Palaeogeography, Palaeoclimatology, Palaeoecology 356–357 (2012) 38–53
Contents lists available at SciVerse ScienceDirect
Palaeogeography, Palaeoclimatology, Palaeoecology
journal homepage: www.elsevier.com/locate/palaeo
Aggradation, incision and interfluve flooding in the Ganga Valley over the past
100,000 years: Testing the influence of monsoonal precipitation
N.G. Roy a, R. Sinha a,⁎, M.R. Gibling b
a
b
Engineering Geosciences Group, Department of Civil Engineering, IIT Kanpur, Kanpur 208016, Uttar Pradesh, India
Department of Earth Sciences, Dalhousie University, Halifax, Nova Scotia, Canada B3H 4R2
a r t i c l e
i n f o
Article history:
Received 24 December 2010
Received in revised form 21 August 2011
Accepted 23 August 2011
Available online 1 September 2011
Keywords:
Valleys
Interfluves
Discontinuities
Monsoon
Aggradation
Incision
Equilibrium profile
Himalayan foreland
Ganga River
Palaeoclimate
a b s t r a c t
In order to investigate fluvial responses to climate forcing, stratigraphic data from seven cores in the Ganga
Valley of India and a valley-margin cliff section 1.3 km long are compared with proxy records for the intensity
of the Southwest Indian Monsoon. Luminescence dates indicate that the strata cover the past ~ 100 ka. Five
aggradational periods separated by incisional episodes are apparent in the valley fill, some of which have correlative strata along the valley margin. During Marine Isotope Stages (MIS) 4 and 5, the valley experienced
fluvial activity, with thin floodplain successions and probable discontinuities along the interfluve margin.
Modest fluvial activity characterised the mid part of MIS 3 at about 37 ka. Aggradation of channel sands preceded by incision is documented for late MIS 3 (23 to 30 ka), as well as for late MIS 2 to early MIS 1 (11–
16 ka) with fluvial dates focused around the Younger Dryas period of monsoon reduction. The most recent
aggradational period took place in the latest Holocene (b 2.5 ka). On the proximal part of the interfluve,
thin deposits that include lacustrine and aeolian beds continued to accumulate until about 26 ka, after
which the interfluve was apparently not inundated by the Ganga and underwent degradation through
gully erosion. There is little evidence of fluvial activity during the Last Glacial Maximum of MIS 2, when
the Ganga appears to have been underfit. The alluvial records fit well with proxy precipitation records, and
the main aggradational periods correspond to times of declining monsoonal strength. Although more difficult
to constrain, the timing of incisional periods corresponds broadly with periods of monsoonal intensification,
and the river appears to have shifted course in its valley mainly during these times.
© 2011 Elsevier B.V. All rights reserved.
1. Introduction
Over timescales of thousands to a few tens of thousands of years, inland river systems undergo phases of aggradation and incision induced
by tectonic, climatic and intrinsic factors. Consequently, fluvial stratigraphy typically displays complexity in architecture and internal heterogeneity (Ambrose et al., 1991), dependent on the interplay between
autocyclic and allocyclic variables. These variables include the magnitude
and rate of channel scour, channel deposition and migration rate, downstream change in grain size and varied sediment flux, channel belt avulsion, base-level change, climatic fluctuations, and tectonic subsidence
patterns and local fault movement (Miall, 1996; Blum and Törnqvist,
2000; Catuneanu, 2006; Gibling, 2006). The effects of these complex variables on sediment preservation form the basis for alluvial-architecture
models that seek to predict subsurface sandbody geometry.
Valleys are prominent landforms in many coastal and inland alluvial
settings, and their cutting and filling exert a major influence on alluvial
architecture. Valley filling and incision are commonly episodic in space
⁎ Corresponding author.
E-mail address: [email protected] (R. Sinha).
0031-0182/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.palaeo.2011.08.012
and time due to changes in external forcing factors, especially stream-power variations linked to the balance between discharge of water and
sediment (Bull, 1991; Bogaart et al., 2003; Strong and Paola, 2008;
Martin et al., 2011). Such cut-and-fill architecture can be studied to understand the rhythmic control of external forcing (Bull, 1991; Bestland,
1997; Weissmann et al., 2002; Gibling et al., 2005; Mack et al., 2006).
Architectural models for valley fills need to be tested for well-dated
Quaternary successions where proxy records for forcing factors are
available. However, only a few modern plains have been studied in sufficient detail to test the models effectively (Blum and Aslan, 2006;
Stouthamer and Berendsen, 2007; Amorosi et al., 2008).
The Ganga Plains in the Himalayan Foreland Basin of India are
underlain by sediment laid down by the Ganga River, among the
world's largest rivers. The plains have a well preserved alluvial record accessible through drilling, and strata are locally exposed in
cliff sections along incised reaches. These records form important
continental archives for understanding landscape dynamics against
the backdrop of late Quaternary environmental changes driven
mainly by hinterland (Himalayan) tectonics and palaeo-monsoonal
dynamics (Gibling et al., 2011). The present paper extends previous
research through a detailed study of the Ganga Valley near Kannauj in
Uttar Pradesh (Fig. 1). In this area, a long stretch of previously
N.G. Roy et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 356–357 (2012) 38–53
undescribed cliffs, ~10–20 m high, allows a high-resolution analysis of
floodplain facies and their architecture at a relatively high elevation
along the valley margin. Seven drill cores with near-complete recovery
down to a depth of 32 m were organised in two traverses across the valley to a distance of 14 km from the cliffs (Fig. 1). Cliff and core strata
were placed in a chronological framework that spans the past ~100 ka
based on 25 samples dated using luminescence methods. Comparison
with available information on palaeo-monsoon intensity suggests that,
to a first order, major monsoonal fluctuations governed fluvial events
of aggradation, incision, and inundation of the interfluve adjoining the
Ganga Valley.
2. Geographic and geomorphic setting of the Ganga Plains
The study area falls in the middle Ganga Plains and is located
~300 km downstream of the Himalayan tectonic front and more than
1200 km upstream of the modern day tidal limit. Of the total catchment
area of 2435 km 2, an area of about 206 km2 (8.5%) is presently glaciated.
The glaciated area was only slightly larger during the LGM (13%) (Owen
et al., 2002), and glacial melt water may have played a modest role in
the Ganga River's water and sediment flux through the Late Quaternary.
The study area experiences monsoonal rainfall of about 60–80 cm per
year. The hinterland of the Ganga receives N200 cm/year and that of
the Ramganga receives 160–180 cm/year. Most of this rainfall is concentrated during July–September, while the rest of the year is dry or
has little rainfall. The monsoonal rainfall therefore exerts an important
control on the hydrology and sediment flux in the basin. The winter
39
temperature varies from 2 to 15 °C and the summer temperature from
25 to 45 °C (Singh, 1994). The hydrology of the Ganga River is strongly
controlled by the monsoonal rainfall, and the proportion of glacial melt
where the Ganga enters the plains, although debatable, has been estimated at 15% (Das Gupta, 1975).
Topographic analysis based on SRTM data (Fig. 2a) and a regional
geomorphic map of the study area (Fig. 3) show that the modern
Ganga River flows in an asymmetric valley (Fig. 2a). Field observations also confirm that the right channel bank is incised, exposing a
15–20 m high cliff line (Fig. 2c) that also marks the right valley margin. The left channel bank borders a flat-lying active floodplain tract
(Fig. 2d) with palaeochannels and meander scars. Fig. 3 suggests
long-term southwest migration of the channel (see also Roy and
Sinha, 2005, 2007). The left valley margin does not form a prominent
geomorphic feature but is delineated where the active floodplain
passes northeastward into an inactive floodplain belt.
Fig. 3 shows five major geomorphic units in the study area, namely
(i) the major active channel belt of the present Ganga and its large
tributary, the Ramganga, currently about 5–7 km and 2.5–3 km
wide, respectively; (ii) the active floodplains of these major channels
within the valley margin; (iii) the active minor channels and floodplains of other smaller tributaries; (iv) wide inactive floodplains
northeast of the Ganga; and (v) a slightly dissected interfluve surface
southwest of the Ganga that extends as far south as the Yamuna River
near the southern margin of the plains (Gibling et al., 2005; Sinha et
al., 2005a). The southern margin is strongly gullied and degraded,
forming irregular cliffs along the river.
a
Head-watershed of
the Ganga River
c
Present Glaciated area (206 km2)
80° 04'
LGM Glacier extension based on Owen et
IRS 1-D Band 4
al.(2002) (333 km2)
Ganga head watershed (total area 2435 km2)
Kannauj
3
1
Hardoi
2
26°50'
6
5
4
7
10 km
Kanpur
b
Fig. 1. (a) Location map of the study area in the western Ganga plains, Uttar Pradesh, India. (b) Band 4 satellite imagery (IRS LISS 3, dated April, 2000) showing active and palaeo-fluvial features having different radiance property. Note drill core location along two transects from left to right valley margin. Dashed line indicates left valley margin. Right valley
margin is very close to the Ganga River. (c) Headwaters of the Ganga River showing the presently glaciated area and the extent of the LG glaciation.
40
N.G. Roy et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 356–357 (2012) 38–53
77o50’
30o
a
Haridwar
b
180
A’
yan
mo
un
tain
fron
t
Delhi
Height (m msl)
A
N
Him
ala
A’
175
Kali
170
165
160
155
0
20
A
60
80
100
120
160
B
B
Altitude (m)
155
Kannauj
Kali
Ganga
B’
Gomti
150
145
140
135
100 km
130
Kanpur
c
40
Distance (km)
B’
Elevation
(m above msl)
Ramganga
Ganga
Gully erosion
81 00
0
20 40 60 80 100 120 140 160 180
Distance (km)
d
Bridge
Bridge
Cliff line
20m
300m
Fig. 2. (a) Digital Elevation Model (DEM) from SRTM source acquired by USGS in Feb, 2000 showing the valley and interfluve. The box shows the area of study. (b) Cross-sections
across the valley show that the Ganga River flows in a broad valley (approximate width 10–15 km) and both the Ganga and its tributaries are incised. (c, d) Field photographs of the
Ganga River close to Kannauj; note that the left bank is incised (left-hand photo) whereas the right bank is a flat plain (right-hand photo).
Present-day Ganga and Ramganga channel belts are braided, but the
Ramganga shows significant sinuosity as well, particularly upstream of
its confluence with the Ganga. The Ganga with a higher ‘braid-channel
ratio’ (Friend and Sinha, 1993) flows in a nearly straight reach near Farrukhabad and a change in flow direction occurs downstream of Fatehgarh from NW–SE to nearly E–W down to its confluence with the
Garra River, after which it resumes its NW–SE trend. The active floodplain of the Ganga is generally wider than that of the Ramganga, but
their widths vary significantly along the channel. Among the active
minor channels, the sinuous Garra River is slightly migratory. A large
part of the study window is occupied by numerous inactive channels
and floodplain features such as meander scroll bars, cut-offs (both
neck and chute cut-offs), and abandoned channel belts.
3. Methods
3.1. Cliff section mapping and raising of drill cores
The valley-margin stratigraphy was studied in an exposed section
1.3 km long and 10–20 m high along the Ganga bank at Mehdipur
Ghat close to Kannauj. Continuous overlapping photographs were
taken from a boat to prepare a photomosaic on which major lithological
units and breaks were traced. Ten stratigraphic logs at carefully chosen
locations were prepared for documentation of grain size, sedimentary
structures and features, recorded on centimetre scale to establish lateral
correlation between logs and across the photomosaic. Based on the
availability of reliable information on sedimentary structures and geometry, the facies interpretations (Table 1) are considered robust.
The valley-fill deposits were studied through seven cores collected along two transects 20 km apart downstream from Kannauj, with
core spacings of 3 to 8 km along each transect (Fig. 1). Three cores
were raised from each transect, and a single core was collected
10 km downstream of transect 2. The core locations were selected
where past fluvial features were observed on the geomorphic map.
Prior to drilling, a ground electrical resistivity survey was performed
with a Schlumberger array to have prior information about the occurrence of major channel bodies before selecting the drill-core locations. Resistivity methods used in the research are outlined in
Yadav et al. (2010). Locations of drill cores were determined by a
hand-held GPS accurate to less than a metre, and core tops were
plotted on the topographic profile based on SRTM-based digital elevation maps.
N.G. Roy et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 356–357 (2012) 38–53
41
Geomorphic Map
Legend
79 55'
Salt encrusted area
Sand hills
Lake area
Water body
Dry channel
Meander cut off
and scroll
4
5
3
0
1
2
25 km
27 00'
Geomorphic Units
Minor ch. +FP
Ch. Belt
Unit1:
Unit3:
Flood plains
Unit2:
Dissected surface
Unit5:
Inactive channels
Unit4:
Fig. 3. Geomorphic units of the Ganga plain between Farrukhabad and Kannauj, based on interpretation of remote sensing images and field checks. 1, active channel belt; 2, active
floodplain (younger and older floodplain); 3, minor active channel and floodplain; 4, inactive floodplain; 5, slightly dissected surface. The left valley margin is represented by the
boundary between Units 2 and 4, and the right margin by the boundary between Units 2 and 5.
Redrawn from Roy and Sinha, 2005, 2007.
Drilling employed a diamond core drilling bit involving a double
barrel core tube. A PVC pipe was inserted inside the core tube. In relatively stiff muddy strata, a slow rotary method was used with close
to 100% core recovery. In unconsolidated sandy strata, a combination
of rotary and percussion methods was used to arrest the core with
slightly lesser core recovery (70–80%) including a slight compaction
due to percussion. The core obtained in the PVC pipes was taken to
the laboratory and split into halves. One half was preserved and the
other half was used for logging and sampling. Geoelectrical logging
(SP and resistivity) was carried out for all drill holes after the coring
was completed. Apart from providing the first-level lithological characteristics, the geoelectric logs also helped in reconstructing facies in
areas of core loss during drilling.
The cores were logged at centimetre scale. Munsell colour chart
(Munsell colour chart, Year 2000) was used to describe colour. Textural description included grain size (visual estimation), compactness, primary structures and lamination, and grading. Special
features include carbonate nodules (kankar) and rhizoconcretions,
and dark nodules probably rich in iron and manganese. Dark
(black), drab (grey), and red-brown soft mottles in silt/clay were
noted, as were reworked carbonate nodules, shell material, burrows
and root fills. The contacts between units were noted as sharp, gradual or wavy. Because the cores were more difficult to characterise and
interpret than the outcrops, a separate facies scheme was utilised
(Table 2), and description and a brief interpretation are presented
together.
3.2. Luminescence dating
Samples for luminescence dating were collected in light-proof PVC
pipes, and the locations were marked on the cliff section and on borehole logs (Figs. 4, 5). From each sample, a coarse grain-size fraction
(63–90 or 90–180 or 180–250 μm) of quartz and feldspar was
extracted and was further processed following standard laboratory
42
N.G. Roy et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 356–357 (2012) 38–53
Table 1
Facies, description and interpretation of valley-margin interfluve sequences, based on exposures at Kannauj.
Litho-facies
Floodplain facies (IF)
IF1. Clayey silt with greater than
30% carbonate nodules.
IF2. Clayey silt with few carbonate
nodules (b5%).
IF3. Silty clay with less than 10%
carbonate nodules
IF4: Sticky clay with rhizoconcretion
and carbonate nodules
IF5: Laminated sandy silt with few
small carbonate nodules.
IF6. Massive sandy silt with N10%
carbonate nodules.
IF7. Calcrete
Aeolian facies
IE: Very fine sand and silt
Channel facies
IC: Very fine micaceous silty
sand, laminated
Description
Interpretation
Dark brown clayey silt. Reddish brown to dark
brown, yellow, grey mottles are present. Large
carbonate nodules (2 to 5 cm), comprised N30%
of the facies. Minor dark mottles and black Fe–Mn
nodules are present.
Pale brown clayey silt, locally silt rich. Drab mottles,
vertical to sub-vertical drab pipes dispersed or in
horizons. Small carbonate nodules (mm size) comprised
b5% of the facies. In places rhizoconcretions comprise
up to 30%.
Yellow brown silty clay with less than 10% kankars and
minor rhizoconcretions. Black and drab mottles and plant
roots are present.
Dark yellow tight silty clay, minor kankars but many
rhizoconcretions and mottles (red-brown) are present,
strongly cohesive. Efflorescence present on weathered
surface. Fragmented gastropod shells at the base.
Dark yellow to pale yellow sandy silt, weakly cohesive
materials with poor lamination preserved. Kankars comprise
b10% of the facies. Large grey mottles and roots. Black Fe–Mn
nodules are common.
Pale yellow massive, tight sandy silt with scattered kankars
(mm size). Yellow brown mottles and dark Fe–Mn nodules present.
Coalesced nodules a few cm to dm in diameter, nodules are
sand to silt particles bound by carbonate cement. Carbonate is
mostly micritic, but locally has coarser sparry fabric. Nodules
interspersed with mud, fine sand and silt. Rhizoconcretions
locally are up to 30 cm long. Kankar nodules are disseminated
in the floodplain mud, while kankar lumps (centimetre-scale-thick)
are laterally amalgamated and persistent carbonate layers between
floodplain layers. The carbonate nodules from floodplain facies
show higher δ13C values compared to those in eolian facies
(average δ13C = − 5.86 in floodplain and δ13C = − 2.76 in eolian facies).
Floodplain deposits with soil carbonate formation.
Variable degrees of pedogenesis. Pedogenesis and
floodplain aggradation occurred alternately.
Commonly occur as thick interfluve floodplain
of a major river.
Floodplain deposits with moderate degree of
pedogenesis as floodplain deposits (b 1 m thick)
of local river.
Shallow water bodies in a large depression and
abandoned channel areas; minor pedogenesis,
but strong carbonate diagenesis.
Floodplain deposits. Weakly pedogenised, Fe–Mn
nodule indicates gleying effect.
The nodular form, evidence for roots, and
intraformational erosion suggest that the
calcretes are largely pedogenic and have
variably attained Stages II–III (near coalescence)
and Stage IV (coalesced nodules) of
development (Machette, 1985).
δ13C indicates that floodplain facies were deposited
with greater water availability as indicated by high
C3:C4 plant ratio, whereas eolian facies were
deposited with less water availability supported
by low C3:C4 plant.
Off-white silt to very fine micaceous sand, moderate sorting of
grains, no recognisable stratification, non-cohesive with minimal
clay and plant roots and few carbonate nodules present.
Windblown sand and silt from nearby source.
Very fine micaceous sand, basal part show curved laminae, finer
particles in the upper part.
Minor channel deposits.
treatments (Stokes, 1992) to isolate quartz and feldspar. A single-aliquot regenerative dose protocol (Murray and Wintle, 2000) was
employed to yield a palaeo-dose estimate from quartz and feldspar
after successfully passing standard tests (Wintle and Murray, 2006)
for reliability. The distribution of palaeo-dose estimates from individual
aliquots was fitted with a normal distribution to obtain the mean
palaeo-dose. The values of palaeo-dose obtained from the feldspar
were underestimated because of anomalous fading (Aitken, 1985;
Huntley and Lamothe, 2001). Consequently, a fading correction was applied to the estimated palaeo-dose based on a single aliquot regenerative dose protocol following the experimental approach of Auclair et
al. (2003). Dose rates were measured using high-resolution gamma
spectrometry with a detector resolution of 2 keV. Other pre-treatments
to minimise overestimation of dose rate included sealing of samples to
avoid 222Rn loss, heating to avoid absorption of radiation by organic
matter and water, and grinding to b2 μm for homogenization. An internal dose rate of 0.06 ± 0.03 mGya−1 for quartz and 0.8 ± 0.3 mGya −1
for feldspar, and a depth-dependent cosmic ray dose contribution
were applied. The dry dose rate was thus computed for these relatively
coarse grains and corrected for the ambient moisture. The age was calculated by dividing the mean palaeo-dose by dose rate for each sample.
Ages up to 28 ka were determined based on optically stimulated
luminescence (OSL) where blue light was used as a stimulation
source. For older sediments, the signal was saturated, and isothermal
luminescence (ITL) and infrared stimulated luminescence (IRSL)
were employed. In addition to 21 finite dates, four saturated samples
yielded minimum OSL dates, and these were recorded as supplementary information. In the case of ITL, the stimulation source is constant
heat (Jain et al., 2007), and for IRSL the stimulation source is infrared
luminescence. For all cores and sections with multiple dates, the dates
are younger upward, without apparent dating anomalies. For samples
that yielded minimum OSL dates, ITL and IRSL methods yielded older
dates or, in one case, a date within error range of the minimum OSL
date.
4. Valley-margin section at Kannauj
4.1. Facies description
The section is composed of mud and sand sheets with decimetreto metre-scale thickness (Fig. 4). Three depositional groups were
identified, namely floodplain (IF) with seven component facies, aeolian (IE) and channel (IC) with one facies each (Table 1). Five stratigraphic units were identified on the photomosaic and field logs (Fig.
4a, b). The units can be correlated for the length of the exposure, although Unit 5 has been eroded to the south.
N.G. Roy et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 356–357 (2012) 38–53
43
Table 2
Luminescence dates. OSL = optically stimulated luminescence; ITL = isothermal luminescence; IRSL = infrared stimulated luminescence.
Dose rate
(Gy/ka)
OSL 2 × Do
(Gy)a
ITL De
(Gy)
OSL age
(ka)b
ITL/IRSLc age
(ka)
(a) Valley margin samples (sample positions are height from river level)
063433
Log 2
Unit 4
1.0
5±4
063434
Log 2
Unit 4
3.60
3±4
063435
Log 2
Unit 3
4.60
3±4
063436
Log 2
Unit 3
5.10
3±4
063443
Log 10
Unit 4
5.00
8±4
063437
Log 2
Unit 1
10.10
37 ± 10
063438
Log 2
Unit 1
11.00
31 ± 10
063439
Log 2
Unit 1
12.10
29 ± 10
063441
Log 9
Unit 1
19.00
29 ± 10
063442
Log 9
Unit 1
20.00
30 ± 10
063446
Log 10
Unit 1
10.00
22 ± 10
4.45 ± 0.20
4.33 ± 0.21
4.33 ± 0.21
4.20 ± 0.20
5.15 ± 0.23
3.62 ± 0.14
3.63 ± 0.15
2.75 ± 0.24
3.35 ± 0.29
3.80 ± 0.16
3.80 ± 0.35
103.32 ± 8.50
105 ± 7
105 ± 7
94.56 ± 12.9
76.8 ± 4.34
69.66 ± 3.62
63.44 ± 3.30
79.70 ± 5.87
69.02 ± 2.72
80.9 ± 4.96
77.22 ± 8.66
153 ± 10
150 ± 1
150 ± 1
–
136 ± 8
150 ± 7
–
245 ± 15
–
–
378 ± 19
N 23 ± 3
N 24 ± 3
N 24 ± 3
N 23 ± 3
N 15 ± 2
N 19 ± 2
N 17 ± 2
N 29 ± 3
N 21 ± 2
N 25 ± 3
N 20 ± 03
34 ± 4
–
35 ± 4
–
26 ± 3
41 ± 4
–
89 ± 10
–
–
100 ± 11
(b) Core samples (sample positions are depth from the surface)
053412
Core 1
Unit 5
7.50
053413
Core 1
Unit 5
12.25
053414
Core 1
Unit 1
32.65
053404
Core 2
Unit 5
3.87
053405
Core 2
Unit 2
23.70
053406
Core 2
Unit 2
31.3
053408
Core 3
Unit 3
21.45
053407
Core 3
Unit 1
9.80
053415
Core 4
Unit 5
3.16
053416
Core 4
Unit 2
13.22
053417
Core 4
Unit 1
28.50
053409
Core 5
Unit 5
3.65
053410
Core 5
Unit 5
9.4
053411
Core 5
Unit 3
18.83
053402
Core 6
Unit 4
3.0
053403
Core 6
Unit 4
13.80
053401
Core 6
Unit 4
19.70
063447
Core 6
Unit 4
21.20
053418
Core 7
Unit 4
5.56
053419
Core 7
Unit 4
22.70
2.47 ± 0.09
1.99 ± 0.08
2.20 ± 0.12
2.60 ± 0.10
2.85 ± 0.11
3.12 ± 0.29
2.12 ± 0.08
2.09 ± 0.08
2.79 ± 0.10
3.42 ± 0.12
3.94 ± 0.16
2.35 ± 0.08
2.37 ± 0.09
3.92 ± 0.14
2.58 ± 0.09
2.38 ± 0.10
3.08 ± 0.12
3.07 ± 0.12
2.41 ± 0.09
3.35 ± 0.13
0.45 ± 0.08
4.32 ± 0.14
121.68 ± 8.10
3.22 ± 0.15
108.72 ± 9.53
55.22 ± 4.49
60.22 ± 2.49
3.40 ± 0.14
2.40 ± 0.06
96 ± 5
65.41 ± 0.89
4.47 ± 0.18
5.93 ± 0.27
91.93 ± 10.16
30.12 ± 1.22
33.61 ± 1.35
41.71 ± 3.81
46.47 ± 2.65
31.40 ± 1.22
66.26 ± 2.08
84 ± 13
–
–
–
–
–
–
–
62 ± 9
227 ± 73
–
–
–
–
–
100 ± 11
–
–
42 ± 0.85
–
0.18 ± 0.02
2.18 ± 0.12
N 55.33 ± 4.81
1.24 ± 0.08
N 38.09 ± 3.71
N 17.69 ± 2.23
28.42 ± 1.70
1.63 ± 0.10
0.86 ± 0.04
N 27.14 ± 1.72
N 16.59 ± 0.77
1.90 ± 0.11
2.51 ± 0.16
N 23.44 ± 2.76
11.67 ± 0.67
14.10 ± 0.89
13.56 ± 1.74
15.12 ± 1.09
13.01 ± 0.73
N 19.81 ± 1.07
–
–
–
–
37 ± 2*
74 ± 3*
–
–
–
–
61 ± 2*
–
–
–
–
–
–
–
–
–
Sample number
a
b
c
Log/core
Unit
Height/depth
(m)
Water content
(%)
39 ± 4
32 ± 4
27 ± 4
27 ± 4
32 ± 4
22 ± 4
32 ± 4
29 ± 4
42 ± 4
42 ± 4
27 ± 4
39 ± 4
37 ± 4
45 ± 4
36 ± 4
16 ± 4
30 ± 4
30 ± 4
44 ± 4
26 ± 4
Represents the saturation limit for saturated samples.
Represents the minimum OSL age for saturated samples.
Age data shown with asterisks are IRSL ages after fading correction.
Unit 1 (base not exposed) is 3.8 m thick with modest lateral variation, and comprises weakly stratified clayey silt (facies IF1and 2).
Scattered to concentrated kankars of various sizes (mm to cm) are
relatively small in the upper part, where rhizoconcretions up to
30 cm long and traces of oxidised rootlets are also present. At the
unit top, the clay is more silty and has vertical drab pipes (maximum
10 cm), mottles, and tiny fragments of gastropod shells.
Unit 2 is composed of sticky clay (IF4) and has a disconformable
contact with Unit 1, defined by a change from clayey silt to clay
with high moisture content. The maximum thickness is 30 cm, and
the unit is locally missing or very thin (2 cm). Kankars (2–3 cm)
(IF7) are common, with some rhizoconcretions (5–10 cm long) and
root traces.
Unit 3 is 1 m to N 2 m thick, consisting mainly of loose, very fine
micaceous sand (IE). Small kankars (2–3 mm) (IF7) are scattered in
the basal part. Two sand sheets, separated by a few centimetres of
sandy silt (IF5), lack internal stratification and are weakly
pedogenised.
Unit 4 is a 2 m-thick, heterogeneous deposit that comprises sandy
silt and silty clay with a few gastropod fragments (IF3 and IF5). Small
kankars (IF7) are scattered through the unit, and discontinuous kankar lumps 5 to 20 cm long are aligned at several horizons (dashed
lines in Fig. 4a). In the northeastern area, a small channel deposit
2 m thick and 5 m wide of very fine sand (IC) is present.
Unit 5 is a rhythmic light and dark interbedding of fine silty sand
(IE) and dark yellow, poorly laminated sandy silt (IF4). Kankar (IF7) is
common as disseminated small nodules and as larger lumps at some
levels. In both Units 4 and 5, mottles are relatively sparse compared
with Unit 1.
4.2. Facies interpretation
Unit 1 represents a distal floodplain environment as inferred from
muddy deposits that are moderately pedogenised. The presence of a
considerable proportion of kankar, some of which forms more concentrated layers, suggests periods of minimal deposition or discontinuities. This unit generally fines up but slightly coarser sediments at
the top suggest that the area was closer to a fluvial conduit, or experienced a high-energy flood peak. The presence of gastropod shells at
the top of Unit 1 suggests standing water.
Unit 2 is interpreted to represent a widespread body of shallow
water, possibly a floodplain marsh. The presence of small rhizoconcretions and root traces suggests that herbaceous or small woody
plants grew in the area, implying shallow water or periodic drying
up. This implies the presence of local floodplain wetlands or ponds
rather than a persistent deeper lake (Glooschenko and Martini,
1987). Unit 2 developed over an irregular floodplain surface that appears to mark a discontinuity.
A major change is marked by Unit 3 which consists of aeolian deposits, the base marking a discontinuity. An aeolian interpretation is
based on the predominance of near-structureless silt to very fine
sand with little evidence of pedogenesis; similar units are locally present across the southern Ganga Plains (see descriptions in Gibling et
al., 2005, 2008).
Unit 4 is made up of stratified floodplain deposits and may indicate a floodplain environment closer to a major conduit. A composition of silty clay, clayey silt and sandy silt, together with poorly
developed palaeosols and the fill of a shallow channel, suggests deposition as a natural levee (Tyler and Ethridge, 1983; Thomas et al.,
44
N.G. Roy et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 356–357 (2012) 38–53
a
b
c
CL
Soil
U5
U4
U3
U4
18 m
10 m
U3
U2
U2
U1
U1
15 m
15 m
Fig. 4. (a) Ganga valley-margin section at Mehdipur Ghat, Kannauj. The section has five major stratigraphic units defined as Units 1 through 5 (U1 to U5) from base to top. The upper
part of the section in places is modified by human activity (Cultural Level, CL) or capped by modern soil. (b, c) Field photographs showing units and facies in the valley margin cliff
section.
2002). Such interbedding of floodplain facies with varied grain size, in
addition to occasional breaks in deposition manifested as kankar
layers, suggests deposition during flood cycles of varied intensity.
The channel body may represent a crevasse through the levee or a
small floodplain channel. Unit 5 is interpreted as a package of alternate aeolian, lacustrine and floodplain deposits, as identified by the
presence of weakly cohesive sandy silt, fine lamination and plant
roots.
4.3. Chronology and architecture
Table 2 lists sample dates from the valley-margin section and from
cores (discussed later). Quartz fractions of 90–180 μm from Unit 1
yielded ITL dates of 89 ± 10 ka from the lower parts and one date of
41 ± 4 ka from the unit top at the position of Log 1. Another sample
from Log 9 position gave a date of 100 ± 11 ka for this unit. Unit 3
yielded an ITL date of 35 ± 4 ka. Unit 4 yielded ITL dates of 34 ± 4 ka
and 26 ± 3 ka, the latter from the topmost beds of the unit. Although
all OSL samples yielded saturated signals, the minimum dates of 15 to
29 ka suggest that the exposed section entirely predates this period.
For Unit 1, dates between 100 and 41 ka indicate that the strata
represent much of Marine Isotope Stages (MIS) 3, 4 and 5. Although
too few dates are available to estimate sedimentation rate, the small
thickness (4 m) of the unit, along with the fine grain size and abundant kankar, suggests slow floodplain aggradation and probably the
presence of discontinuities. Although the topmost beds were not
dated, the bulk of sediment accumulation in Units 2–5 (8 m) apparently represents a relatively short interval of more rapid
N.G. Roy et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 356–357 (2012) 38–53
accumulation, with much coarser sediment, between about 39 and
23 ka (including error ranges), late in MIS 3 through to early MIS 2.
The repeated intercalation of thin aeolian and lacustrine deposits in
Unit 5 may reflect climatic fluctuation. At present, there is no indication of deposition since about 23 ka, and the cliff-line may have been
degrading through this time.
5. Ganga Valley cores
5.1. Facies description and interpretation
Three channel facies are present (Table 3), namely sandy silt/silty
sand (VC1), fine to very fine sand (VC2) and medium sand (VC3).
The coarser channel sands have ‘salt and pepper’ texture due to the
abundance of dark, dense minerals, which is characteristic of Himalayan-derived sediment. Among the channel facies, the finer grained
facies VC1 generally occurs as relatively thin bodies up to 3 m thick
within muddy successions and contains mottles and sparse carbonate
nodules. These bodies are inferred to represent small tributary or
floodplain channels. In contrast, the fine- to medium-grained facies
VC2 and VC3 form continuous sand bodies up to 10 m thick that are
interpreted as the deposits of major trunk rivers.
45
Floodplain facies form only a modest proportion of most drill cores
and consist mainly of silty clay with pedogenic features that include
dark brown, yellow brown, and grey mottles, and carbonate nodules
that are disseminated or locally concentrated in bands. The floodplain
facies are classified into grey silty clay (VF1), pale yellow silty clay
(VF2), brown silty clay (VF3), and prominent bands of carbonate nodules (VF4) (Table 2). Aeolian deposits (facies VE) consist of structureless, soft clay-rich silt that varies from light yellowish brown to pale
yellow. Carbonate-cemented layers and thin clayey silt and silt laminae are characteristic, probably reflecting the presence of low
mounds of wind-blown material. This facies appears more clay rich
and better stratified than inferred aeolian facies in the cliff section,
and the aeolian interpretation must be considered provisional.
Lacustrine deposits (facies VL) were identified as a 4-m-thick, organic-rich layer in one log (core 7) above thick channel sand, and are
present in other logs as a thin layer above channel fills. The lacustrine
interpretation is based on the sheet-like geometry, the fine lamination, and the distinctive lithology of dark olive grey, stiff and sticky
clay. This facies is similar to the floodplain marsh deposits in Unit 2
in the cliff section. Carbonate nodules and locally prominent mottling
point to a moderate degree of pedogenesis. This facies is attributed to
temporarily or permanently abandoned channels, probably under
backswamp/marshy conditions.
a
Legend- valley margin
>28 ka valley edge
Late Holocene
Channel
Legend- valley fill
Channel/floodplain
(MIS 3-2)
Pre-LGM
Channel
MIS 4/5 succession
Fig. 5. Drill core facies logs along transect 1 (a), transect 2 (b) and longitudinally along the valley (c) showing stratigraphic units and lateral correlation. The modern Ganga Valley is
shown by a solid thick line.
46
N.G. Roy et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 356–357 (2012) 38–53
b
Legend- valley margin
MIS 3 valley edge
Pre-Holocene
Channel
Late
Holocene
Channel
Legend- valley fill
Channel/floodplain
(MIS 3-2)
Channel margin/
Floodplain
(MIS 3-2)
Intra- valley
Floodplain (MIS 2)
MIS 4/5 succession
Fig. 5 (continued).
5.2. Chronology and architecture
Fig. 5 shows seven drill core logs in two cross-valley transects (Fig. 5a, b) and one transect along the valley (Fig. 5c;
see locations in Fig. 1b). The cores consist of channel bodies
separated by floodplain, lacustrine and probable aeolian deposits that cover a time span from about 74 ka to 0.1 ka
(Table 2). Five stratigraphic units were identified on the basis
of lithology and age. In contrast to the sheet-like nature of valley-margin units, the cored units appear to be lensoid and locally incised into underlying units, making unit designation
less reliable. Although the seven cores are insufficient for a
complete valley-scale correlation, some units can be traced between the two transects.
Unit 1 (cores 1, 2, 4 and 5, base not penetrated) consists of 2 m of
fine to very fine channel sand (facies VC2). The channel sand yielded
IRSL dates of 74 ± 3 ka and 61 ± 2 ka, and a minimum OSL date of
55 ka (Table 2) suggests that this channel activity corresponds to topmost MIS 5 and MIS 4. The considerable difference in dates from cores
2 and 4 suggests more than one episode of channel activity, possibly
with associated incision. Unit 1 is present in cores close to the modern, southern Ganga Valley margin, but in neither transect were the
lowermost channel sands fully penetrated, and the unit may be widespread below the modern Ganga Valley.
Unit 2 consists of alternate floodplain (VF1–VF3) and aeolian (VE)
beds. Thin sand layers (splays) yielded an IRSL date of 37 ± 2 ka
which corresponds to late in MIS 3, and two OSL samples provided
minimum dates of 27 ka and 23 ka (Table 2). Grey mottles indicate
a weak degree of pedogenesis. In core 2, a sharp break within Unit 2
is marked by the presence of a concretion zone at 25 m depth. Moderately pedogenised muds suggest slow deposition or periods of
non-deposition. In core 2, a kankar horizon at the 25 m level
(Fig. 5a) above 3 m of undated floodplain deposits may represent a
significant discontinuity. Although shown within Unit 2, fine sediments below this kankar level are provisionally taken to mark the termination of MIS 4/5 deposits.
New channel activity during MIS 3 (37 ka) is recorded from thin
sands below a thicker (3 m) channel body in core 2 (Fig. 5a). Channel
deposits are also recorded prior to 27 ka (a minimum OSL date) in
core 4 (Table 2). Both of these channel sediments are capped with
N.G. Roy et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 356–357 (2012) 38–53
47
Longitudinal Transect
c
Pre-Holocene
Channel
?
Late-Holocene
Channel
Intra-valley
Floodplain/channel
(MIS 2)
Pre-LGM
Channel
>20 ka
?
?
Fig. 5 (continued).
thick, strongly pedogenised overbank deposits 7 m thick, suggesting
reduced sedimentation late in MIS 3. In contrast, floodplain deposits
lower in Unit 2 show less intense pedogenesis, probably within cumulative floodplain soils. Similar weak pedogenic activity during
floodplain aggradation has been documented from eastern Ganga
Plains interfluves (Mohindra et al., 1992; Sinha and Friend, 1994).
Unit 3 consists of 3 m of medium sand (VC1) and we interpret
this as the top of a thick channel body that extends below the
base of core 3, the only place where it is recorded. An OSL date
of 28 ± 2 ka from core 3 (Table 2) suggests deposition late in
MIS 3. The channel sands lie close to the left margin of the present Ganga Valley and their elevation indicates that they are incised into older strata.
Unit 4 constitutes one or more major channel bodies composed of
very fine to fine sand (VC1 and VC2), recorded in cores 6 and 7 only.
In core 6, two distinct sand bodies are separated by 5 m of floodplain
mud (VF2). Five dates bracket the channel sediments between 15 and
11 ka (Table 2), representing MIS 2 to early MIS 1 (latest Pleistocene
to earliest Holocene) and suggesting that 20 m of strata were deposited in as little as about 4000 years. The channel body is interpreted as
a braided-fluvial deposit, and it lies close to the left valley margin
where it appears deeply incised (Fig. 5b).
Unit 5 is represented by a sand body (VC1–VC3) more than
15 m thick, capped by thin muddy deposits that were probably
laid down within abandoned channels. Seven dates range between
2.5 ± 0.16 ka and 0.18 ± 0.02 ka (Table 2), representing late Holocene channel activity. These channel deposits at the top of the
succession are widespread and were penetrated in five cores, variously incised into Units 2, 3 and 4 (Fig. 5b). The sediments
resemble those of the modern river, and they are interpreted as
braided-fluvial deposits.
6. Synthesis of Ganga Valley and valley-margin records, in comparison with the monsoon record
The Ganga Valley-margin exposures and subsurface deposits are
illustrated schematically in Fig. 6, along with proxy records for monsoon intensity based on stacked core records for the Arabian Sea
(Clemens and Prell, 2003) and modelling results for the SW Indian
Monsoon (Prell and Kutzbach, 1987). The latter curve is closely linked
to insolation trends, and changes in modelled intensity approach 30%
of present values. The two proxy records differ in detail but show
broadly similar high- and low-intensity periods, fluctuating over periods of a few thousand to about 20,000 years. The fluctuations
show general accord with marine isotope stages determined from offshore cores (Waelbroeck et al., 2002; Fig. 6).
In interpreting our records, we have drawn strongly on the range
of dates provided by the 13 finite OSL dates and 9 ITL and IRSL dates
to establish periods of channel aggradation and incision. Although
more dates are necessary to provide a comprehensive understanding,
the dates were obtained from precisely documented strata, commonly to test the age of key boundaries, and they suggest several distinct
periods of fluvial activity, which are especially clear from about 30 ka
onwards.
The interfluve successions are characterised by overbank fines,
whereas the valley succession comprises channel bodies and fines,
and the relationship between them is a key issue. The valley-margin
deposits have a sheet-like geometry and aggradational style, in
48
N.G. Roy et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 356–357 (2012) 38–53
Table 3
Facies group, facies, description, association and interpretation of cores below the Ganga Valley.
Facies
Floodplain facies
VF1. Grey silty-clay
VF2. Pale yellow silty-clay
VF3. Brown silty clay
(with/without mottles)
VF4. Carbonate nodules
Lacustrine facies
VL. Dark silty clay
Aeolian
VE. Yellow clayey silt
Channel facies
VC1. Yellow silty sand
VC2. Very fine sand and
fine sand
VC3. Medium sand
Description
Abundance and associated facies
Interpretation
Silty clay, structureless, moderately tough,
sticky due to high clay concentration, rare
mottles. Tiny kankar nodules and Fe–Mn black
coated nodules (2–3 mm size) are rare. Munsell
colour brown grey (2.5Y6/2) to grey (5Y6/1)
with a greenish tinge.
Soft silty clay with high clay proportion, sticky
and tough with uncommon reddish brown and
yellowish brown mottles in Core-5. Kankars
from 1 mm to 2 cm are scattered and fairly
abundant. Munsell colour: pale yellow silty
clay (5Y8/4 to 5Y7/3)
Soft clay and silt mixture; cohesive, sticky
nature — due to relatively enrichment in clay.
Mottles are common to few, scattered to
concentrated. Black mottles are common
and yellow brown mottles are frequent.
Fe–Mn nodules and kankars are abundant.
Munsell colour: brown silty clay
(2.5Y 5/4 to 2.5Y 6/4)
Layers of concentrated carbonate nodules
(diameter varies from 2 to 3 cm up to 5 cm),
rounded to angular. Most nodules are
carbonate-cemented sand.
Present in three cores (5, 2, 4), in units
with maximum thickness up to 1 m (Core-2),
forms incomplete cycle; commonly truncated
below base of channel sand.
Floodplain deposits; dark colour suggests
deposition under reducing condition
Abundant facies, present in Core-5 and
Core-4. Maximum thickness is 2.7 m. This
facies frequently interrupted by kankar zones
(20–30 cm thick). Commonly associated with
silty sand (VC1) and locally clayey silt (VE).
Floodplain deposits, weakly pedogenised
with soil carbonate nodules
Maximum thickness 4 m, associated with
silty sand, clayey silt facies.
Floodplain deposits; moderate to strong
pedogenesis, with abundant soil carbonate
formation, Fe–Mn nodules indicate gleying
effect.
Layers maximum 50 cm thick associated
with facies VF1, VF2 and VC1.
Layered concretion, probably of
groundwater origin.
Sheet like geometry and fine lamination,
stiff and sticky clay, maximum 1 cm diameter
kankar nodules sparsely distributed, mottles
are uncommon to common.
Munsell colour: dark olive grey (5Y3/2)
Maximum thickness 3 m, associated with
fine channel sand facies in Core-7, in other
cores thickness is a few centimetres.
Floodplain deposits with low degree of
pedogenesis formed under reducing
conditions created due to ponding within
channels. Presence of carbonate nodules
points to moderate degree of pedogenesis.
Silt with few clay, non-cohesive, low skewness
relative to other facies, small kankars sparsely
distributed, rare black nodules and yellow
brown mottles. Carbonate cemented bands,
and thin lamination of clayey silt and silt
material are characteristic. Munsell colour:
light yellowish brown (2.5Y 6/3, 2.5Y6/4) to
pale yellow (2.5Y7/3; 5Y/8/3)
Maximum thickness 3.5 m in the Cores-1 and
-5; mainly associated with floodplain facies
Weakly pedogenised sheets of massive
windblown massive sheet sand; thin
lamination probably reflects the presence
of low mounds of wind-blown material.
Silty and very fine sand, with minimal clay,
slightly micaceous, mottles and sparse
carbonate nodules. Units fine upwards.
Maximum thickness 3 m in Core 4 associated
with floodplain facies.
Very fine and fine sand mixture. Sharp break
occurred with medium sand facies. Small
kankars are sparsely distributed.
Medium sand, off-white to white, mixed with
fine and very fine sand, highly micaceous,
presence of heavy minerals gives ‘salt and
pepper’ appearance. Small kankars sparsely
distributed. Generally well sorted.
Maximum thickness 7.4 m in Core-6, commonly
associated with medium sand facies, and
channel fill top.
Maximum thickness of 10 m in Core-3, at the
base of the core reworked carbonate gravel
present.
Coarse sediment body associated with
small channel, probably a tributary or
floodplain channel. Dark yellow silt
resembling eolian facies suggests that
deposition occurred in drier climatic phase.
Thick channel body indicates large trunk
channel, probably braided deposits.
contrast to the lensoid geometry of the sub-valley deposits which implies both aggradation and incision. Although the sheet-like form of
the valley-margin deposits may in part reflect the orientation of the
cliff exposure parallel to the river course, the two successions differ
considerably in facies and are largely coeval. Thus, the two components must be viewed as representing discrete but related geomorphic regions, and the correlation between them potentially yields
considerable insight into fluvial dynamics through time. Additionally,
our cores from the Ganga Valley allow some understanding of autogenic channel relocation within the valley area.
Thick channel deposits of large trunk river
laid down under high energy condition.
Rare thin beds intercalated with floodplain
deposits are interpreted as splays from
major channel.
Five major aggradational periods can be recognised, based on
dates and error bars. They are shown as schematic cross-sections (I
to V) in Fig. 6, and their inferred age ranges are shown alongside
the proxy records for climate change. Probable periods of incision
are also shown.
6.1. Period I: MIS 5 to early MIS 3 (111–59 ka)
The lowermost floodplain deposits (Unit 1) of the valley-margin
section are bracketed between 100 and 41 ka (Fig. 4a), a period of
N.G. Roy et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 356–357 (2012) 38–53
49
Fig. 6. Ganga valley-fill stratigraphic architecture model for the Kannauj area, showing periods of incision and aggradation compared with proxy climatic records and Marine
Isotope Stages (MIS), from about 110 ka to present. Summer Monsoon Factor is from Clemens and Prell (2003), based on stacked Arabian Sea core records, and the modelled
SW Indian Monsoon record is from Prell and Kutzbach (1987). The boundaries between marine isotope stages are from Waelbroeck et al. (2002). Aggradational and incisional periods are outlined more fully in the text, which also compares the record near Kannauj with other palaeoclimate records in Asia.
nearly 60,000 years. Floodplain development during this period was
interrupted by multiple pedogenic events, possibly indicating slow
sedimentation or short-term discontinuities, and more intense pedogenesis is apparent in the topmost beds.
Below the modern valley, the oldest channel bodies (partly penetrated in Unit 1) yielded finite dates of 74 ka and 61 ka, and thus appear to be in part time-equivalent to the basal floodplain unit of the
valley margin (Fig. 6). However, a difference in elevation of 25–
30 m is recorded between the base of floodplain deposits (Unit 1) in
the cliff section (Fig. 4a) and the top of channel deposits of Unit 1
below the valley (Fig. 5a, b). This elevation difference suggests the
presence at this time of an incised valley with channels that were capable of intermittently flooding the interfluve to the south. Minor discontinuities in the valley-margin deposits are inferred from the
varied intensity of pedogenesis and may reflect the episodic nature
of this inundation. Over the past century, low- to high-stage seasonal
fluctuations of the Ganga and Yamuna rivers have often been extreme
(in excess of 20 m: Gibling et al., 2005), and the highest flows may inundate parts of the proximal interfluves, depending on the local
height of the cliffs.
An alternative interpretation is that alluvium had built up in the
area of the modern valley to a higher level than that presently
recorded, allowing frequent interfluve inundation. Valley incision
may have subsequently removed much of this sediment and detached
the interfluve from the trunk channel, reducing sediment supply and
initiating interfluve pedogenesis.
Cliff sections along the modern Yamuna River, 70 km southwest of
the study area, also expose a thick floodplain succession dating back
to ~100 ka, with prominent discontinuities (Gibling et al., 2005;
Sinha et al., 2005b; Tandon et al., 2006). Although Unit 1 in the
Ganga Valley was incompletely penetrated, the range in ages for
channel bodies suggests multiple channel events and possibly incisional discontinuities, perhaps in accord with modelled strong fluctuations in monsoon intensity in MIS 4 and 5 (Fig. 6).
In general, we infer prolonged channel activity during MIS 5 and 4
and at least some interfluve flooding, followed by a reduction in channel activity associated with pedogenesis of floodplain muds. A similar
period of persistent channel activity in MIS 5/4 was inferred for strata
in the Belan River at the southern margin of the Ganga Plains (Gibling
et al., 2008).
6.2. Period II: MIS 3 (45–30 ka)
Deposits that correspond to MIS 3 are represented by Unit 2 of the
valley-fill, where a splay associated with minor channel fills yielded a
date of 37 ka. Probable aeolian strata of pale, friable silt are prominent
in this unit. During late MIS 3, the valley-margin succession records
thin, non-fluvial deposits (floodplain marsh and aeolian sediments
of Units 2 and 3). This suggests that flow through the channel was either drastically reduced or that the trunk river migrated away from
the southern valley margin, reducing sediment supply to the area. A
period of floodplain degradation, possibly associated with reduced
50
N.G. Roy et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 356–357 (2012) 38–53
sediment supply, is recorded in the Yamuna valley ~70 km south of
Kanpur (Fig. 1a) between 30 and 40 ka, manifested as the cutting of
deep gullies and their subsequent filling with colluvium (Gibling et
al., 2005). Period II may correspond with reduced monsoonal activity,
although the two proxy records show some differences through this
period.
6.3. Period III: latest MIS 3 and early MIS 2 (30–23 ka)
Renewed fluvial activity is recorded in the cores late in MIS 3.
High-energy channel activity (gravelly sands) dated at 28 ka are
recorded near the left margin of the present Ganga Valley (Fig. 6),
and their relative depth suggests that incision occurred prior to alluviation. This channel activity is time-equivalent to a channel fill
recorded at Bithur 40 km downstream from the study area by Sinha
et al. (2007a).
In the valley-margin succession, levee deposits of Unit 4 yielded
dates of 34 and 26 ka, the latter from a small channel body. There is
some uncertainty as to whether these correspond to Periods II or III,
but the unit top (at least) is attributed to Period III because the younger date is time-equivalent to channel deposits below the valley, and
because the presence of renewed overbank deposition along the valley margin is consistent with high discharge late in MIS 3. As with
Stage I, the deposition of sediments in both geomorphic areas may reflect extreme floods from an incised valley or alluviation within the
valley to an elevation higher than presently recorded.
Goodbred (2003) suggested that high-intensity floods would have
been active during a period of strong solar insolation at about 31 ka,
associated with large-scale sediment production from the Himalayan
source area (see also Sharma and Owen, 1996; Taylor and Mitchell,
2000). The proxy and modelled monsoon curves suggest that the
south Asian climate was wetter late in MIS 3 or in earliest MIS 2
(Fig. 6).
No channel deposits are recorded below the Ganga Valley between 28 ka and 15 ka, suggesting that this interval, which includes
the global Last Glacial Maximum (LGM), experienced reduced fluvial
activity. Relatively strong pedogenesis is noted below much younger
deposits in the topmost muddy parts of Unit 2 in cores 2, 4 and 5,
and an interval of soil formation may accord with reduced activity, although these strata are not constrained by dates.
The valley-margin succession is capped by aeolian, lacustrine and
floodplain deposits of Unit 5. Although undated, these suggest a cooler and drier climate towards the LGM. Aeolian accumulation on the
valley margin may reflect “underfit” of the Ganga River during the
LGM and exposure of the river bed, from which silt was preferentially
picked up by winds. Similar relationships have been documented
from Bithur (lake and aeolian deposits at ~27 ka; Gibling et al.,
2005) in the Ganga Valley and in the Belan valley (aeolian deposits
with broken gastropod shells at ~17–20 ka; Williams et al., 2006;
Gibling et al., 2008). Evidence of reduced siliciclastic supply in the
Ganga–Brahmaputra delta (Goodbred, 2003; Weber et al., 2003) in
the later part of MIS 3 is consistent with reduced hinterland supply
or possibly some sediment storage in inland valley areas.
6.4. Period IV: late MIS 2 and early MIS 1(16–11 ka)
Channel aggradation was recorded in Unit 4 of cores 6 and 7,
where five finite OSL dates in the topmost 20 m of strata range between 15.1 and 11.7 ka. The presence of muds between channel bodies suggests channel switching in a braided depositional
environment. These thick braided-channel deposits imply high sediment supply from the hinterland. This may reflect landscape instability due to monsoon intensification following the LGM (Overpeck et
al., 1996; Fig. 6), inducing extensive slope failure and transport of hillside regolith to the valley floor (Pratt et al., 2002). The valley margin
appears to have been largely inactive and possibly degrading during
both Periods IV and V, as indicated by ITL dates and by minimum
OSL dates N15 ka, although later gully erosion (Fig. 4) may have removed much of the record.
Dates of 13.0 ± 0.7 ka and 11.7 ± 0.7 ka for the top of the Unit 4
channel bodies in cores 6 and 7 suggest that the channels aggraded
rapidly during a period of reduced discharge between 13 and 11 ka.
This interval corresponds to the Younger Dryas, for which lake records at Sanai Tal 200 km east of Kanpur document a period of reduced monsoonal strength between 13.4 and 12.2 ka (Sharma et al.,
2004). Elsewhere in the Ganga Plains, evidence for reduced monsoonal strength and decreased discharge at this time includes the presence
of aeolian and lacustrine strata at Bithur on the Ganga Valley margin
with dates from the topmost part at ~ 13 ka (Srivastava et al., 2003;
Gibling et al., 2005), gully erosion and aeolian accumulation at
Mawar on the Sengar River between 13 and 9 ka (Gibling et al.,
2005), aeolian sands in incised channel in the Belan valley between
12.2 and 13.4 ka (Gibling et al., 2008), and aeolian sands east of
Delhi dated at 12.3 ka (Tandon et al., 2006). Cool, dry conditions associated with the Younger Dryas event have been widely identified in
other Asian continental deposits (Porter, 2001).
The alluviation episode followed an earlier period of channel incision (Fig. 6). This lowering of the channel's equilibrium profile accords with a period of monsoon intensification following the Last
Glacial Maximum — a pattern observed across much of southern
Asia (Tandon et al., 2006).
6.5. Period V: MIS 1 (2.7 ka to present)
Late Holocene channel bodies aggraded across much of the present valley area, and sediments accumulated within a short period
of 2.7 to 0.1 ka, based on seven OSL dates. The base of these deposits
lies adjacent to much older strata and directly overlies MIS 3/2 deposits, suggesting that valley incision preceded filling. Rapid accumulation during the late Holocene was also observed in the Yamuna
valley, 70 km to the southwest, where a 20 m core penetrated a
12 m channel-margin succession of interbedded sand and clay, yielding two dates of 2 ka or younger from the lower strata (Sinha et al.,
2005b, 2009). Late Holocene aggradation is inferred to reflect a pronounced decline in monsoon intensity and transport capacity since
the mid Holocene (Overpeck et al., 1996; Fig. 6). In the case of the Yamuna succession, accumulation may be associated with a recent
onset of Yamuna river deposition in the area (Sinha et al., 2005b,
2009).
During this period, there is no indication of deposition along the
valley margin, which was probably undergoing degradation as
recorded in the modern gullied landscape adjacent to the valley.
The river's equilibrium profile was apparently so low in the post-LGM
period that the river was not able to overtop the interfluve valley
margin. Southward migration is manifested by the Ganga's present
position at a cliff line along the southern valley margin, indicating
widening of the valley floor.
7. Discussion
The Ganga Valley in the study area is 1200 km inland from the
tidal limit, and it is unlikely that depositional architecture was affected by sea-level changes in the past to such a distance inland (Sinha et
al., 2005b; Tandon et al., 2006; Sinha et al., 2007a,b; see also Blum and
Törnqvist, 2000). The valley deposits are entirely terrestrial facies of
the foreland-basin fill, and no marine beds have been encountered.
There is no recorded history of seismic events from this region, nor
indications of seismic-related deformation in core or outcrop as observed at other sites (Gibling et al., 2005), and basement faults are
buried under a thick alluvial cover. Some workers (Singh et al.,
1997; Agarwal et al., 2002) suggested a tectonic influence on major
river incision due to forebulge activity, but this has not been
N.G. Roy et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 356–357 (2012) 38–53
established by regional geological mapping or convincing geomorphic
evidence. It is unlikely therefore that the Ganga Valley fine-scale architecture in this area was controlled by tectonic activity and sea-level change during the late Quaternary.
Paola (2000) suggested that, over time scales of 10 3–10 5 years
(Milankovitch time scale), patterns of deposition for foreland-basin
settings reflect cyclical change in water and sediment supply due to
climate shifts; over longer time scales, the influence of basin subsidence and other tectonic factors may be more apparent. Flume experiments and field studies in the Himalayan Foreland Basin and
elsewhere have identified sequences generated as a result of alternate
low and high discharge, without any change in subsidence rates,
(Bull, 1991; Milana, 1998; Weissmann et al., 2002), and an analysis
of successions in southern Asia supports Paola's assertion (Gibling
et al., 2011).
Quaternary records suggest that the strength of the Southwest Indian Monsoon has varied significantly on a timescale of centuries to
millennia. Evidence of such variations is provided by well dated, multi-proxy records from cores in the Arabian Sea (Clemens and Prell,
2003; Rao et al., 2008) and Bay of Bengal (Chauhan et al., 2004), continental fluvial records (Gibling et al., 2005, 2008; Tandon et al., 2006;
Sinha et al., 2007a,b), and lake records (Sharma et al., 2004; Prasad
and Enzel, 2006) from northern India, as well as from records on
the Tibetan Plateau (van Campo and Gasse, 1993). For the western
Ganga Plains, Gibling et al. (2005, 2008), Tandon et al. (2006), and
Sinha et al. (2007a) inferred strong correlation of fluvial sedimentation with variation of monsoonal rainfall regime, and a similar conclusion was reached on a basinal scale by Goodbred (2003). Based on the
continental alluvial record from the southern margin of Thar Desert in
northwest India, Juyal et al. (2006) showed strong monsoonal variability over the last 130 ka, and Bryson and Swain (1981) documented a period of strong early Holocene precipitation, decreasing by
about one third by the late Holocene.
Such fluctuations in monsoon intensity should have resulted in
significant change in effective discharge of the Ganga River, causing
rise and fall of the river's equilibrium profile. These discharge changes
are likely to be manifested in alluvial architecture in the valleys and
on interfluves, especially if the magnitude of change was sufficient
for the system to cross a geomorphic threshold. The availability of a
robust proxy record for monsoonal fluctuations extending back beyond 100 ka has allowed us to compare climatic events with alluvial
events of incision and aggradation in the Ganga River area at Kannauj.
Our results suggest that, to a first order, alluvial events correspond
well with monsoonal fluctuations, and this correspondence is especially evident for the well dated record after 30 ka.
A particularly provocative conclusion is that several well-dated
periods of channel aggradation correspond to times of decreasing
monsoonal precipitation (Fig. 6): during latest MIS 3 and earliest
MIS 2 (Period III), during the Younger Dryas (Period IV), and during
the late Holocene (Period V). These are inferred to record decreased
discharge and rise in the river's equilibrium profile. In all these
cases, elevation considerations indicate that aggradation was preceded by incision, suggesting that periods of increasing monsoonal precipitation were associated with high river energy and a lowering of
the equilibrium profile. Spatial considerations suggest that these
were also times of channel relocation within the valley (Figs. 5, 6).
Occurrences of aeolian and lacustrine deposits, as well as intensity
of pedogenesis, also play an important role in “reading” palaeoclimatic changes. Earlier alluviation and incisional episodes have less
well-constrained durations, and are less readily compared with monsoonal changes.
For the latest Holocene aggradation (Period V), low effective discharge of the river channel is in accord with a wealth of data across
Asia that documents severe weakening of Southwest Indian Monsoon, evident in records of Arabian Sea upwelling (Gupta et al.,
2005; Thamban et al., 2007) and in historical records from north
51
India (Pandey et al., 2003). Low percentage of Globigerina bulloides
in sediments from the Oman coastal margin indicate reduced upwelling and a monsoonal minimum about 2000 years B.P. (Gupta et al.,
2005). Aridity records based on dolomite percentage from the Arabian Sea margin (Sirocko et al., 1993), biological productivity off the Somali margin (Ivanocho et al., 2005), precipitation records off the
Indus river mouth (Staubwasser et al., 2003) and from the southwest
Indian margin (Sarkar et al., 2000) further suggest a declining monsoon about 2000 years B.P. Such a climatic change would have reduced sediment transport capacity of the river system, triggering
aggradation. This inference is borne out impressively by the presence
of more than 15 m of late Holocene sands below a considerable part
of the Ganga Valley, deposited within a period of less than 2000 years.
Numerous studies have demonstrated that periods of increased
precipitation are marked by river incision (Porter et al., 1992; Jones
and Schumm, 1999). Pratt et al. (2002) observed that early Holocene
incision in the central Nepal Himalaya was driven by enhanced monsoon precipitation, although some deposition was also evident. Although enhanced precipitation commonly results in higher
discharge, increased sediment supply during the same period may reduce the stream's erosive power and the ability to incise its thalweg,
at least locally.
The High Himalaya source area of the Ganga and its tributaries
contains large glaciers. Meltwater apparently contributes only modest volumes to Ganga flow (15%, according to Das Gupta, 1975), but
meltwater contributions could have been greater in the past. The
areal extent and timing of glacial events in the Himalaya and Tibetan
Plateau are poorly defined, but compilation of cosmogenic isotope
dates for moraines and glacially eroded surfaces indicate that the
mountain belt experienced numerous glacial advances from MIS 5
onwards (Owen et al., 2002, 2008). However, glacier ice during the
LGM covered about 539 km 2 of the headwater area of the Ganga
River catchment (Fig. 1c), about 22% of the total area. For the central
Himalayan area in the Ganga headwaters, Owen et al. (2008)
recorded a general correlation between glacial advances and the
monsoon intensity peaks identified by Prell and Kutzbach (1987), including advances during a peak monsoon condition in MIS 3 (Fig. 6).
This correlation suggests that alluvial events are likely to be a
consequence of variations both in direct monsoonal runoff and
in meltwater supplied to the plains. In the western Himalaya, correlation is less clear, due to the availability of precipitation supplied by westerly winds. We conclude that 1) alluvial events on
the Ganga Plains are likely to reflect in part meltwater contributions, 2) meltwater pulses are likely to coincide with enhanced
monsoonal runoff, and 3) meltwater was probably not the dominant discharge factor, based on estimates of modern meltwater
contributions and past glacier area.
Episodes of aggradation and incision may be attributed to variation in transport capacity of river systems and/or sediment supply
from the hinterland (Julien, 1995; Houben, 2003) linked to changes
in catchment physiography and external forcing. Assuming that
catchment physiography and valley gradient remained relatively
unchanged during the late Quaternary, a change in the balance between sediment supply and river discharge (from hinterland and adjacent plains) should change the river's equilibrium profile on the
plains. The terms capacity-limited system and supply-limited system
have been used to explore this balance (Brown and Keough, 1992;
Lewin, 1992; Macklin, 1992; Julien, 1995; Bisson and Montgomery,
1996; Montgomery and Buffington, 1997). In a capacity-limited system, the amount of transported load is limited by water supplied to
the river, whereas in a supply-limited system, sediment in the river
is controlled by sediment availability. Goodbred (2003) emphasised
the importance of upland supply in a source-to-sink consideration
of the Ganga–Brahmaputra system, and strong variations in sediment
flux from the Himalaya are indicated (see, for example, Pratt et al.,
2002). We are currently not able to discern which events in the
52
N.G. Roy et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 356–357 (2012) 38–53
Ganga Plains reflect supply- or transport-limited scenarios. However,
late Holocene channel aggradation followed a period of high sediment supply in the Ganga Delta (Goodbred, 2003), suggesting that
deposition reflected transport-limited conditions.
Despite the availability of data from both valley-margin and
intra-valley areas, it is difficult to be certain when the Ganga River
occupied a channel attached to the adjoining alluvial tract or floodplain, and when the river occupied a valley that was detached and
rarely or never flooded the adjoining interfluves (attachment terminology from Gibling et al., 2005). Strong and Paola (2008) distinguished stratigraphic valleys (the valley-form erosional surface
preserved in the stratigraphic record) from topographic valleys (the
topography at an instant in time). As noted by Gibling et al.
(2011), the Ganga Valley in the Kannauj to Kanpur area (Fig. 1) is
a stratigraphic valley that contains smaller channel bodies or valley
fills, as illustrated in Fig. 6, with internal sequence boundaries
(bases of channel bodies and carbonate-rich palaeosols) and facies
discontinuities. It is not certain that deposition commenced in a
large, deeply incised hollow, as shown schematically in Fig. 6.
More probably, the body of alluvium laid down by the Ganga is underlain by a composite valley-fill unconformity (Blum and Aslan,
2006), and long-term southward migration of the Ganga has created
a composite, diachronous valley fill with dimensions that greatly exceed the instantaneous topography. The relocation of channels within the valley area, evident from the cores (Figs. 5, 6), may in part
record the influence of the Ramganga and other tributaries (Fig. 1)
in modifying the trunk-river position due to confluence dynamics
(Roy and Sinha, 2005, 2007).
8. Conclusions
The exposed cliff section and valley cores from the southern Ganga
Plains provide an integrated picture of valley and proximal interfluve
development for the Ganga River at Kannauj over the last 100 ka.
Comparison with proxy records for monsoon strength, which show
strong fluctuations driven largely by solar insolation, indicate a first-order fit between monsoonal fluctuations and alluvial events, with a
tendency for incision during periods of monsoonal intensification
and aggradation as monsoon strength decreased. The valley cores record episodes of aggradation during Marine Isotope Stages (MIS) 4
and 5; the mid part of MIS 3; late MIS 3 (23 to 30 ka); late MIS 2 to
early MIS 1 (11–16 ka), especially around the Younger Dryas; and in
the latest Holocene (b2.5 ka). The Last Glacial Maximum was a time
of greatly reduced fluvial activity, and no channel sediments have
yet been identified that correspond with this period.
Along the southern valley margin, bordering a wide interfluve,
there was modest accumulation of floodplain, lacustrine and aeolian
deposits, punctuated by discontinuities, through to early MIS 2. The
Ganga probably varied in its degree of attachment to the interfluve,
periodically inundating the adjoining, more elevated alluvial tract
but at other times unable to flood this zone. Thus, the Ganga probably
fluctuated from being a channel with a broad floodplain to a valley
confined by an elevated plain to the south. There is no indication
that the Ganga has inundated its southern interfluves since the end
of MIS 3.
We infer that the Ganga River experienced strong variation in its
equilibrium profile over relatively short periods (centuries to thousands of years), linked to varied sediment and water discharge from
the Himalayan mountains and the lowland alluvial plain.
Acknowledgements
This paper is a part of the Ph.D. thesis of NGR at Indian Institute of
Technology Kanpur, and he acknowledges the Institute for the research fellowship for his tenure of stay. A major support for drilling
came from a research project funded to RS from the Department of
Science and Technology, New Delhi, and their generous support is
thankfully acknowledged. NGR carried out the luminescence work
at Nordic Luminescence Laboratory at Risø, Denmark, and the help received from Andrew Murray and Mayank Jain for the dating work was
invaluable. The support for this visit and dating costs were met from a
Discovery Grant from the Natural Sciences and Engineering Research
Council of Canada to MRG. This is a contribution to IGCP 582 on Tropical Rivers.
References
Agarwal, K.K., Singh, I.B., Sharma, M., Sharma, S., 2002. Extensional tectonic activity in
the cratonward parts (peripheral bulge) of the Ganga Plain foreland basin, India.
International Journal of Earth Science 91, 897–905.
Aitken, M.J., 1985. Thermoluminescence Dating. Academic Press, London. 395 pp.
Ambrose, W.A., Tyler, N., Parsley, M.J., 1991. Facies heterogeneity, pay continuity, and
infill potential in barrier island, fluvial, and submarine-fan reservoirs: examples
from the Texas Gulf Coast and Midland Basin. The three-dimensional facies architecture of terrigenous clastic sediments and its implications for hydrocarbon
discovery and recovery: SEPM Concepts in Sedimentology and Paleontology, 3,
pp. 13–21.
Amorosi, A., Pavesi, M., Ricci Lucchi, M., Sarti, G., Piccin, A., 2008. Climatic signature of
cyclic fluvial architecture from the Quaternary of the central Po Plain, Italy. Sedimentary Geology 209, 58–68.
Auclair, M., Lamothe, M., Huot, S., 2003. Measurement of anomalous fading for feldspar
IRSL using SAR. Radiation Measurements 37, 487–492.
Bestland, E.A., 1997. Alluvial terraces and paleosols as indicators of early Oligocene climate change (John Day Formation, Oregon). Journal of Sedimentary Research 67,
840–855.
Bisson, P., Montgomery, M.R., 1996. Valley segments, stream reaches and channel
units. In: Hauer, F.R., Lamberti, G.A. (Eds.), Methods in Stream Ecology. Academic
Press, San Diego, pp. 23–52.
Blum, M.D., Aslan, A., 2006. Signatures of climate vs. sea-level change within incised
valley-fill successions: Quaternary examples from the Texas Gulf Coast. Sedimentary Geology 190, 177–211.
Blum, M.D., Törnqvist, T.E., 2000. Fluvial responses to climate and sea-level change: a
review and look forward. Sedimentology 47, 2–48.
Bogaart, P.W., Van Balen, R.T., Kasse, C., Vandenberghe, J., 2003. Process-based modelling of fluvial system response to rapid climate change — I. model formulation and
generic applications. Quaternary Science Reviews 22, 2077–2095.
Brown, A.G., Keough, M., 1992. Holocene floodplain metamorphosis in the Midlands,
United Kingdom. Geomorphology 4, 433–445.
Bryson, R.A., Swain, A.M., 1981. Holocene variations of monsoon rainfall in Rajasthan.
Quaternary Research 16, 135–145.
Bull, W.B., 1991. Geomorphic Responses to Climate Change. Oxford Univ, Press, New
York. 326 pp.
Catuneanu, O., 2006. Principles of Sequence Stratigraphy. Elsevier, Amsterdam. 375 pp.
Chauhan, M.S., Sharma, C., Singh, I.B., Sharma, S., 2004. Proxy records of late Holocene
vegetation and climate changes from Basaha Jheel, central Ganga plain. Journal of
the Paleontological Society of India 49, 27–34.
Clemens, S.C., Prell, W.L., 2003. A 350,000 year summer-monsoon multi-proxy stack
from the Owen Ridge, Northern Arabian Sea. Marine Geology 201, 35–51.
Das Gupta, S.P., 1975. The Upper Gangetic Flood Plain: A Regional Survey. National
Atlas Organization Monograph, Calcutta. 194 pp.
Friend, P.F., Sinha, R., 1993. Braiding and meandering parameters. In: Best, J.L., Bristow,
C.S. (Eds.), Braided Rivers: Geological Society of London Special Publication, 75,
pp. 105–111.
Gibling, M.R., 2006. Width and thickness of fluvial channel bodies and valley fills in the
geological record: a literature compilation and classification. Journal of Sedimentary Research 76, 731–770.
Gibling, M.R., Tandon, S.K., Sinha, R., Jain, M., 2005. Discontinuity-bounded alluvial sequences of the southern Gangetic plains, India: aggradation and degradation in response to monsoonal strength. Journal of Sedimentary Research 75, 369–385.
Gibling, M.R., Sinha, R., Roy, N.G., Tandon, S.K., Jain, M., 2008. Quaternary fluvial and eolian deposits on the Belan River, India: paleoclimatic setting of Paleolithic to Neolithic archeological sites over the past 85,000 years. Quaternary Science Reviews
27, 392–411.
Gibling, M.R., Fielding, C.R., Sinha, R., 2011. Alluvial valleys and alluvial sequences:
towards a geomorphic assessment. In: Davidson, S., Leleu, S., North, C.P.
(Eds.), From River to Rock Record: The Preservation of Fluvial Sediments
and their Subsequent Interpretation. Tulsa, Oklahoma: SEPM Special Publication, 97, pp. 423–447.
Glooschenko, W.A., Martini, I.P., 1987. Vegetation of river-influenced coastal marshes
of the southwestern end of James Bay, Ontario. Wetlands 7, 71–84.
Goodbred Jr., S.L., 2003. Response of the Ganges dispersal system to climate change: a
source-to-sink view since the last interstade. Sedimentary Geology 162, 83–104.
Gupta, A.K., Das, M., Anderson, D.M., 2005. Solar influence on the Indian summer monsoon during the Holocene. Geophysical Research Letters 32, 1–4.
Houben, P., 2003. Spatio-temporally variable response of fluvial systems to LatePleistocene climate change: a case study from central Germany. Quaternary Science Reviews 22, 2125–2140.
N.G. Roy et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 356–357 (2012) 38–53
Huntley, D.J., Lamothe, M., 2001. Ubiquity of anomalous fading in K-feldspars and the
measurement and correction for it in optical dating. Canadian Journal of Earth Sciences 38, 1093–1106.
Ivanocho, T.S., Goneshram, R.S., Brummer, G.A., Ganssen, G., Jung, S.J.A., Moreton, S.G.,
Kroon, D., 2005. Variation in tropical convection as an amplifier of global climate
change at the millennial scale. Earth and Planetary Science Letters 235, 302–314.
Jain, M., Bøtter-Jensen, L., Murray, A.S., Essery, R., 2007. A peak structure in isothermal
luminescence signals in quartz: origin and implications. Journal of Luminescence
127, 678–688.
Jones, L.S., Schumm, S.A., 1999. Causes of avulsion: an overview. In: Smith, N.D., Rogers,
J. (Eds.), Fluvial Sedimentology VI: Special Publication International Association of
Sedimentologists, 28, pp. 171–178.
Julien, P.Y., 1995. Erosion and Sedimentation, 2nd Edition. CUP, Cambridge. 392 pp.
Juyal, N., Chamyal, L.S., Bhandarib, S., Bhushana, R., Singhvi, A.K., 2006. Continental record of the southwest monsoon during the last 130 ka: evidence from the southern
margin of the Thar Desert, India. Quaternary Science Reviews 25, 2632–2650.
Lewin, J., 1992. Alluvial sedimentation style and archaeological sites: the lower
Vyrnwy, Wales. In: Needham, S., Macklin, M.G. (Eds.), Alluvial Archaeology in Britain. Oxbow, Oxford, pp. 103–109.
Machette, M.N., 1985. Calcic soils of the southwestern United States. In: Weide, D.L.
(Ed.), Soil and Quaternary Geology of the southwestern. United States, Geological
Society of America Special Paper 203, pp. 1–21.
Mack, G.H., Seager, W.R., Leeder, M.R., Perez-Arlucea, M., Salyards, S.L., 2006. Pliocene
and Quaternary history of the Rio Grande, the axial river of the southern Rio
Grande rift, New Mexico, USA. Earth-Science Reviews 79, 141–162.
Macklin, M.G., 1992. Holocene river environments in Prehistoric Britain, human interaction and impact. Quaternary Proceedings 7, 521–530.
Martin, J., Cantelli, A., Paola, C., Blum, M., Wolinsky, M., 2011. Quantitative modeling of the
evolution and geometry of incised valleys. Journal of Sedimentary Research 81, 64–79.
Miall, A.D., 1996. The Geology of Fluvial Deposits. Springer-Verlag, New York. 582 pp.
Milana, J.P., 1998. Sequence stratigraphy in alluvial settings: a flume-based model with
applications to outcrop and seismic data. American Association of Petroleum Geologist Bulletin 82, 531–545.
Mohindra, R., Parkash, B., Prasad, J., 1992. Historical geomorphology and pedology of
the Gandak megafan, Middle Gangetic plains, India. Earth Surface Processes and
Landforms 17, 643–662.
Montgomery, D.R., Buffington, J.M., 1997. Channel reach morphology in mountain
drainage basins. Bulletin Geological Society of America 109, 596–611.
Murray, A.S., Wintle, A.G., 2000. Luminescence dating of quartz using an improved single aliquot regenerative-dose protocol. Radiation Measurements 32, 57–73.
Overpeck, J., Anderson, D., Trumbore, S., Prell, W., 1996. The southwest Indian monsoon over the last 18000 years. Climate Dynamics 12, 213–225.
Owen, L.A., Finkel, R.C., Caffee, M.W., 2002. A note on the extent of glaciation throughout the Himalaya during the global Last Glacial Maximum. Quaternary Science Reviews 21, 147–157.
Owen, L.A., Caffee, M.W., Finkel, R.C., Seong, Y.B., 2008. Quaternary glaciation of the Himalayan–Tibetan orogen. Journal of Quaternary Science 23, 513–531.
Pandey, D.N., Gupta, A.K., Anderson, D.M., 2003. Rainwater harvesting as an adaptation
to climate change. Current Science 85, 46–59.
Paola, C., 2000. Quantitative models of sedimentary basin filling. Sedimentology 47,
121–178.
Porter, S.C., 2001. Chinese loess record of monsoon climate during the last glacial–interglacial cycle. Earth-Science Reviews 54, 115–128.
Porter, S.C., An, Z., Zheng, H., 1992. Cyclic Quaternary alluviation and terracing in a nonglaciated drainage basin on the north flank of the Qinling Shan, central China. Quaternary Research 38, 157–169.
Prasad, S., Enzel, Y., 2006. Holocene paleoclimates of India. Science 66, 442–453.
Pratt, B., Burbank, D.W., Heimsath, A., Ojha, T., 2002. Impulsive alluviation during early
Holocene strengthened monsoons, central Nepal Himalaya. Geology 30, 911–914.
Prell, W.L., Kutzbach, J.E., 1987. Monsoon variability over the past 150,000 years. Journal of Geophysical Research 92 (D7), 8411–8425.
Rao, V.P., Kessarkar, P.M., Patil, S.K., Ahmad, S.M., 2008. Rock magnetic and geochemical record in a sediment core from the eastern Arabian Sea: diagenetic and environmental implications during the late Quaternary. Palaeogeography,
Palaeoclimatology, Palaeoecology 270, 46–52.
Roy, N.G., Sinha, R., 2005. Alluvial geomorphology and confluences dynamics in the
Gangetic plains, Farrukhabad–Kannauj area, Uttar Pradesh, India. Current Science
88, 2000–2006.
Roy, N.G., Sinha, R., 2007. Understanding confluence dynamics in the alluvial Ganga–
Ramganga valley, India: an integrated approach using geomorphology and hydrology. Geomorphology 92, 182–197.
Sarkar, A., Ramesh, R., Somayajulu, B.L.K., Agnihotri, R., Jull, A.J.T., Burr, G.S., 2000. High
resolution Holocene monsoon record from the eastern Arabian Sea. Earth and Planetary Science Letters 177, 209–218.
Sharma, M.C., Owen, L.A., 1996. Quaternary glacial history of NW Garhwal, Central
Himalayas. Quaternary Science Reviews 15, 335–365.
53
Sharma, S., Joachimski, M., Sharma, M., Tobschall, H.J., Singh, I.B., Sharma, C., Chauhan,
M.S., Morgenroth, G., 2004. Lateglacial and Holocene environmental changes in
Ganga plain, Northern India. Quaternary Science Reviews 23, 145–159.
Singh, R.L., 1994. India: A Regional Geography. National Geographical Society of India,
Varanasi.
Singh, I.B., Rajagopalan, 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.
Sinha, R., Friend, P.F., 1994. River systems and their sediment flux, Indo-Gangetic
plains, northern Bihar, India. Sedimentology 41, 825–845.
Sinha, R., Jain, V., Prasad, B.G., Ghosh, S., 2005a. Geomorphic characterisation and diversity of the fluvial systems of the Gangetic plains. Gemorphology 70, 207–225.
Sinha, R., Tandon, S.K., Gibling, M.R., Bhattacharjee, P.S., Dasgupta, A.S., 2005b. Late
Quaternary geology and alluvial stratigraphy of the Ganga basin. Himalayan Geology 26, 223–240.
Sinha, R., Bhattacharjee, P., Sangode, S.J., Gibling, M.R., Tandon, S.K., Jain, M., Godfrey,
D., 2007a. Valley and interfluve sediments in the southern Ganga plains, India: exploring facies and magnetic signatures. Sedimentary Geology 201, 386–411.
Sinha, R., Kumar, R., Sinha, S., Tandon, S.K., Gibling, M.R., 2007b. Late Cenozoic fluvial
successions in northern and western India: an overview and synthesis. Quaternary
Science Reviews 26, 2801–2822.
Sinha, R., Kettanah, Y., Gibling, M.R., Tandon, S.K., Jain, M., Bhattacharjee, P.S., Dasgupta, A.S.,
Ghazanfari, P., 2009. Craton-derived alluvium as a major sediment source in the Himalayan Foreland Basin of India. Geological Society of America Bulletin 121, 1596–1610.
Sirocko, F., Sarnthein, M., Erienkeuserg, H., Lange, H., Arnold, M., Duplessy, J.C., 1993.
Century-scale events in monsoonal climate over the past 24,000 years. Nature
364, 322–324.
Srivastava, P., Singh, I.B., Sharma, M., Singhvi, A.K., 2003. Luminescence chronometry
and Late Quaternary geomorphic history of the Ganga Plain, India. Palaeogeography, Palaeoclimatology, Palaeoecology 197, 15–41.
Staubwasser, M., Sirocko, F., Grootes, P.M., Segl, M., 2003. Climate change at the 4.2 ka
BP termination of the Indus Valley civilization and Holocene south Asian monsoon
variability. Geophysical Research Letters 30, 1425.
Stokes, S., 1992. Optical dating of young (modern) sediments using quartz: results from a
selection of depositional environments. Quaternary Science Reviews 11, 153–159.
Stouthamer, E., Berendsen, H.J.A., 2007. Avulsion: the relative roles of autogenic and allogenic processes. Sedimentary Geology 198, 309–325.
Strong, N., Paola, C., 2008. Valleys that never were: time surfaces versus stratigraphic
surfaces. Journal of Sedimentary Research 78, 579–593.
Tandon, S.K., Gibling, M.R., Sinha, R., Singh, V., Ghazanfari, P., Dasgupta, A., Jain, M., Jain,
V., 2006. Alluvial valleys of the Gangetic Plains, India: causes and timing of incision.
In: Dalrymple, R.D., Leckie, D.A., Tillman, R. (Eds.), Incised Valleys in Time and
Space: SEPM Special Publication, 85, pp. 15–35.
Taylor, P.J., Mitchell, W.A., 2000. The quaternary glacial history of the Zanskar range,
north-west Indian Himalaya. Quaternary International 65–66, 81–99.
Thamban, M., Kawahata, H., Rao, V.P., 2007. Indian summer monsoon variability during
the Holocene as recorded in sediments of the Arabian Sea: timing and implications.
Journal of Oceanography 63, 1009–1020.
Thomas, J.V., Parkash, B., Mohindra, R., 2002. Lithofacies and palaeosol analysis of the
Middle and Upper Siwalik Groups (Plio–Pleistocene), Haripur–Kolar section,
Himachal Pradesh, India. Sedimentary Geology 150, 343–366.
Tyler, N., Ethridge, F.G., 1983. Depositional setting of the Salt Wash Member of the
Morrison Formation, southwest Colorado. Journal of Sedimentary Petrology 53,
67–82.
Van Campo, E., Gasse, F., 1993. Pollen and diatom-inferred climatic and hydrological
changes in Sumxi Co Basin (Western Tibet), since 13000 yr B.P. Quaternary Research 39, 300–313.
Waelbroeck, C., Labeyrie, L., Michel, E., Duplessy, J.C., McManus, J.F., Lambeck, K., Balbon, E., Labracherie, M., 2002. Sea-level and deep water temperature changes derived from benthic foraminifera isotopic records. Quaternary Science Reviews 21,
295–305.
Weber, M.E., Wiedicke-Hombach, M., Kudrass, H.R., Erlenkeuser, H., 2003. Bengal fan
sediment transport activity and response to climate forcing inferred from sediment physical properties. Sedimentary Geology 155, 361–381.
Weissmann, G.S., Mount, J.F., Fogg, G.E., 2002. Glacially-driven cycles in accumulation
space and sequence stratigraphy of a stream-dominated alluvial fan, San Joaquin
Valley, California, U.S.A. Journal of Sedimentary Research 72, 240–251.
Williams, M.A.J., Pal, J.N., Jaiswal, M., Singhvi, A.K., 2006. River response to Quaternary
climatic fluctuations: evidence from the Son and Belan valleys, north-central India.
Quaternary Science Reviews 25, 2619–2631.
Wintle, A.G., Murray, A.S., 2006. A review of quartz optically stimulated luminescence
characteristics and their relevance in single-aliquot regeneration dating protocols.
Radiation Measurements 41, 369–391.
Yadav, G.S., Dasgupta, A.S., Sinha, R., Lal, T., Srivastava, K.M., Singh, S.K., 2010. Shallow
sub-surface stratigraphy of interfluves inferred from vertical electric soundings in
Western Ganga plains, India. Quaternary International 227, 104–115.