz and K-feldspar fractions. The Quartz fraction was further etched with 40% HF for 80 min followed by 12N HCl for 30 min to remove the alpha skin. The separated mineral grains were mounted as a monolayer on stainless steel discs using Silkospray. For silt-rich sediments, the analysis was carried out on 4 –11 Am polyminerallic samples extracted through a sequential pretreatment of the sample with 1N HCl, 30% H2O2 followed by deflocculation using 0.01N Na-oxalate and eventual extraction of the desired grain size using Stokes settling times appropriate for a 6-cm column of acetone. The extracted fine silt fraction was resuspended in acetone and equal volumes were pipetted onto aluminium discs kept in small glass cylinders with acetone, and finally dried at 50 jC in the dark. Blue-Green light Stimulated Luminescence (BGSL) and Infrared light stimulated luminescence (IRSL) measurements were carried out on a Riso TL-DA-15 reader. IR stimulation used TEMT-484 infrared diode arrays and a filtered halogen lamp was used for BGSL. For IRSL, the detection optics comprised Corning 7-59 and Schott BG-39 filters coupled to an EMI 9635 QA photomultiplier tube, whereas for BGSL, the detection optics channel comprised two Hoya U-340 and Schott BG-39 filters. Beta irradiation was made using a 25-
P. Srivastava et al. / Geomorphology 54 (2003) 279–292
mCi 90Sr/90Y beta source. The paleodose (P) was estimated using the additive dose method and a late light subtraction was carried out. The dose rate estimation relied on thick source ZnS (Ag) alpha counting technique for elemental concentration of Uranium, Thorium and NaI (TI) gamma ray spectrometry for Potassium. Radioactive equilibrium was assumed in the decay series. Though this assumption may not always hold true for fluvial sediments, nearly 50% contribution to the dose by potassium provides some cushion against gross errors in the ages and corresponding interpretations based on this assumption. Quartz BGSL ages on sample no. CH1, CH2, GP20 and ME1 were determined after elevated temperature IR cleaning at 220 jC for 5 min (Jain and Singhvi, 2001) with the assumption that IR cleaning does not affect the Quartz BGSL. Ages on samples CH3, GP21 and ME2 were estimated using IRSL analysis of fine-grain polyminerallic aliquots. The choice of a fine-grain or coarse-grain analysis was dictated by the mean grain size of the sediment.
6. Results and discussion Table 2 provides data on radioactivity, paleodose, dose rate and the ages. In the following, the importance of these ages is discussed. Fig. 6 shows the typical IRSL shine down curve, paleodose plateau and growth curve for the sample GP-21. 6.1. Evolution chronology of ponds Chunauti Tal: Two dates on the sandy, fining upward sequence do not indicate depositional hiatus. Basal age of the sandy unit was 11 F 2 ka and the top
287
of this unit was dated to 8 F 3 ka. The base of the overlying muddy unit was dated to 7 F 1 ka. These ages suggest that the channel activity was established prior to 11 ka and continued until 8 ka and the channel abandonment occurred some time during f 8 –7 ka. Uganna Tal: The channel sand gave an age of 13 F 3 ka, showing that this channel was initiated before 13 ka. The age of the initiation of the muddy unit was 7 F 2 ka, indicating that the abandonment of the channel and the formation of this pond occurred some time around 7 ka. Misa Tal: The sandy unit of this pond was saturated with water, which precluded sampling. The age from the muddy unit from just above the sand was 6 F 1 ka, indicating that here, the channel abandonment had occurred prior to 6 ka. The chronostratigraphic data suggest that the channel activity in the region began before 13 ka and continued at least up to 8 ka. These channels were abandoned some time during 8 – 6 ka. Point bar deposits of paleochannels, preserved as alluvial ridges, are overlain by brownish-coloured, well-sorted aeolian sand in the region, which also suggests a phase of channel abandonment during 7– 5 ka (Fig. 4). This means that during z 13 ka, there was a phase when many smaller channels, capable of carrying very fine sand (3– 4B) developed on the upland interfluve terrace. Such small channels remained active up to 8 ka and thereafter; some of the channels became abandoned, giving rise to the ponds. Some of the larger channels, now seen as abandoned channel belts, were active between >13 and 8 ka and became abandoned around 8 ka forming large linear lakes. While channels were once operating simultaneously, their size and order should have determined the exact timing of their abandonment. Occurrence of aeolian
Table 2 Radioactivity, paleodose, dose rate and ages obtained from sequences of ponds Sample no.
Depth (m)
Mineral
Technique
U (ppm)
Th (ppm)
K (%)
Dose rate (Gy/ka)
P (Gy)
Age (ka)
CH-1 CH-2 CH-3 GP-20 GP-21 ME-1 ME-2
2.70 0.75 0.35 1.0 0.5 2.1 1.5
quartz quartz polyminerallic quartz polyminerallic quartz polyminerallic
BGSL BGSL IRSL BGSL IRSL BGSL IRSL
5.4 F 0.3 3.9 F 0.8 4.3 F 0.9 2.5 F 0.9 13.7 F 1 8.5 F 0.4 8.9 F 1.4
19.6 F 6.7 21.2 F 6.5 21.4 F 7.7 6.5 F 2.9 13.8 F 7.8 13.0 F 6.9 19.1 F 8.6
2.1 F 0.1 2.1 F 0.1 2.1 F 0.1 1.8 F 0.1 1.8 F 0.1 1.3 F 0.1 3.3 F 0.2
4.2 F 0.5 4.3 F 0.6 5.9 F 1.1 2.1 F 0.4 6.1 F 1.2 5.0 F 0.7 6.2 F 0.8
45.6 F 5.6 35.1 F 12.4 40.3 F 2.5 26.8 F 2.8 41.9 F 6.4 30.1 F 1.9 14.3 F 2.2
11.0 F 2.0 8.0 F 3.0 7.0 F 1.0 13.0 F 3.0 7.0 F 2.0 6.0 F 1.0 2.0 F 0.5
288
P. Srivastava et al. / Geomorphology 54 (2003) 279–292
Fig. 6. Typical IRSL (A) IRSL output with stimulation time. N refers to signal from sample as received and the other curves. N + nGy refer to samples additionally irradiated in the laboratory with calibrated beta source, (B) growth of luminescence with dose and (C) paleodose vs. stimulation time.
sand in the region also suggests a phase of decreased water budget and a more arid climate leading to channel desiccation. Table 3 provides a compilation of chronologies of channel abandonment in the area. The luminescence dates are comparable with the calibrated 14C-dates on shells given by Singh (1996) and Agarwal et al. (1992), which probably shows that the net effect of diagenetic alteration and reservoir age is minimal and the radiocarbon ages, at least in such cases are acceptable (Timothy, 1986). The distribution pattern of the ponds and their sedimentary sequence shows that they evolved due to the abandonment. Analogous evidence of pond formation in the central Ganga Plain has been provided by Agarwal et al. (1992) and Singh (1996). Although the size and the order of the
abandoned channels determined the initial shape and size of ponds, continuing siltation and erosion of individual catchments have since modified their dimensions. Thus, clustered ponds in a linear depression are most likely to be due to subsequent siltation leading to the fragmentation of the larger ponds and the formation of smaller water bodies. Similarly, varying pond densities in different regions may reflect variation in palaeo drainage density, but this aspect needs a confirmation. 6.2. Geomorphic evolution In the following, an attempt is made to understand the formation of ponds, alluvial ridges and the drainage system of interfluve areas of the central Ganga
P. Srivastava et al. / Geomorphology 54 (2003) 279–292 Table 3 Chronological studies inferring phases of channel abandonment in central Ganga Plain Geomorphic event
Present work
Agarwal et al. (1992)a
Singh (1996)a
Srivastava et al. (2000)b
Channel abandonment (ka) Channel activity (ka)
8–6
f8
f8
7–5
>8
>7
>13 – 8
–
a Based on calibrated 14C-dates on molluscan shells from muddy sediments overlying immediately above the sandy unit in sequence of Misa tal. b Based on luminescence dates on alluvial ridges in the central Ganga Plain.
Plain and their response to Late Pleistocene – Holocene climatic change and other processes. Subwatersheds in the region, originating within the alluvial plain, are usually of fourth order, 1– 3 m deep and < 30 –50 m wide and are supported by precipitation and ground water. Water in the Ganga basin is provided by precipitation and glacial melt. Discharge of Ganga River at Hardwar has f 15% contribution from glacial melt and the rest is contributed by precipitation (Das Gupta, 1975). The winter rainfall in the region accounts for f 10% of the total rainfall. This implies that the groundwater table in the Ganga Plain depends largely upon the summer monsoon rainfall. A drop in the water table consequently leads to drying of dependent shallow channels. Thus, during 8 –6 ka, when channel abandonment occurred and formation of ponds began, a lowering of groundwater table due to decreased rainfall and glacial melting should have occurred. In view of glacial component being a smaller fraction and that the ages refer to events occurring 10 – 12 ka after the Last Glacial Maximum (LGM), it is considered that lowering of groundwater reflects a phase of reduced precipitation/ monsoon regime. Based on the studies on groundwater calcrete in the central Ganga Plain, Agarwal et al. (1992) suggested a significant lowering of the water table during the Mid-Holocene. There is evidence of tens of kilometre-scale undulations in the study area, producing higher parts with 2 –10 m relief. These undulations were also responsible for formation of centripetal drainage basins with ponds of various sizes (Srivastava, 1998). The field evidence for kilometre-scale warping and undulations includes
289
centripetal drainage basins, warped areas undergoing gully erosion, sudden disappearance of abandoned channel segments, positioning of fluvial channels with Bhur sand ridges on high ground and several-metres high levees adjacent to arcuate ponds (earlier channels). It has been argued that tectonic activity producing regional warping took place in pulses between 8 and 5 ka and played an important subordinate role to climatic change in the formation of lakes and ponds (Singh, 2001, 2002). Widespread tectonic activity in Ganga Plain during the Mid-Holocene has been documented (Singh et al., 1997). The tectonic warping also affected the ground water table and helped formation of calcrete bands in warped areas around 5 ka. This also led to origin of alkaline soils in the upland interfluve areas (Singh, 2002). Singh (1996) argued that the channel abandonment during 8 –6 ka could be linked to increasing base level of the main rivers in response to mid-Holocene high sea level. The dense network of channels in the central Ganga Plain formed during lowered base level of the Late Pleistocene (Singh, 2002). The tributaries responsible for the formation of ponds do not show features supporting avulsion. All the major rivers are incised by 10 –20 m in the upland interfluve surface and possess wide river valleys with terraces, in which they flow as underfit rivers. These rivers exhibit evidence of river morphology changes (river metamorphosis) and shifting within their respective river valleys. Thus, the smaller river channels (that eventually transformed to ponds) flowing on the upland interfluve surface cannot be formed by avulsion of the major incised rivers. Based on the foregoing, it is proposed that during the time span of >13 – 8 ka, a dense network of channels developed on the interfluve surface of the central Ganga Plain. Increased monsoon rainfall in 12– 8 ka caused effective sediment – water movement through this channel network. Tectonic events of warping during 8 – 5 ka and increasing aridity caused disruption of many drainages, especially the smaller tributaries, forming centripetal drainage areas with linear ponds and lakes. The larger rivers and some tributaries were able to cut through the tectonic uplifts and survived. The weakening of SW monsoon during middle to late Holocene (Kutzbach, 1981; Van Campo, 1986; Sirocko et al., 1991; Overpeck et al., 1996) caused desiccation, segmentation of large linear
290
P. Srivastava et al. / Geomorphology 54 (2003) 279–292
lakes into a series of ponds. The arid climate and tectonically enhanced relief caused increased sediment supply into the ponds around 5 ka. Lowering of the water table and increased aridity initiated fragmentation and drying of the ponds. For the mid-Holocene, evidence from western Indian lakes indicates lake desiccation f 4.8 ka BP (Swain et al., 1983; Singh et al., 1990). Increased dune building activity f 5 ka in the Thar Desert also indicates a more arid climate in the region (Thomas et al., 1999). The present inference of a reduced hydrologic budget in the mid-Holocene thus accords with the regional climatic records. The exact timing of the events in the different regions is however likely to be different on account of differences in geomorphic thresholds for lake hydrological changes, dune building activity, vis-a-vis alluvial processes in the Ganga valley. Thus, the same overall climatic forcing via the reduced summer rainfall manifested itself differently at different spatial locations. Two dates of 6 ka at 2.2 m and 2 ka at 1.5 m in the Misa Tal gives additional information. Sedimentation of the basal 70 cm (2.2 –1.5 m) occurred in 4 ka ( f 17.5 cm/ka) whereas 1.5 m of the sedimentation occurred in only f 2 ka (i.e. 75 cm/ka). Such a fourfold enhanced siltation rate during the past 2 ka should be due to enhanced erosion linked to anthropogenic activity. Archaeological evidence in the form of habitational mounds adjacent to Misa Tal indicates human settlement at Misa Tal at least for 1700 years (unpublished data). In the study region of the Ganga– Gomati interfluve, few archaeological sites are dated around 3.5 ka. However, large-scale inhabitation of this region was initiated between 3 and 2.5 ka and continued later without interruption. Most of these habitation sites are located on the mounds adjacent to the ponds (Singh, 2002). The agricultural activity of these settlers must have resulted in increased rates of soil erosion and silting up of the lakes and ponds. At Sarai Nahar Rai, an archaeological site about 100 km east of the area, palynological studies of lake sediments revealed cereal pollen grains in the middle to late Holocene ( f 4.5 ka BP) (Gupta, 1976). The archaeological studies indicate that during the late Holocene human activity in this region increased to the extent that such processes overtook the landscape changes caused by climate alone.
In summary, the present studies show that in the central Ganga Plain, formation and evolution of microgeomorphological features (ponds) was controlled by changes in the hydrological regime, which in turn was climate/monsoon controlled. Tectonic activity and base level changes played subsidiary, supportive roles. In some cases, e.g. at Misa Tal, human activity became the key factor in the landscape evolution (siltation), overtaking the climate-induced changes of the last about 2 ka. The sequence of events in the evolution of ponds was as follows: 1. Around 13 ka with climate shifting towards a warm and humid phase, the surface water budget should have been higher due to glacial melting of Himalayan glaciers and reestablishment of SW monsoon during k13 –11 ka. This resulted in the formation of numerous subwatersheds with high stream frequency and density on the interfluve surface. Such conditions prevailed till 8 ka. 2. During 8 –6 ka, large lakes were formed due to disruption of drainage system due to tectonic warping and higher rainfall. 3. During 6 – 5 ka, increased aridity set in that enhanced channel abandonment. The water table lowered and as a consequence a large number of small subwatersheds and small channels (that primarily survived on monsoon rains for their water budget), decayed, forming smaller ponds. The Monsoon, though weaker, was able to support these water bodies, at least seasonally. With increasing aridity, deposition of aeolian sand over the exposed point bars of abandoned channels occurred leading to the formation of the alluvial ridges in the region. 4. Around 5 ka, aridity was accentuated and the ponds started shrinking and becoming abandoned due to increased rates of siltation. Enhancement of relief due to tectonic activity accelerated the erosion in the watersheds of the ponds. Siltation of ponds was further accelerated by human impacts that became significant by f 2 ka.
7. Conclusions The present study demonstrates that in certain situations, even microgeomorphological features respond
P. Srivastava et al. / Geomorphology 54 (2003) 279–292
to global changes and that their sedimentary successions can help palaeoclimatic reconstruction. The ponds in the central Ganga Plain developed due to disruption of drainage systems, in response to climatic (monsoon) changes, supported by tectonic activity and possible base level changes. Lowering of the water table and increased aridity in the Gangetic Plains caused further fragmentation and desiccation of the ponds. These events occurred during 8 –5 ka and have a clear monsoon connection. The enhanced siltation of ponds during the last 2 ka indicates human impact and increased erosion in the central Ganga Plain. Acknowledgements PS and SS acknowledge Council of Scientific and Industrial Research, New Delhi, for research fellowship. Dr. A.K. Tangri, Uttar Pradesh Remote Sensing Application Centre, Lucknow is thanked for help with the remote sensing data. Constructive comments by Prof. Vishwas Kale, Pune, Prof. Avijit Gupta and an anonymous reviewer helped in improving the presentation. UKS thanks the Department of Geology, Kumaun University for working facilities. References Agarwal, A.K., Rizvi, M.H., Singh, I.B., Kumar, A., Chandra, S., 1992. Carbonate deposits in Ganga Plain. In: Singh, I.B. (Ed.), Gangetic Plain: Terra Incognita. Geology Department, Lucknow University, Lucknow, pp. 35 – 43. Aitken, M.J., 1985. Thermoluminescence Dating. Academic Press, London. 267 pp. Aitken, M.J., 1998. An Introduction to Optical Dating. Academic Press, London. 359 pp. Bishop, P., Godley, D., 1984. Holocene palaeochannels at Sisatchanalai, north – central Thailand: ages, significance and palaeoenvironmental indications. Holocene 4, 21 – 41. Das Gupta, S.P., 1975. The Upper Gangetic Flood Plain: A Regional Survey. National Atlas Organization Monograph, Calcutta. 194 pp. Gupta, H.P., 1976. Holocene palynology from meander lake in the Ganga valley, district Pratapgarh U.P. Palaeobotanist 25, 109 – 119. Jain, M., Singhvi, A.K., 2001. Limits to depletion of green light stimulated luminescence in feldspars: implications for quartz dating. Radiation Measurements 33, 883 – 892. Kutzbach, J.E., 1981. Monsoon climate of Early Holocene: climate experiment with the earth’s orbital parameters for 9000 years ago. Science 214, 59 – 61.
291
Overpeck, J., Anderson, D., Trumbore, S., Prell, W., 1996. The southwest Indian monsoon over the last 18,000 years. Climate Dynamics 12, 213 – 225. Rotnicki, K., 1991. Retro-diction of discharges—discharges of meandering and sinuous alluvial rivers and its implications. In: Starkel, L., Gregory, K.J., Thornes, J.B. (Eds.), Temperate Palaeohydrology: Fluvial Processes in the Temperate Zone During the last 15,000 years. Wiley, Chichester, pp. 430 – 471. Singh, I.B., 1996. Geological evolution of Ganga Plain—an overview. Journal of the Palaeontological Society of India 41, 99 – 137. Singh, I.B., 2001. Proxy records of neotectonics, climate changes and anthropogenic activity in the late quaternary Ganga Plain. National symposium on the roles of earth sciences in integrated development and related societal issues. Special Publication Geological Survey of India 65 (1), xxxiii – 1. Singh, I.B., 2002. Late quaternary evolution of the Ganga Plain and proxy records of climatic change, neotectonics and anthropogenic activity. Pragdhara. Journal of the U. P. State Archeological Department 12, 1 – 25. Singh, G., Wasson, R.J., Agarwal, D.P., 1990. Vegetational and seasonal climatic changes since the last full glacial in the Thar Desert, northwestern India. Review of Palaeobotany and Palynology 64, 351 – 358. Singh, I.B., Rajagopaln, G., Agarwal, K.K., Srivastava, P., Sharma, M., Sharma, S., 1997. Evidence of middle to late Holocene neotectonic activity in the Ganga Plain. Current Science 73, 1114 – 1117. Singh, I.B., Shukla, U.K., Srivastava, P., 1998. Point bar complex of an exhumed channel in upland interfluve of Ganga Plain. Journal of the Geological Society of India 51, 315 – 322. Singh, I.B., Srivastava, P., Sharma, S., Sharma, M., Singh, D.S., Rajagopalan, G., Shukla, U.K., 1999. Upland interfluve (Doab) deposition: alternative model to muddy overbank deposits. Facies 40, 197 – 210. Sirocko, F., Sarnthein, M., Lange, H., Erlenkeuser, H., 1991. Atmospheric summer circulation and coastal upwelling in the Arabian Sea during the Holocene and last glaciation. Quaternary Research 36, 72 – 93. Someshwarrao, M., Bisaria, B.K., Singhvi, A.K., 1997. A feasibility study towards absolute dating of Indo – Gangetic alluvium using thermoluminescence and infrared-stimulated luminescence techniques. Current Science 72, 663 – 669. Srivastava, P., 1998. Sedimentology and geomorphology of interfluve areas of central Ganga Plain. PhD thesis. Geology Department. Lucknow University, Lucknow, India. 141 pp. Srivastava, P., Shukla, U.K., Mishra, P., Sharma, M., Sharma, S., Singh, I.B., Singhvi, A.K., 2000. Luminescence chronology and facies development of Bhur sands in the interfluve region of central Ganga Plain, India. Current Science 78, 498 – 503. Srivastava, P., Sharma, M., Singhvi, A.K., 2003a. Luminescence chronology of incision and channel pattern changes in River Ganga at Kanpur, India. Geomorphology (this issue). Srivastava, P., Singh, I.B., Singhvi, A.K., 2003b. Luminescence chronometry and late quaternary geomorphic history of the Ganga Plains. Palaeocol. Palaeogeograph. Palaeoclimatol. (in press).
292
P. Srivastava et al. / Geomorphology 54 (2003) 279–292
Swain, A.M., Kutzbach, J.E., Hastenrath, S., 1983. Estimate of Holocene precipitation for Rajasthan, India, based on pollen and lake level data. Quaternary Research 19, 1 – 17. Thomas, M.F., 2000. Late Quaternary environmental changes and the alluvial record in humid tropical environments. Quaternary International 72, 23 – 36. Thomas, J.V., Kar, A., Kailath, A.J., Juyal, N., Rajaguru, S.N., Singhvi, A.K., 1999. Late Pleistocene – Holocene history of aeo-
lian accumulation in the Thar Desert, India. Zeitschrift fu¨r Geomorphologie N.F. 116, 181 – 194. Timothy, Y., 1986. Studies of non-marine molluscs for the selection of shell samples for radiocarbon dating. Radiocarbon 28, 457 – 463. Van Campo, E., 1986. Monsoon fluctuations in two 20,000-Yr B.P. Oxygen-isotope/pollen records off southwest India. Quaternary Research 26, 376 – 388.
Late Pleistocene–Holocene hydrologic changes in the interfluve areas of the central Ganga Plain, India Pradeep Srivastava a,1, Indra Bir Singh b, Shikha Sharma b, Uma Kant Shukla c, Ashok Kumar Singhvi a,* a
Planetary and Geoscience Division, Physical Research Laboratory, Navrangpura, Ahmedabad-380 009, India b Department of Geology, Lucknow University, Lucknow-226007, India c Department of Geology, Kumaun University, Nainital-263 002, India Received 11 October 2000; received in revised form 4 November 2002; accepted 5 December 2002
Abstract Abandoned channel belts, ponds and point bar deposits of palaeochannels in the interfluve regions of the central Ganga Plain suggest changes in the morphohydrologic conditions during the Latest Pleistocene – Holocene. Stratigraphy of these ponds comprises channel sand at the base overlain by shell-bearing clayey silt. The contact of the two facies marks the phase when channels converted into standing water bodies. Point bar deposits of some palaeochannels are overlain by oxidised aeolian sand, indicating that the channel abandonment possibly occurred due to the desiccation and aridity in the region. Optically stimulated luminescence (OSL) chronometry of the pond sediments suggests that the deposition of the basal channel sand started before 13 ka and continued up to f 8 ka. The ponds formed around 8 – 6 ka when the channel activity ceased. Evidence from the point bar deposits also indicates that the fluvial activity in the region ended sometime during 7 – 5 ka. This was followed by aeolian aggradation. The present study thus suggests that the hydrologic conditions in the Gangetic plains, i.e. initiation of channels and their abandonment, formation of microgeomorphologic features such as ponds and their eventual siltation, were controlled largely by climatic changes (i.e. monsoon changes) supported by tectonic activity. For the past 2 ka, increasing human and related agricultural activity has substantially accentuated the natural siltation rate of ponds. D 2003 Elsevier Science B.V. All rights reserved. Keywords: Ganga Plain; Interfluve; Ponds; Alluvial ridges; Luminescence dating; Holocene climate change
1. Introduction Abandoned channel belts, ponds and palaeo point bar deposits have been useful in palaeohydrologic * Corresponding author. Tel.: +91-79-6302129x4366; fax: +9179-6301502. E-mail address: [email protected] (A.K. Singhvi). 1 Present address: Department of Geography, University of Georgia, Athens, GA 30602, USA.
reconstructions. Palaeo river changes in response to climate changes have been investigated in detail in South East Asia, Australia, Africa and South America (Bishop and Godley, 1984; Rotnicki, 1991; Thomas, 2000). However, in the Indian context, such studies have been minimal. The Ganga Plain in northern India is a major component of the Himalayan foreland basin and is one of the largest, fluvially controlled depositional systems of the world. This region has three distinc-
0169-555X/03/$ - see front matter D 2003 Elsevier Science B.V. All rights reserved. doi:10.1016/S0169-555X(02)00361-6
280
P. Srivastava et al. / Geomorphology 54 (2003) 279–292
tive geomorphic surfaces of regional extent. It has been suggested that these surfaces formed as a consequence of climatic change and associated base level conditions during the Late Pleistocene –Holocene (Singh, 1996). The Upland Interfluve Surface is the most prominent and widespread geomorphic surface of the Ganga Plain. In the present study, geomorphic and sedimentological analysis of the sediments of abandoned channel belts, associated point bar deposits (alluvial ridges), ponds and lakes on this surface along with their luminescence dating were used to reconstruct palaeohydrologic changes through time.
quantities of sediment load. Locally referred to as Tals, ponds are small, closed water bodies that act as sediment sinks for the rest of the area. Alluvial ridges occur as sandy mounds within abandoned channel belts and often preserve the point bar deposits of the paleochannels (Pc). In the present study (spanning f 4000 km2 area), three representative ponds and two alluvial ridges from Gomati – Sai and Sai –Ganga River interfluves were examined and based on their sedimentary successions, luminescence chronology and the regional/ global climatic record, palaeohydrologic changes in the region are inferred.
2. Geomorphic setting
3. Morphology and sedimentology of ponds
The study area spans part of rivers Ghaghara – Gomati, Gomati – Sai and Sai – Ganga interfluves (Figs. 1 and 2). Satellite data (Fig. 1a) show that, (i) the river Ganga (SW corner of the imagery) is strictly braided and has f 12 km broad flood plain with meander scars on its river valley terrace (T1 terrace) and, (ii) the Sai and the Gomati Rivers (extreme NE corner) are meandering streams with narrower ( f 1 km) flood plains. Regional slope of the area is NW –SE. A cross-section along A – AV shows that River Ganga is incised by f 25 m into the upland terrace (T2 terrace), the Gomati up to 10 m and the Sai up to f 7 m (Fig. 1b). The region between the two river valleys occurs as a raised 20 – 50 km broad interfluve area on which smaller subwatersheds (up to fourth order), abandoned channels belts and alluvial ridges occur. The inference that interfluves (T2 terrrace) are the oldest geomorphic surfaces is based on the fact that all the other surfaces are either superimposed or incised into it. Earlier luminescence ages from near the top of this surface in the region suggest that its aggradation continued at least until 7 ka (Srivastava et al., 2003a). The sedimentary sequence of this surface comprises fine sand and silt with calcrete layers, suggesting its formation due to the activation and abandonment of smaller river channels, lake activity and sheet flows on flat surfaces (Singh et al., 1999). Presently, f 30% of the interfluve area is being drained by small, highly sinuous, underfit channels that are incapable of transporting any appreciable
The dimensions of the ponds are variable and their depth ranges between 1 and 5 m. They are local sinks, where sediment – water transfer into the pond occurs by sheet flow and minor rills. No delta building takes place and no outflow channels that could drain out the pond water, exist. The water budget of the ponds shows seasonality such that during the summer months some of the ponds dry out. Based on an analysis of shapes and sizes of 1388 ponds, three basic pond types were identified (Figs. 2 and 3): 1. Type 1 ponds have an area of >0.1 – 2.7 km2. These ponds are perennial and occur as chains of arcuate and linear lakes associated with linear belts of abandoned channels and their cutoff meanders. Occasionally, circular or subcircular shapes also occur. About 14% of the ponds accounting for f 50% of the total ponded area belong to this category. 2. Type 2 ponds occur as small and arcuate, linear or oval-shaped water bodies and, have an area of f 0.05 –0.1 km2. About 7% of the ponds occupying >20% of the total ponded area belong to this category. 3. Type 3 ponds occur in clusters, at times scattered throughout the area and are generally located in linear depressions. Such ponds have an area of < 0.05 km2 and f 79% of the ponds covering f 30% of total ponded area belong to this group.
P. Srivastava et al. / Geomorphology 54 (2003) 279–292
281
Fig. 1. (A) Satellite data showing the general geomorphology of the area. Note the occurrence of ponds as parts of paleochannels (Pc) in interfluve areas (T2). T1 is a river valley terrace showing a meander scroll plain (SP). T0 is an active flood plain. (B) Cross-section along A – AV. Note the channel incision and raised interfluve areas and position of ponds and alluvial ridges.
Pond density, i.e. (total number of ponds/square kilometre) varies from 0.73 to 0.1 and higher pond density is accompanied by clustered occurrences of
ponds of Type 3 Stratigraphy of the larger ponds comprises a sandy unit at the base overlain by lensoid shell layer (marl deposit), followed by f 2 – 3 m
282
P. Srivastava et al. / Geomorphology 54 (2003) 279–292
Fig. 2. Geomorphology of the study area. Water bodies viz, Chunauti Tal, Uganna Tal and Misa Tal, marked A, B and C, respectively, are studied in detail for the stratigraphy and chronology.
thick clayey silty sediment. The shell layer marks the initiation of pond activity and has been radiocarbon dated to 9 – 8 ka BP (Agarwal et al., 1992; Singh, 1996). In the present study, the ponds with clear geomorphic evidences of an association with abandoned channel belt (sinuous and linear geometry)
were selected for sedimentological study and are discussed below (Fig. 1). Chunauti Tal: It is a linear pond f 125 m wide and f 1000 m long. The basal unit comprises micaceous fine sand grading into a 1.7-m-thick greyish yellow silty fine sand unit. The sandy unit is capped
P. Srivastava et al. / Geomorphology 54 (2003) 279–292
283
Fig. 3. Pond types in the region. The dotted line represents the depressed areas connecting two smaller ponds.
by 0.75-m-thick black ferrugenised clayey silt with sporadic gastropod and bivalve shells. Uganna Tal: It is an arcuate pond inscribing an island, has a length of f 1200 m and width of f 150 m. In a 1-m deep trench, the basal 30 cm is micaceous fine sand (3– 4 B), overlain by f 25 cm bioturbated, ferrugenised silty clay. This is capped by 45-cm- thick black clayey silt. Misa Tal: It is an arcuate pond with a width of f 150 m and a length of f 1500 m with a stratigraphic sequence comprising basal micaceous fine sand overlain by f 60 cm unit of shell-bearing clayey silt. This silt is overlain by 1.8-m-thick black clayey silt with gastropod shells. The clay percentage varies up to 30%. Overall, the pond sequences comprise grey coloured micaceous sands at the base followed by lenticular units of shell and clayey sediments. The facies association of sand overlain by shell-bearing clayey silt suggests that an active channel carrying fine sand was abandoned and converted into ponds in which sedimentation of a clay-rich shell-bearing unit occurred. The sediment type, the depositional environment of pond-fill deposits and their geomorphic pattern (arcuate, linear) indicates their ori-
gin as a consequence of channel abandonment in the region. The presence of bioturbated ferruginized silty clayey units indicates sedimentation under stagnant water conditions. Table 1 Lithofacies of alluvial ridges Lithofacies
Thickness (m)
Textural attributes
Cross-bedded sand
0.5 – 1.5
Parallel laminated sand Rippled fine sand
0.5 – 2.5
grey, large-scale trough and planar cross-bedded, moderate mottling, occasional clay balls grey to yellowish grey, individual laminae 0.5 – 1.0 cm thick, moderately mottled grey to yellowish grey, present in form of lenses in different facies; ripple wavelength V 20.0 cm and height up to 5.0 cm grey-buff, grey, extensive mottling; both animal and plant burrows are visible yellow reddish yellow, highly ferruginized, no physical structures but occasional burrows are visible
0.2 – 0.3
Bioturbated fine sand
1.0 – 3.0
Well-sorted ferruginized sand
1.0 – 3.0
284
P. Srivastava et al. / Geomorphology 54 (2003) 279–292
4. Alluvial ridges Alluvial ridges are another important features associated with the abandoned channel belts. Facies, palaeohydrology and luminescence chronology of these features have been reported by Singh et al. (1998) and Srivastava et al. (2000). Morphology, sedimentology and chronology of these features are described below. The ridges exist on the upland interfluve surface and are of oval, arcuate and linear shaped with a positive relief of 2 –5 m and can be traced up to 0.25 –
2.5 km. Table 1 and Fig. 4 summarise the lithological details on the alluvial ridges. Overall, the stratigraphy shows a fining upwards point bar sequence at the base overlain by well-sorted, oxidised aeolian sand. Based on the geometry of the sand body and the grain-size changes in a similar point bar deposit (Gangaganj ridge, Fig. 4), Singh et al. (1998) inferred that a gradual decline of water discharge (from 447 to 7 m3/s) occurred with eventual abandonment of the channel. Presence of wind blown sand over point bar deposits suggests a period of enhanced aridity. A similar inference has been made in the western Ganga
Fig. 4. Lithological succession and chronology of alluvial ridges (a) at Gangaganj and (b) at Gahira bypass.
P. Srivastava et al. / Geomorphology 54 (2003) 279–292
Fig. 5. Palaeowatershed. Probable water divide demarcated on the basis of pattern of slope marked with thick arrows.
285
286
P. Srivastava et al. / Geomorphology 54 (2003) 279–292
Plain. Slope and geomorphic analysis of the Survey of India toposheets (1:50,000 scale) indicates a palaeowatershed with 55 alluvial ridges within the palaeo water-divide (Fig. 5).
5. Luminescence dating Luminescence dating of sediments relies on the fact that, during their predepositional weathering and transport, daylight exposure of minerals constituting the sediments bleaches their geological luminescence to a near zero residual level (I0). Thus, at the time of their sedimentation, the grains have zero or near zero luminescence. On burial, further daylight exposure ceases and reaccumulation of the luminescence signal starts due to radiation exposure arising from the decay of natural radioactivity viz. 232Th, 238U and 40K in the sediment. A small contribution additionally comes from the cosmic rays. Thus, reaccumulation of the luminescence signal continues until the excavation and sample collection. The total luminescence signal (Inat) comprises the initial residual level I0 and the luminescence (Id) acquired since burial (i.e. Inat = Id + I0). Id can be related to the age of burial via the annual rate of luminescence induction by the relation, Age ¼ Paleodose=Annual dose ¼ P=DT Where P is the equivalent dose or the paleodose representing the laboratory radiation dose that induces luminescence intensity Id. DT represents the total annual radiation dose and is computed using the elemental concentrations of natural radioactivity. The event dated by the luminescence technique is the most recent transport and burial event (see e.g. Aitken, 1985, 1998). Aeolian sediments are transported by saltation and/ or suspension over long distances and consequently experience direct daylight exposure for an extended duration. This enables a reasonable assumption that their geological luminescence is bleached to an optimum residual level. Contrastingly, fluvially transported sediments experience a wide variety of daylight exposure before deposition. This is due to attenuation and modification of daylight and its spectrum, by the water column, turbulence, the grain size and the sediment load. Thus, in general, the daylight
bleaching of fluvially transported sediments is heterogeneous and partial. The advent of Optically Stimulated Luminescence (OSL) dating technique with rapid bleachability of OSL signal helped in minimising some of these problems. The feasibility of using infrared stimulated luminescence dating (IRSL) in Ganga Plain was first demonstrated by Someshwarrao et al. (1997) and Srivastava et al. (2003b) have applied these methods in the region. In the present study, samples from the basal channel sand and overlying clayey units were dated to develop a chronology of the events of channel activity, their abandonment and the development of the ponds. Sample nos. CH 1 – 3 were collected from Chunauti Tal sequence, GP 20 – 21 from Uganna Tal and ME 1 – 2 from Misa Tal. Fig. 1 provides the sample locations and stratigraphic logs. All the samples were pretreated with 10% HCl and 30% H2O2 to remove carbonates and organic matter. The samples were then dry sieved to obtain the 105– 150 Am grain size fraction. Density separation using Na-polytungstate (q = 2.58 g/cm3) was carried out to separate Quartz and K-feldspar fractions. The Quartz fraction was further etched with 40% HF for 80 min followed by 12N HCl for 30 min to remove the alpha skin. The separated mineral grains were mounted as a monolayer on stainless steel discs using Silkospray. For silt-rich sediments, the analysis was carried out on 4 –11 Am polyminerallic samples extracted through a sequential pretreatment of the sample with 1N HCl, 30% H2O2 followed by deflocculation using 0.01N Na-oxalate and eventual extraction of the desired grain size using Stokes settling times appropriate for a 6-cm column of acetone. The extracted fine silt fraction was resuspended in acetone and equal volumes were pipetted onto aluminium discs kept in small glass cylinders with acetone, and finally dried at 50 jC in the dark. Blue-Green light Stimulated Luminescence (BGSL) and Infrared light stimulated luminescence (IRSL) measurements were carried out on a Riso TL-DA-15 reader. IR stimulation used TEMT-484 infrared diode arrays and a filtered halogen lamp was used for BGSL. For IRSL, the detection optics comprised Corning 7-59 and Schott BG-39 filters coupled to an EMI 9635 QA photomultiplier tube, whereas for BGSL, the detection optics channel comprised two Hoya U-340 and Schott BG-39 filters. Beta irradiation was made using a 25-
P. Srivastava et al. / Geomorphology 54 (2003) 279–292
mCi 90Sr/90Y beta source. The paleodose (P) was estimated using the additive dose method and a late light subtraction was carried out. The dose rate estimation relied on thick source ZnS (Ag) alpha counting technique for elemental concentration of Uranium, Thorium and NaI (TI) gamma ray spectrometry for Potassium. Radioactive equilibrium was assumed in the decay series. Though this assumption may not always hold true for fluvial sediments, nearly 50% contribution to the dose by potassium provides some cushion against gross errors in the ages and corresponding interpretations based on this assumption. Quartz BGSL ages on sample no. CH1, CH2, GP20 and ME1 were determined after elevated temperature IR cleaning at 220 jC for 5 min (Jain and Singhvi, 2001) with the assumption that IR cleaning does not affect the Quartz BGSL. Ages on samples CH3, GP21 and ME2 were estimated using IRSL analysis of fine-grain polyminerallic aliquots. The choice of a fine-grain or coarse-grain analysis was dictated by the mean grain size of the sediment.
6. Results and discussion Table 2 provides data on radioactivity, paleodose, dose rate and the ages. In the following, the importance of these ages is discussed. Fig. 6 shows the typical IRSL shine down curve, paleodose plateau and growth curve for the sample GP-21. 6.1. Evolution chronology of ponds Chunauti Tal: Two dates on the sandy, fining upward sequence do not indicate depositional hiatus. Basal age of the sandy unit was 11 F 2 ka and the top
287
of this unit was dated to 8 F 3 ka. The base of the overlying muddy unit was dated to 7 F 1 ka. These ages suggest that the channel activity was established prior to 11 ka and continued until 8 ka and the channel abandonment occurred some time during f 8 –7 ka. Uganna Tal: The channel sand gave an age of 13 F 3 ka, showing that this channel was initiated before 13 ka. The age of the initiation of the muddy unit was 7 F 2 ka, indicating that the abandonment of the channel and the formation of this pond occurred some time around 7 ka. Misa Tal: The sandy unit of this pond was saturated with water, which precluded sampling. The age from the muddy unit from just above the sand was 6 F 1 ka, indicating that here, the channel abandonment had occurred prior to 6 ka. The chronostratigraphic data suggest that the channel activity in the region began before 13 ka and continued at least up to 8 ka. These channels were abandoned some time during 8 – 6 ka. Point bar deposits of paleochannels, preserved as alluvial ridges, are overlain by brownish-coloured, well-sorted aeolian sand in the region, which also suggests a phase of channel abandonment during 7– 5 ka (Fig. 4). This means that during z 13 ka, there was a phase when many smaller channels, capable of carrying very fine sand (3– 4B) developed on the upland interfluve terrace. Such small channels remained active up to 8 ka and thereafter; some of the channels became abandoned, giving rise to the ponds. Some of the larger channels, now seen as abandoned channel belts, were active between >13 and 8 ka and became abandoned around 8 ka forming large linear lakes. While channels were once operating simultaneously, their size and order should have determined the exact timing of their abandonment. Occurrence of aeolian
Table 2 Radioactivity, paleodose, dose rate and ages obtained from sequences of ponds Sample no.
Depth (m)
Mineral
Technique
U (ppm)
Th (ppm)
K (%)
Dose rate (Gy/ka)
P (Gy)
Age (ka)
CH-1 CH-2 CH-3 GP-20 GP-21 ME-1 ME-2
2.70 0.75 0.35 1.0 0.5 2.1 1.5
quartz quartz polyminerallic quartz polyminerallic quartz polyminerallic
BGSL BGSL IRSL BGSL IRSL BGSL IRSL
5.4 F 0.3 3.9 F 0.8 4.3 F 0.9 2.5 F 0.9 13.7 F 1 8.5 F 0.4 8.9 F 1.4
19.6 F 6.7 21.2 F 6.5 21.4 F 7.7 6.5 F 2.9 13.8 F 7.8 13.0 F 6.9 19.1 F 8.6
2.1 F 0.1 2.1 F 0.1 2.1 F 0.1 1.8 F 0.1 1.8 F 0.1 1.3 F 0.1 3.3 F 0.2
4.2 F 0.5 4.3 F 0.6 5.9 F 1.1 2.1 F 0.4 6.1 F 1.2 5.0 F 0.7 6.2 F 0.8
45.6 F 5.6 35.1 F 12.4 40.3 F 2.5 26.8 F 2.8 41.9 F 6.4 30.1 F 1.9 14.3 F 2.2
11.0 F 2.0 8.0 F 3.0 7.0 F 1.0 13.0 F 3.0 7.0 F 2.0 6.0 F 1.0 2.0 F 0.5
288
P. Srivastava et al. / Geomorphology 54 (2003) 279–292
Fig. 6. Typical IRSL (A) IRSL output with stimulation time. N refers to signal from sample as received and the other curves. N + nGy refer to samples additionally irradiated in the laboratory with calibrated beta source, (B) growth of luminescence with dose and (C) paleodose vs. stimulation time.
sand in the region also suggests a phase of decreased water budget and a more arid climate leading to channel desiccation. Table 3 provides a compilation of chronologies of channel abandonment in the area. The luminescence dates are comparable with the calibrated 14C-dates on shells given by Singh (1996) and Agarwal et al. (1992), which probably shows that the net effect of diagenetic alteration and reservoir age is minimal and the radiocarbon ages, at least in such cases are acceptable (Timothy, 1986). The distribution pattern of the ponds and their sedimentary sequence shows that they evolved due to the abandonment. Analogous evidence of pond formation in the central Ganga Plain has been provided by Agarwal et al. (1992) and Singh (1996). Although the size and the order of the
abandoned channels determined the initial shape and size of ponds, continuing siltation and erosion of individual catchments have since modified their dimensions. Thus, clustered ponds in a linear depression are most likely to be due to subsequent siltation leading to the fragmentation of the larger ponds and the formation of smaller water bodies. Similarly, varying pond densities in different regions may reflect variation in palaeo drainage density, but this aspect needs a confirmation. 6.2. Geomorphic evolution In the following, an attempt is made to understand the formation of ponds, alluvial ridges and the drainage system of interfluve areas of the central Ganga
P. Srivastava et al. / Geomorphology 54 (2003) 279–292 Table 3 Chronological studies inferring phases of channel abandonment in central Ganga Plain Geomorphic event
Present work
Agarwal et al. (1992)a
Singh (1996)a
Srivastava et al. (2000)b
Channel abandonment (ka) Channel activity (ka)
8–6
f8
f8
7–5
>8
>7
>13 – 8
–
a Based on calibrated 14C-dates on molluscan shells from muddy sediments overlying immediately above the sandy unit in sequence of Misa tal. b Based on luminescence dates on alluvial ridges in the central Ganga Plain.
Plain and their response to Late Pleistocene – Holocene climatic change and other processes. Subwatersheds in the region, originating within the alluvial plain, are usually of fourth order, 1– 3 m deep and < 30 –50 m wide and are supported by precipitation and ground water. Water in the Ganga basin is provided by precipitation and glacial melt. Discharge of Ganga River at Hardwar has f 15% contribution from glacial melt and the rest is contributed by precipitation (Das Gupta, 1975). The winter rainfall in the region accounts for f 10% of the total rainfall. This implies that the groundwater table in the Ganga Plain depends largely upon the summer monsoon rainfall. A drop in the water table consequently leads to drying of dependent shallow channels. Thus, during 8 –6 ka, when channel abandonment occurred and formation of ponds began, a lowering of groundwater table due to decreased rainfall and glacial melting should have occurred. In view of glacial component being a smaller fraction and that the ages refer to events occurring 10 – 12 ka after the Last Glacial Maximum (LGM), it is considered that lowering of groundwater reflects a phase of reduced precipitation/ monsoon regime. Based on the studies on groundwater calcrete in the central Ganga Plain, Agarwal et al. (1992) suggested a significant lowering of the water table during the Mid-Holocene. There is evidence of tens of kilometre-scale undulations in the study area, producing higher parts with 2 –10 m relief. These undulations were also responsible for formation of centripetal drainage basins with ponds of various sizes (Srivastava, 1998). The field evidence for kilometre-scale warping and undulations includes
289
centripetal drainage basins, warped areas undergoing gully erosion, sudden disappearance of abandoned channel segments, positioning of fluvial channels with Bhur sand ridges on high ground and several-metres high levees adjacent to arcuate ponds (earlier channels). It has been argued that tectonic activity producing regional warping took place in pulses between 8 and 5 ka and played an important subordinate role to climatic change in the formation of lakes and ponds (Singh, 2001, 2002). Widespread tectonic activity in Ganga Plain during the Mid-Holocene has been documented (Singh et al., 1997). The tectonic warping also affected the ground water table and helped formation of calcrete bands in warped areas around 5 ka. This also led to origin of alkaline soils in the upland interfluve areas (Singh, 2002). Singh (1996) argued that the channel abandonment during 8 –6 ka could be linked to increasing base level of the main rivers in response to mid-Holocene high sea level. The dense network of channels in the central Ganga Plain formed during lowered base level of the Late Pleistocene (Singh, 2002). The tributaries responsible for the formation of ponds do not show features supporting avulsion. All the major rivers are incised by 10 –20 m in the upland interfluve surface and possess wide river valleys with terraces, in which they flow as underfit rivers. These rivers exhibit evidence of river morphology changes (river metamorphosis) and shifting within their respective river valleys. Thus, the smaller river channels (that eventually transformed to ponds) flowing on the upland interfluve surface cannot be formed by avulsion of the major incised rivers. Based on the foregoing, it is proposed that during the time span of >13 – 8 ka, a dense network of channels developed on the interfluve surface of the central Ganga Plain. Increased monsoon rainfall in 12– 8 ka caused effective sediment – water movement through this channel network. Tectonic events of warping during 8 – 5 ka and increasing aridity caused disruption of many drainages, especially the smaller tributaries, forming centripetal drainage areas with linear ponds and lakes. The larger rivers and some tributaries were able to cut through the tectonic uplifts and survived. The weakening of SW monsoon during middle to late Holocene (Kutzbach, 1981; Van Campo, 1986; Sirocko et al., 1991; Overpeck et al., 1996) caused desiccation, segmentation of large linear
290
P. Srivastava et al. / Geomorphology 54 (2003) 279–292
lakes into a series of ponds. The arid climate and tectonically enhanced relief caused increased sediment supply into the ponds around 5 ka. Lowering of the water table and increased aridity initiated fragmentation and drying of the ponds. For the mid-Holocene, evidence from western Indian lakes indicates lake desiccation f 4.8 ka BP (Swain et al., 1983; Singh et al., 1990). Increased dune building activity f 5 ka in the Thar Desert also indicates a more arid climate in the region (Thomas et al., 1999). The present inference of a reduced hydrologic budget in the mid-Holocene thus accords with the regional climatic records. The exact timing of the events in the different regions is however likely to be different on account of differences in geomorphic thresholds for lake hydrological changes, dune building activity, vis-a-vis alluvial processes in the Ganga valley. Thus, the same overall climatic forcing via the reduced summer rainfall manifested itself differently at different spatial locations. Two dates of 6 ka at 2.2 m and 2 ka at 1.5 m in the Misa Tal gives additional information. Sedimentation of the basal 70 cm (2.2 –1.5 m) occurred in 4 ka ( f 17.5 cm/ka) whereas 1.5 m of the sedimentation occurred in only f 2 ka (i.e. 75 cm/ka). Such a fourfold enhanced siltation rate during the past 2 ka should be due to enhanced erosion linked to anthropogenic activity. Archaeological evidence in the form of habitational mounds adjacent to Misa Tal indicates human settlement at Misa Tal at least for 1700 years (unpublished data). In the study region of the Ganga– Gomati interfluve, few archaeological sites are dated around 3.5 ka. However, large-scale inhabitation of this region was initiated between 3 and 2.5 ka and continued later without interruption. Most of these habitation sites are located on the mounds adjacent to the ponds (Singh, 2002). The agricultural activity of these settlers must have resulted in increased rates of soil erosion and silting up of the lakes and ponds. At Sarai Nahar Rai, an archaeological site about 100 km east of the area, palynological studies of lake sediments revealed cereal pollen grains in the middle to late Holocene ( f 4.5 ka BP) (Gupta, 1976). The archaeological studies indicate that during the late Holocene human activity in this region increased to the extent that such processes overtook the landscape changes caused by climate alone.
In summary, the present studies show that in the central Ganga Plain, formation and evolution of microgeomorphological features (ponds) was controlled by changes in the hydrological regime, which in turn was climate/monsoon controlled. Tectonic activity and base level changes played subsidiary, supportive roles. In some cases, e.g. at Misa Tal, human activity became the key factor in the landscape evolution (siltation), overtaking the climate-induced changes of the last about 2 ka. The sequence of events in the evolution of ponds was as follows: 1. Around 13 ka with climate shifting towards a warm and humid phase, the surface water budget should have been higher due to glacial melting of Himalayan glaciers and reestablishment of SW monsoon during k13 –11 ka. This resulted in the formation of numerous subwatersheds with high stream frequency and density on the interfluve surface. Such conditions prevailed till 8 ka. 2. During 8 –6 ka, large lakes were formed due to disruption of drainage system due to tectonic warping and higher rainfall. 3. During 6 – 5 ka, increased aridity set in that enhanced channel abandonment. The water table lowered and as a consequence a large number of small subwatersheds and small channels (that primarily survived on monsoon rains for their water budget), decayed, forming smaller ponds. The Monsoon, though weaker, was able to support these water bodies, at least seasonally. With increasing aridity, deposition of aeolian sand over the exposed point bars of abandoned channels occurred leading to the formation of the alluvial ridges in the region. 4. Around 5 ka, aridity was accentuated and the ponds started shrinking and becoming abandoned due to increased rates of siltation. Enhancement of relief due to tectonic activity accelerated the erosion in the watersheds of the ponds. Siltation of ponds was further accelerated by human impacts that became significant by f 2 ka.
7. Conclusions The present study demonstrates that in certain situations, even microgeomorphological features respond
P. Srivastava et al. / Geomorphology 54 (2003) 279–292
to global changes and that their sedimentary successions can help palaeoclimatic reconstruction. The ponds in the central Ganga Plain developed due to disruption of drainage systems, in response to climatic (monsoon) changes, supported by tectonic activity and possible base level changes. Lowering of the water table and increased aridity in the Gangetic Plains caused further fragmentation and desiccation of the ponds. These events occurred during 8 –5 ka and have a clear monsoon connection. The enhanced siltation of ponds during the last 2 ka indicates human impact and increased erosion in the central Ganga Plain. Acknowledgements PS and SS acknowledge Council of Scientific and Industrial Research, New Delhi, for research fellowship. Dr. A.K. Tangri, Uttar Pradesh Remote Sensing Application Centre, Lucknow is thanked for help with the remote sensing data. Constructive comments by Prof. Vishwas Kale, Pune, Prof. Avijit Gupta and an anonymous reviewer helped in improving the presentation. UKS thanks the Department of Geology, Kumaun University for working facilities. References Agarwal, A.K., Rizvi, M.H., Singh, I.B., Kumar, A., Chandra, S., 1992. Carbonate deposits in Ganga Plain. In: Singh, I.B. (Ed.), Gangetic Plain: Terra Incognita. Geology Department, Lucknow University, Lucknow, pp. 35 – 43. Aitken, M.J., 1985. Thermoluminescence Dating. Academic Press, London. 267 pp. Aitken, M.J., 1998. An Introduction to Optical Dating. Academic Press, London. 359 pp. Bishop, P., Godley, D., 1984. Holocene palaeochannels at Sisatchanalai, north – central Thailand: ages, significance and palaeoenvironmental indications. Holocene 4, 21 – 41. Das Gupta, S.P., 1975. The Upper Gangetic Flood Plain: A Regional Survey. National Atlas Organization Monograph, Calcutta. 194 pp. Gupta, H.P., 1976. Holocene palynology from meander lake in the Ganga valley, district Pratapgarh U.P. Palaeobotanist 25, 109 – 119. Jain, M., Singhvi, A.K., 2001. Limits to depletion of green light stimulated luminescence in feldspars: implications for quartz dating. Radiation Measurements 33, 883 – 892. Kutzbach, J.E., 1981. Monsoon climate of Early Holocene: climate experiment with the earth’s orbital parameters for 9000 years ago. Science 214, 59 – 61.
291
Overpeck, J., Anderson, D., Trumbore, S., Prell, W., 1996. The southwest Indian monsoon over the last 18,000 years. Climate Dynamics 12, 213 – 225. Rotnicki, K., 1991. Retro-diction of discharges—discharges of meandering and sinuous alluvial rivers and its implications. In: Starkel, L., Gregory, K.J., Thornes, J.B. (Eds.), Temperate Palaeohydrology: Fluvial Processes in the Temperate Zone During the last 15,000 years. Wiley, Chichester, pp. 430 – 471. Singh, I.B., 1996. Geological evolution of Ganga Plain—an overview. Journal of the Palaeontological Society of India 41, 99 – 137. Singh, I.B., 2001. Proxy records of neotectonics, climate changes and anthropogenic activity in the late quaternary Ganga Plain. National symposium on the roles of earth sciences in integrated development and related societal issues. Special Publication Geological Survey of India 65 (1), xxxiii – 1. Singh, I.B., 2002. Late quaternary evolution of the Ganga Plain and proxy records of climatic change, neotectonics and anthropogenic activity. Pragdhara. Journal of the U. P. State Archeological Department 12, 1 – 25. Singh, G., Wasson, R.J., Agarwal, D.P., 1990. Vegetational and seasonal climatic changes since the last full glacial in the Thar Desert, northwestern India. Review of Palaeobotany and Palynology 64, 351 – 358. Singh, I.B., Rajagopaln, G., Agarwal, K.K., Srivastava, P., Sharma, M., Sharma, S., 1997. Evidence of middle to late Holocene neotectonic activity in the Ganga Plain. Current Science 73, 1114 – 1117. Singh, I.B., Shukla, U.K., Srivastava, P., 1998. Point bar complex of an exhumed channel in upland interfluve of Ganga Plain. Journal of the Geological Society of India 51, 315 – 322. Singh, I.B., Srivastava, P., Sharma, S., Sharma, M., Singh, D.S., Rajagopalan, G., Shukla, U.K., 1999. Upland interfluve (Doab) deposition: alternative model to muddy overbank deposits. Facies 40, 197 – 210. Sirocko, F., Sarnthein, M., Lange, H., Erlenkeuser, H., 1991. Atmospheric summer circulation and coastal upwelling in the Arabian Sea during the Holocene and last glaciation. Quaternary Research 36, 72 – 93. Someshwarrao, M., Bisaria, B.K., Singhvi, A.K., 1997. A feasibility study towards absolute dating of Indo – Gangetic alluvium using thermoluminescence and infrared-stimulated luminescence techniques. Current Science 72, 663 – 669. Srivastava, P., 1998. Sedimentology and geomorphology of interfluve areas of central Ganga Plain. PhD thesis. Geology Department. Lucknow University, Lucknow, India. 141 pp. Srivastava, P., Shukla, U.K., Mishra, P., Sharma, M., Sharma, S., Singh, I.B., Singhvi, A.K., 2000. Luminescence chronology and facies development of Bhur sands in the interfluve region of central Ganga Plain, India. Current Science 78, 498 – 503. Srivastava, P., Sharma, M., Singhvi, A.K., 2003a. Luminescence chronology of incision and channel pattern changes in River Ganga at Kanpur, India. Geomorphology (this issue). Srivastava, P., Singh, I.B., Singhvi, A.K., 2003b. Luminescence chronometry and late quaternary geomorphic history of the Ganga Plains. Palaeocol. Palaeogeograph. Palaeoclimatol. (in press).
292
P. Srivastava et al. / Geomorphology 54 (2003) 279–292
Swain, A.M., Kutzbach, J.E., Hastenrath, S., 1983. Estimate of Holocene precipitation for Rajasthan, India, based on pollen and lake level data. Quaternary Research 19, 1 – 17. Thomas, M.F., 2000. Late Quaternary environmental changes and the alluvial record in humid tropical environments. Quaternary International 72, 23 – 36. Thomas, J.V., Kar, A., Kailath, A.J., Juyal, N., Rajaguru, S.N., Singhvi, A.K., 1999. Late Pleistocene – Holocene history of aeo-
lian accumulation in the Thar Desert, India. Zeitschrift fu¨r Geomorphologie N.F. 116, 181 – 194. Timothy, Y., 1986. Studies of non-marine molluscs for the selection of shell samples for radiocarbon dating. Radiocarbon 28, 457 – 463. Van Campo, E., 1986. Monsoon fluctuations in two 20,000-Yr B.P. Oxygen-isotope/pollen records off southwest India. Quaternary Research 26, 376 – 388.