Introduction

The Haryana plain in the Himalayan foreland basin is the drainage divide among the Indus plain in the west and the Ganga plain in the east (Fig. 1). This plain is limited by the fault-bounded rivers, i.e., Yamuna River flowing along the Yamuna fault (Kumar et al. 1996) in the east and the Ghaggar River following the Ghaggar fault (Manchanda and Hilwig 1981; Singhai et al. 1991) in the west, the Thar Desert from the southwestern part, and the Himalaya in the northeastern boundary. Except the northern (Himalayan piedmont zone) and southern (Aravalli piedmont zone) boundaries, the area is almost flat and filled with Quaternary sediments. The basement below this plain shows a broad gently sloping surface toward northeast (Narula et al. 2000). A climatic trend of decreasing rainfall and increasing temperature towards southwest is observed (Fig. 2). Most part of the study area falls under arid to semiarid climatic zone with an average rainfall of 455 mm/year.

Fig. 1
figure 1

Location map of the study area (shown by a yellow line), with major geomorphic provinces (base map source: www.glcf.umiacs.umd.edu)

Full size image
Fig. 2
figure 2

Rainfall and temperature diagram of the study area (after Central Ground Water Board, India, 2004)

Full size image

This region has been affected by neotectonism and climatic changes during the Holocene as observed in the adjoining Ganga plains (Sharma et al. 2004; North and Wariswick 2007). Compression from the southwest has developed several longitudinal (parallel to the Himalayan trend) faults which have segmented the region into five fault-bounded tectonic blocks. Tilting and sagging of these blocks have controlled the geomorphology, sedimentation process, and hydro-geomorphology of the region to a great extent. Large rivers such as the Yamuna, the Ghaggar, and the Sutlej and many small rivers too have changed their courses in response to neotectonism by which a network of paleochannels formed defining the paleohydro-geomorphology of the region. Though some workers suggest that the Yamuna River, presently a tributary of the Ganga River, was flowing through the Haryana plains (Oldham 1886; Oldham 1893; Raike 1968; Wilhelmy 1966; Kar and Ghose 1984), subsurface evidence and chronology of changes of its courses are lacking (Wilhelmy 1966). The “Saraswati River,” highly extolled in the Rigveda (the oldest religious text available to mankind), has lost its course in this plain (Kalyanraman 1999) and is a major interest to researchers of geology and archeology since the publication of a research paper (Oldham 1886) as on its banks flourished the Vedic culture, one of the oldest civilizations (6000–3000 b.c.) (Radhakrishna 1999). Though the reasons for the extinction of the mighty Saraswati River are highly debated, the reporting of the lineament controlled paleochannels at depth of 20–25 m below the surface along the suspected course of the Saraswati River may further intensify the research in recent future.

Haryana plains (Fig. 1) were investigated for mapping geomorphological and tectonic features and degree of soil developments using an integrated approach such as remote sensing (satellite imageries), field checking of these features and ground-penetrating radar (GPR) studies at a few selected places, and optically stimulated luminescence (OSL) dating techniques. Neotectonic aspects were studied by using digital elevation models (DEM), an association of terminal fans and changes in modern/paleo-drainage patterns to locate faults. GPR studies were carried out at selected locations to study the nature of inferred faults. Using GPR, an attempt was also made to infer the absence/presence of the paleochannels in the areas of the suspected paleo-course of the “Lost Saraswati River” along the lineament.

Major landforms recognized from the area are Himalayan Piedmont, fluvial plains (Old Yamuna Plains I–III, Old Sutlej Plains I–II, Old and Young Katha Plains), Eolian Plain, Terminal Fans, Aravalli Hills, and associated Pediments and Piedmont. Based on the degree of soil development and OSL ages, these soil-geomorphic units have been grouped into six members (QIMS-I to VI) (Quaternary Indus Morphostratigraphic Sequence) of a morphostratigraphic sequence: QIMS-VI 9.86–5.38 Ka, QIMS-V 5.38–4.45 Ka, QIMS-IV 4.45–3.60 Ka, QIMS-III 3.60–2.91 Ka, QIMS-II < 2.91–1.52 Ka, and QIMS-I < 1.52 Ka. Nine sub-parallel, NW-SE trending longitudinal faults, i.e., Ambala Faults-I and II, Markanda Fault, Dhund Fault, Karnal Fault, Patiala Fault, Jind Fault, Rohtak Fault, and Hissar Fault, were identified using drainage, paleo-drainage, and DEMs. These longitudinal faults are probable thrust splays off the Himalayan Frontal Thrust or independent entities developed due to compression from the southwest. Based on this study, major upheavals in the Proto-history of the NW India can be inferred to have been taking place at about 3000 b.c., probably caused by tectonic activity and climatic change to drier conditions. Second major upheaval occurred at about 2000 b.c. leading to the breakdown of the drainage system of the region caused by tectonic one.

Regional studies of geomorphic features and their relationship with the neotectonic activities of the region are limited (Singhai et al. 1991) and hence this study is an attempt in this direction.

Methodology

Mapping of geomorphological units and soil using remote sensing

Remote sensing and geographic information system (GIS) techniques have successfully been used in mapping soils in the adjoining part of the Ganga plain (Singh et al. 2006; Bhosle et al. 2008; Pati et al. 2011b). The following LANDSAT satellite multi-spectral scanning (MSS) data freely accessed from the web ( www.glcf.umiacs.umd.edu/ ) have been used in the present investigations:

Image MSS

Path/row

Date

(a) WRS-P/R-1

158/039

23 Feb 03

(b) WRS-P/R-1

158/040

31 Jan 03

A false color composite (FCC) was generated by coding bands 4, 3, and 2 as red, green, and blue for MSS images, respectively. LANDSAT data are already geo-referenced in UTM coordinates and were converted into polyconic and geographic projections, as required.

Using the image elements (Gupta 1991) such as tone, texture, pattern, size, shape, and association, a number of landforms like river floodplains, Piedmont, paleochannels, terminal fans, old river plains, and eolian plain were identified. These were checked in subsequent field works. A total of 25 soil-geomorphic units with soil properties varying within small ranges were mapped within the study area (Fig. 3a).

Fig. 3
figure 3

a Boundary of soil-geomorphic units superimposed on a mosaic of Landsat MSS FCC images (WRS-1, Path-158, Row 39 and WRS-1, Path 158, Row 40, 25 Jan and 23 Feb 2003). OdPt, Oldest Piedmont; OdPt-I, Old Piedmont-I; OdPt-II, Old Piedmont-II; Apt, Active Piedmont; Ygpt, Young Piedmont; YgSjPn, Young Sutlej Plain; OdSjPn, Old Sutlej Plain; OdYPn-I, Old Yamuna Plain-I; OdYPn-II, Old Yamuna Plain-II; OdYPn-III, Old Yamuna Plain-III; OdKaPn, Old Katha Plain; YgKaPn, Young Katha Plain; YnCgTFn-I, Young Chautang Terminal Fan-I; YnCgTFn-II, Young Chautang Terminal Fan-II; YnCgTFn-III, Young Chautang Terminal Fan-III; KrTFn, Karnal terminal Fan; StTFn, Sonipat Terminal Fan; ShFn, Sahibi Fan; MaTFn, Markanda Terminal Fan; OdYTFn, Old Yamuna Terminal Fan; FlPn, fluvial plain; Fi-Al-Pn, fluvio-eolian plain; Al-Pn, eolian plain; ArHPn, Aravalli Hill piedmont plain. b Study area is divided into six morphostratigraphic sequences, namely QIMS-I–VI based on OSL age and soil development. QIMS, Quaternary Indus Morphostratigraphic Sequence

Full size image

Digital elevation model

Digital elevation models (DEMs) were prepared by manually digitizing the point heights from the Survey of India topographic sheets on a scale of 1:50,000 which have a resolution of ± 2 m heights (Bhosle et al. 2009). The topographic sheets were geo-referenced with the same coordinates as those of the MSS images. A total of 6579 spot heights taken from 105 topographic sheets were digitized by using ArcGIS 9 software and a DEM was prepared by using linear rubber stretching interpolation approach, using ERDAS IMGINE8.5 software. Then the MSS image was draped upon the prepared DEM and a map was prepared for a regional overview of the area. Vertical exaggerations of 600 and 1200 were used to prepare 3D maps to identify possible faults and landforms and faults (Fig. 4), and lineaments (Fig. 5), respectively. 2D topographic profiles (cross-sections) generated from the DEMs were used to locate possible faults by a break in slope (Fig. 6) in such zones and to find out the differences in elevation across the inferred faults.

Fig. 4
figure 4

DEM with vertical exaggerations of 600 shows the fault-bounded tectonic blocks and major faults

Full size image
Fig. 5
figure 5

Study area shows two sets of lineaments (D-I and D-II). D-I is older than D-II

Full size image
Fig. 6
figure 6

Topographic profiles drawn across the Ambala Faults I and II (AF-I and AF-II), Patiala Fault (PtF), Markanda Fault (MaF), Dhand Fault (DhF), Karnal Fault (KrFt), Jind Fault (JdF), and Hissar Fault (HrF). Subfigure on the left shows locations of the topographic profiles on DEM

Full size image

DEMs of small areas around the terminal fans and possible faults were prepared using kriging interpolation method of SURFER-8 program (Golden Software Inc. 2003) to produce near-realistic 3D views (Fig. 7) with vertical exaggerations of about 200 to 400 and the best possible viewing positions and sun-setting (Table 1), which produced most pleasing results. DEMs show some artifact morpho-features like significant breaks in ‘slopes’ and ‘cliffs’, which are indicative of faults (Bhosle et al. 2009).

Fig. 7
figure 7

Digital elevation models prepared for terminal fans by the SURFER-8 program (Golden Software, Inc. 2003). MaTFn, Markanda Terminal Fan; ACgTFn, Active Chautang Terminal Fan; KrTFn, Karnal terminal Fan; StTFn, Sonipat Terminal Fan; ShFn, Sahibi Fan; KaTFn, Katha Terminal Fan

Full size image
Table 1 Terminal fans and associated faults in the area (vertical exaggerations and light positions used for preparations of DEMs for different terminal fans included; modified after Acharya et al. 2012)
Full size table

Field investigations

Detailed fieldwork was carried out for the study of the morphological properties, nature of sediments, and degree of soil development of various geomorphic units (Fig. 8) mapped by remote sensing. Field impression of tectonic features and their correlation with geomorphology was established during the field work. In addition, boundaries of the geomorphic units, soil character, and properties such as salt-efflorescence and water-logging conditions were cross-checked in the field. Possible locations for ground-penetrating radar (GPR) profiling were also marked during the field work.

Fig. 8
figure 8

Field photographs: a floodplain, b Himalayan Piedmont, c triangular facets marked by erosion common in tectonically active areas (Piedmont), d sand and gravel alternating layers in Piedmont, e recently deposited mass flow debris in front of Piedmont indicating seismic activity, f exposed section of the Sutlej megafan in the western part of the area

Full size image

Soils are indicators of tectonism in any area and hence have successfully been used to study neotectonic activities in the Ganga plains (Srivastava et al. 1994; Kumar et al. 1996; Singh et al. 2006; Bhosle et al. 2008; Pati et al. 2011b). Fifty-six pedons (Fig. 9) from the different geomorphic units were studied in detail in the field for their soil morphology and degree of soil development. Sediment samples from different geomorphic units were collected in one-end closed iron pipes for optical stimulated luminescence (OSL) dating. Samples from C-horizons of soil profiles were collected for OSL dating and suitable locations were selected for collecting samples from the base of each of the geomorphological units. For this purpose, brick kilns were targeted as open pits are available for the study of sediment properties and sample collection for OSL dating. Few pits were dug at a few places for collecting samples for OSL dating.

Fig. 9
figure 9

Distribution of OSL samples collected from the study area

Full size image

Optical stimulated luminescence dating of soils

Luminescence dating of sediments relies on the fact that a few minutes of sunlight exposure that the sediments certainly receive during weathering and transportation is sufficient to zero the pre-depositional luminescence signal. On burial, re-growth of signal occurs due to ambient radiation. The dose of radiations received per unit time by the decay of ambient radioactive elements and cosmic rays is the dose rate (R). Age determination (burial period) involves the determination of the total absorbed paleodose (P), by measuring signal (luminescence) emitted under infra-red stimulation. Fifty-three samples collected from all the 25 geomorphic units were dated by OSL technique. In the present study, 4–11 μm polymineralic fractions were used and luminescence was measured by stimulation with infrared rays using Corning-7-58 and BG-39 filters coupled to EMI-(9635QA PMT) and shine-down curves were prepared (Fig. 10). The sediment samples were given incremental irradiation (additive dose method) and the signal was measured each time. Paleodose is determined by extending backwards the curve between total dose given and measured luminescence (Fig. 10). A plot of equivalent dose versus shine downtime was prepared to check for the existence of “shine plateau” within 2–9 s of diode exposure (Fig. 10). The annual dose was determined by measuring the 238U, 232Th, and 40K. Average cosmic ray contribution was taken as 150 μGy/annum (at a depth of around 1 m) (Aitken 1985). Moisture percentage was measured by heating the sediment samples at 100 °C.

Fig. 10
figure 10

a OSL shine down curves, b growth curves, and c equivalent dose plateaus for some typical samples from the study area

Full size image

Thus, age can be determined by the following equation:

$$ \mathrm{Age}=\mathrm{Paleodose}\ \left(\mathrm{P}\right)/\mathrm{Annual}\ \mathrm{dose}\ \mathrm{rate}\ \left(\mathrm{R}\right) $$

Experimental results including the paleodose, dose rate, and age of the samples collected from different geomorphic units are given in Table 2.

Table 2 Radioactivity values, depth of IRSL samples, equivalent dose, and absolute ages of dated samples from different morphostratigraphic members (QIMS-VI–I; modified after Acharya et al. 2012)
Full size table

Major landforms in the study area

Using the MSS images, topographic maps, DEM, and detailed field work, major landforms like floodplains of different rivers, old fluvial plains, fluvio-eolian plains, terminal fans, piedmont zone, and pediment were identified (Fig. 3a, Table 3). These were further subdivided into old and young, depending on the OSL ages and degree of soil developments (Fig. 3b).

Table 3 Different landforms and their characteristic features identified in the Haryana plain with the help of digital elevation models (DEM) and field work
Full size table

Morphostratigraphy of the study area

Based on the OSL ages, the geomorphic units were grouped into six members of a morphostratigraphic sequence (Frye and Willman 1962) members QIMS-I to VI, with ages of ≤ 1.5 Ka, 1.5–2.9 Ka, 2.9–3.6 Ka, 3.6–4.4 Ka, 4.4–5.3 Ka, and 5.3–9.3 Ka, respectively (Fig. 11, Table 3). Though five members (QIMS-II–VI) have ages between 5.3 and 1.5 Ka in small ranges, the degree of soil development provides more degree of accuracy to the morphostratigraphy in reconstructing the evolutionary history of the region. Active floodplains of different rivers were not dated and were taken as < 1.5 Ka and have been included in QIMS-I.

Fig. 11
figure 11

Bar diagrams for 53 IRSL ages from different soil-geomorphic units with ages arranged in decreasing order. Tectonically stable periods are marked by flat or gently sloping portions of the curve joining tops of the bars. Sharp breaks in the curved are used to mark boundaries between different members of the morphostratigraphic sequence

Full size image

Variation in soil properties in the morphostratigraphic sequence

In general, the B-horizon thickness and that of solum, the degree of development of ped-structure increases with increasing age from QIMS-II to QIMS-VI.

Plots of the total clay content and pedogenic clay content with depth are shown in Fig. 12 and the degree of illuvial translocation has been assessed by calculating clay accumulation index (CAI) (Levine and Ciolkosz 1983). The CAI for QIMS-II to QIMS-VI varies from 227 to 484, 166–485, 186–714, 300–855, and 885–1147, respectively. This systematic variation suggests an increase of pedogenic clays with the increasing age. The younger soils have varied textures from sandy loam to silty loam classes. With increasing age, soils tend to change their textures to loam due to the breakdown of rock fragments and weathering of unstable minerals.

Fig. 12
figure 12

Variation of solum thickness with increasing age of soil. Distribution of clay content in soil profiles with age

Full size image

The member QIMS-II and III soils show very weak to moderate pedality, whereas the soils of the older members QIMS-IV to V show moderate to well-developed peds in B-horizons. In the upper and the lower horizons (A and C), soils show weak to no ped development. Argillans are thick to medium thick in the QIMS-V and VI soils. These are patchy, thin in QIMS-IV and III soils.

Lineaments

DEM with a vertical exaggeration of 1200 prepared by ERDAS IMAGINE 8.5 software brings out clearly two sets of lineaments D-I and D-II, striking in NE and ENE directions, respectively (Fig. 5). ‘D’ stands for the Drishadvati River, which was used to flow in this region. Based on the cross-cutting relations, the lineament D-II seems to be younger than D-I. Shifting of the Drishadvati River from the lineament D-I to D-II and subsequently away from the lineament D-II has been observed over a short period, as the oldest sediments associated with D-I and D-II are of similar OSL ages.

Structural features in the study area and the mechanism of development

The Ganga plain, since its origin, has experienced frequent geomorphic changes due to tectonic activities (Mohindra et al. 1992). Finite element modeling of the Ganga plain by Parkash et al. (2000) suggests compression from SW (Fig. 13) develops the longitudinal faults (parallel to subparallel with the Himalayan trend) and extensional transverse normal faults (an angle to the Himalayan trend). Using global positioning system (GPS), Sridevi (2004) suggested the northeastward movement of the Indian plate creates compression from the southwest. The above works support the mechanism of development of the longitudinal and transverse faults in the Ganga plain studied by Singh et al. (2006), Bhosle et al. (2008, 2009), and Pati et al. (2011a). In the present study area, eight longitudinal faults have been identified running parallel to subparallel to the Himalayan trend. The Ambala-I and II, Markanda, Patiala, Jind, Rohtak, and Hissar faults, which run almost in NW–SE direction, sub-parallel to the Himalayan trend are longitudinal in nature (Fig. 14), whereas the faults bounding the study area, i.e., Ghaggar and Yamuna Faults, transverse to the Himalayan trend, are tear faults in nature (Sahoo et al. 2000; Prabhat et al. 2014). The Ghaggar fault controls the course of the Ghaggar River, consists of a number of segments, and as a whole, it shows a curvilinear trend with convexity toward the southeast. The Yamuna faults which were earlier recognized (Kumar et al. 1996) follow a curvilinear pattern with convexity to the southwest and control the course of the Yamuna River (Fig. 15b).

Fig. 13
figure 13

Finite element model (Parkash et al. 2000) shows compression from SW develops longitudinal faults in the study area. GPS movement was taken from Sridevi (2004)

Full size image
Fig. 14
figure 14

Distribution of faults in the study area. Location of areas for which DEMs were prepared for terminal fans

Full size image
Fig. 15
figure 15

Major faults identified in the study area. a Standard FCC (ETM+ image, 7-4-2 band combination). The satellite image shows drainage anomalies across faults (shown in enlarged view in rectangles a-b-c-d-e-f) in the study area. b Distribution of epicenters of earthquakes in the study area (source—http://asc-india.org/seismi/seis-cdh.htm). AF-I and II, Ambala Faults-I and II; MaF, Markanda Fault; DhF, Dhund Fault; KrF, Karnal Fault; PtF, Patiala Fault; JdF, Jind Fault; RkF, Rohtak Fault; HrF, Hissar Fault

Full size image

Methodology used to identify faults

As the area is covered by Quaternary sediments, direct exposure of faults on the surface is not evident. Therefore, the methodology adopted by Goldsworthy and Jackson (2000), Singh et al. (2006), Bhosle et al. (2009), and Pati et al. (2011b) to map faults in the flat area was used here. Geomorphic markers such as offset and convergent drainage, drastic change in width or direction of a river over a short distance and formation of terminal fans were used to map possible faults. In addition, high vertically exaggerated DEMs (exaggeration 600), topographic profiles (Fig. 6) in the directions right angles to strike faults, distribution of earthquake epicenters, and GPR were used for the same. During this exercise, special attention was paid to boundaries of soil-geomorphic units, which may coincide along faults. Topographic breaks showing lateral continuity over tens of kilometers were taken to confirm the presence of a fault.

Drainage characteristics around faults

Due to aridity, natural surface drainage is very limited in the present area. Therefore, limited evidence are existing of the fault-influenced drainage.

Stream convergence occurs on the upthrown blocks of the normal and reverse faults due to upliftment and development of anticlines (Gawthorpe and Leeder 2000; Pati et al. 2011a). In the present study, convergent drainage is observed on the upthrown blocks of the Ambala, Karnal, and Markanda faults. Figure 15a shows a standard false color composite (FCC) prepared from Enhanced Thematic Mapper Plus (ETM+) image using 7-4-2 band combination. (The light and bright green colors represent grasslands and healthy vegetation, respectively. Pink areas represent barren soil; orange and brown represent sparsely vegetated areas. Dry vegetation appears as orange and water as blue. Urban areas appear in shades of magenta.) The image shows drainage anomalies observed in the study area across the faults. Smaller rectangles (a-b-c-d-e-f) represent an enlarged view of smaller areas within the study area where streams show convergence and offsetting in the satellite image.

Stream offset from the course around faults is a general feature in tectonically active basins (Gawthorpe and Leeder 2000; Singh et al. 2006; Bhosle et al. 2009). Smaller channels offset more as compared to larger rivers for the same relief across the concerned fault (Pati et al. 2011b). In the present study, offset of streams has been recorded around Amabala, Patiala, Markanda, and Hissar faults (Fig. 15a). Paleochannels show terminating and offsetting characteristics around Jind fault (Fig. 16).

Fig. 16
figure 16

Satellite imagery after a heavy rain showing paleochannels around Jind Fault filled with water and shows paleochannel morphology

Full size image

Due to lack of dense drainage network in the plain accompanied by artificial conversion of some of the paleochannels into canals, morphometric analysis is restricted to a few parameters. The parameters calculated from available drainage network have been summarized in Table 4. The obtained values imply that the plain is tectonically active. Major river channels, i.e., Ghaggar and Yamuna, offset along Markanda and Ambala-II faults by 9.2 and 4.5 km, respectively. Such an offsetting by a few kilometers distance along other faults is observed in other rivers also (Table 4). The channel offset indicates relief modification due to faults in the plain. Channel width and sinuosity of the Yamuna River have been calculated at 45 points along the river course. It is observed that average width of the channel decreases downstream (0.82 km in zone A, 0.49 km in zone B, and 0.48 km in zone C; Fig. 3). Sinuosity of the Yamuna River increases downstream which can be attributed to southwards decrease of slope. Sinuosity of the river apparently changes across the Patiala and Karnal fault. Up to Patiala fault, the sinuosity is 1.13; between Patiala fault and Karnal fault, it is 1.19; and from Karnal fault to Hissar fault, it increases to 1.27. Other streams across the faults show higher sinuosity and confluence on northern block as compared to its southern counterpart. Previous studies (Jackson and Leeder 1994; Gawthorpe and Leeder 2000; Singh et al. 2006; Bhosle et al. 2009; Pati et al. 2011a; Verma et al. 2017) show that rivers generally have higher sinuosity and exhibit convergence on the upthrown block, thereby implying northern blocks to be upthrown. As discussed in section “Structure of the area and nature of faults”, the subsurface nature of these faults is curved thrust splay. These morphometric parameters are thus another evidence suggesting the same. However, it should be kept in mind that structural and tectonic constraints are not always the only factors governing the drainage morphology; local topography, and soil and anthropogenic activities also play a significant role in creating anomalies.

Table 4 Channel offsets, sinuosity, and number of confluence points of different streams across the faults identified in the study area
Full size table

Based on the aforementioned criteria, nine faults (Ambala-I and II, Patiala, Markanda, Dhandh, Karnal, Jind, Rohtak, and Hissar Faults), six terminal fans (Fig. 14, Table 5) and five tectonic blocks, i.e., Piedmont, Jind-Rohtak, Aravalli, Punjab, and Saharanpur blocks (Fig. 4), were identified. Detail characteristics of the faults are given in Table 5.

Table 5 Major structural features delineated and their characteristic features based on surface drainage anomalies in the Haryana plain
Full size table

Distribution of earthquakes

Major parts of Haryana state fall under seismic Zone-IV (http://asc-india.org/seismi/seis-cdh.htm). Most of the earthquakes in this region originate at shallow depth, though a few earthquakes recorded from this region originated from intermediate depth. Earthquake epicenters from the study area were plotted on the map showing faults (Fig. 15b). It was observed that the region between the Rohtak and Hissar faults is the most active. The Rohtak Fault seems to be most active followed by Hissar fault. Also, the activity of the Jind fault and Ambala fault-II caused a few earthquakes. All these recorded earthquakes are of low to moderate magnitude limited around 5 Mb indicating frequent strain release around this area. Though focal mechanism solutions of these earthquakes are not available, the geological setting of the basement and the adjacent Himalaya suggests these earthquakes are originating along the flexural bulge of the basement. Recurrence of earthquakes along the basement fault suggests seismically active nature of the buried Precambrian basement ridge. The associated tectono-geomorphological features and the soil chronosequence suggest it to be seismically active through the Holocene. Very recently, a research paper has been communicated (Patel et al., communicated in International Journal of Earth Sciences) correlating the focal depth and frequency of occurrence of earthquakes of the region with the corresponding segment of the Himalayan front.

GPR studies

Numerous studies have been carried out during last 20–25 years by different agencies to trace the courses of the palaeo river Saraswati in the western Rajasthan and adjoining states of Haryana, Punjab, and Gujarat (Ghose et al. 1979; Yashpal et al. 1980; Ramasamy et al. 1991; Kar 1995; Sahai 1999; Rao 1999; Raghav 1999; Valdiya 2002; Gupta et al. 2004; Bhadra et al. 2009; Saini and Mujtaba 2010; Shukla and Thakkar 2014; Khonde et al. 2017). From the above studies, the possible paths of the “Lost Saraswati” have been identified. In the present study, GPR technique has been used to map the subsurface nature of one such path and found to be corroborating to the existing literature (Figs. 17 and 18).

Fig. 17
figure 17

GPR profile across the fluvial plain (two paleochannels are indicated by blue lines, faults areas are encircled in red line). Faults along the channel indicate the channel to be fault controlled

Full size image
Fig. 18
figure 18

GPR profile across Old Yamuna Plain-II shows buried channel below 20 m depth indicating the presence of the river. Presence of faults in the profile along the channel indicates the channels were fault controlled

Full size image

GPR profiling was carried out across the lineament controlled paleochannels (Fig. 17) suspected to be of the “Lost Saraswati.” The previous studies identified lost Saraswati River as a major Himalayan-origin river with a wide channel in the plain. The river must have carried a large amount of coarse to medium sand which it deposited along the active stream and silt along the sides, i.e., rest of the main channel (Fig. 19). The GPR possibly picked up one of those smaller paleochannels (~ 300 m wide in the present case) within the main channel (which is estimated to be a few kilometers wide). Such river systems typically show multi-story valley type architecture as shown by the Ganga valley (Gibling et al. 2011). Also, GPR profiles were carried out across six out of nine longitudinal faults inferred from remote sensing and GIS studies (Ambala-I and II, Patiala, Markanda, Rohtak, and Hissar Faults) to know their subsurface nature and continuity. These faults are not represented by single fault plane but are sets of a number of faults, and the majority of faults show downthrown side in direction inferred from the DEMs and topographic profiles. Two paleochannels have been identified in GPR profiles, i.e., on the fluvial plain (Fig. 17) and on the Old Yamuna Plain-II (Fig. 18).

Fig. 19
figure 19

Schematic cross section of delineated paleochannel in the study area

Full size image

Subsurface Interface Digital GPR system (SIR-3000), manufactured by Geophysical Survey Systems Inc. (GSSI), was used in this investigation. Shielded antennae were used to avoid ringing inherent in the hardware (Audru et al. 2001). Profiling was carried out in distance mode using a survey wheel (model 620; GSSI). In addition, 100 MHz antennas were used in the investigation which provides 20–25 cm resolution which is a good tradeoff between depth of penetration and resolution (Bristow and Jol 2003). Good reflections were obtained up to depths of 15 m. Surveys with the perpendicular broad-side position of the antennas (Jol and Bristow 2003) were used. Also, common offset method, which provides good results for sedimentological studies in the subsurface (Neal 2004), was used. Collected GPR data were processed using GSSI computer-based RADAN-5 software. Filter bandwidth 40/50 to 150/180 MHz were used as suggested by Fisher et al. (2000) for 100 MHz antenna and other processing techniques were used to a bare minimum, unless the data warranted.

Integration and interpretation of results

Structure of the area and nature of faults

All the nine faults in Haryana plains inferred from remote sensing and GPR studies are parallel to the Himalayan trend and are thus longitudinal in nature (Parkash et al. 2000).

Major faults in the western Ganga Plains (Yamuna fault, Deoha-Ganga Fault, Ghaggar Fault) trend N–S or NNE–SSW in the northern region, turn to N–S in the central region, and take ESE direction in the south, thus giving these faults a curvilinear shape. Finite element modeling of stress conditions in the region (Parkash et al. 2000), with compression from SW, inferred from tilting of large blocks and confining of basin in the west by the N–S trending subsurface Aravalli Ridge and southern Peninsula in the south with E–W trending contact, produced stress lines very similar to the major faults, which were called as longitudinal faults (Kumar et al. 1996). More recent finite element modeling for the Ganga plain and Haryana plains produces a stress pattern with major deformation contours in Haryana running parallel to the Himalayan trend (Parkash et al. 2000). This is due to the fact that Haryana basin is much narrower than Upper Ganga Plains. Also, Himalayas striking WNW–ESE by the side of the Ganga Plains take a turn along NW–SE to the west of the Yamuna River. Thus, SW compression almost impinges perpendicular to the foreland basin/Himalayas, forming Himalayan-trend parallel longitudinal faults in the Haryana plains.

Major faults of the Haryana Plains are longitudinal in nature and formed by compression from a direction perpendicular to them. The most probable explanation is that they (except for Hissar fault) are thrust splays off the Himalayan Frontal Thrust or independent thrust faults developed due to compression. Their near surface normal fault nature in the shallow depth as observed in GPR profiles is probably due to the fact these thrusts are curved in nature and these seem nearly normal faults near the surface (Fig. 20). Similar E–W trending shallow-depth longitudinal reverse faults was well reported by Pati et al. (2011a) in the Middle Ganga plain, from the western part of the Munghyr–Saharsa ridge (Eastern Ganga plain). The present area falls in the similar geological setting and hence these faults seem to be reverse faults. Southernmost Hissar Fault is very similar to Southern Boundary Fault (Parkash and Kumar 1991) between the southern Peninsula and the Ganga Plains. Irrespective of their nature, these faults have strong control over the fluvial geomorphic system of present and past.

Fig. 20
figure 20

Diagrammatic sketch showing different longitudinal faults in the Haryana plain as thrust splays off the Himalayan Frontal Thrust (HFT)

Full size image

Morphological features like various drainage patterns (convergent and offset drainages), initiation of new streams, development of terminal fans, breaks in slopes, and the presence of breaks in slope in DEMs have been used for recognition of normal faults in the Upper Ganga plain (Singh et al. 2006; Bhosle et al. 2009). The same features have been useful in deciphering possible thrusts in the present case, as near-surface ramp anticlines forming topographic highs develop in case of thrusts similar to the development of anticlines on the upthrown sides of normal faults (Gawthorpe and Leeder 2000). However, the exact nature of the faults at depth can be confirmed only by subsurface techniques like seismic surveys.

Tectonics, sedimentation, and terminal fans

A great significance has been attached to the presence of distributary streams in formation of terminal fans (North and Wariswick 2007). Except for the one-third proximal part of the Young Chautang Terminal Fan-I, all other terminal fans in this area with semi-arid climate were deposited by distributary streams. Similar is the case in for the adjoining Ganga–Yamuna interfluve with semiarid climate, lying further east. However, further east, the sub-humid Deoha/Ganga–Ghaghara interfluve (Singh et al. 2006) and Ghaghara–Kosi regions are marked by terminal fans deposited by braided/meandering streams (Pati et al. 2011b, 2012). Thus, terminal fans formed by distributary stream systems are confined to semiarid climate, with a few exceptions.

All the terminal fans in the major parts of the Upper Ganga Plains were found to have been formed by the involvement of the whole of the inland streams on the downthrown blocks of normal faults (Singh et al. 2006; Bhosle et al. 2008). However, more recent studies (Pati et al. 2011b) in the Middle Ganga Plains found that splays from a large river like the Gandak could also generate “splay terminal fans” (Gandak Splay Terminal Fans) on the underthrust blocks of some thrusts. In fact, the classical Markanda Terminal Fan is indeed a splay terminal fan, as the Markanda River flowing in a southerly direction was constrained to flow westwards by the E–W trending Markanda fault. The Markanda Terminal fan was developed on the downthrown block to the south of the Markanda fault by splays from the Markanda River. Similarly, the Old Yamuna Terminal Fan was developed on the downthrown block south of the Karnal fault by splays from the Yamuna River, in the process of withdrawing from this area. Similarly, splays from the Yamuna River, when it was flowing along the modern Chautang River course (see Section 9.4), could have created Karnal (splay) Terminal Fan across the Karnal fault and Sonipat (splay) Terminal Fan might have been developed by splays from the Yamuna River from its present position. The Young Chautang Terminal Fans I–III were formed by the involvement of the whole of the Chautang River.

Role of climate in geomorphic changes

Some broad ideas are available about climatic changes during the Holocene in the present area, the adjoining Thar Desert to the southwest, Ganga plains to the east, and oceans. Studies of cores from the Bay of Bengal (Goodbread and Kuehl 2000; Goodbread 2003) and lacustrine cores from the Upper Ganga Plain (Sharma et al. 2004) indicate that the earlier half of the Holocene was marked by very wet and warm climate and in the later half the climate turned wet and warm. Also, studies of lacustrine deposits from the adjoining Thar Desert show that the periods > 10,000, 10,000–5000, and 5000–3000 years b.p. were marked by cold and severe dry climate, slightly wetter climate, and wet climate, respectively. Later drier conditions came in and continued with minor modifications till present (Singh 1971; Singh et al. 1972, 1974, 1990).

The present study indicates that since ~ 4.1 Ka, terminal fans related to faults were formed by streams with distributary pattern. As discussed in Section 9.4, period 4.1–3.8 Ka was a time of tectonic activity, which also caused shifting away of the Yamuna from the Haryana plain due to upliftment and tilting, forming upland, followed by the development of many terminal fans due to the activity of longitudinal faults. Thus, the role of any climatic changes is ruled out in this major change in drainage pattern of the area.

The lost Saraswati and Drishadvati rivers, and evolution of drainage in the Haryana plain

As obvious from soil-geomorphic map (Fig. 3), GPR studies (Figs. 17 and 18), and the hydro-geomorphology of the region from the Vedic time to present (Fig. 21), the Yamuna had flowed through Haryana Plains, its deposits were covered by later terminal fans or eolian deposits, so obvious surficial evidence of the presence of this large river are not expected. In Jind–Rohtak Block, archeological sites of pre-Harappan to post-Harappan period follow straight lines trending NE–SW (Fig. 22), following the Old Yamuna courses, as these cultures were essentially riverine in nature. It is very likely that the rivers were flowing in this direction during these periods. An indication of flow of the large Yamuna River in the same NE–SW direction inferred from lineament map confirms this aspect. But major confirmation of the Yamuna River flowing through Haryana Plains comes from GPR studies of the subsurface paleochannels, suggesting the presence of a large river in the area in the recent past.

Fig. 21
figure 21

Distribution of drainage in Haryana and adjoining areas in Vedic time and at present (after Joshi et al. 1984)

Full size image
Fig. 22
figure 22

The Proto-historic sites lie in straight lines (marked in red), which on northward extension meet the modern Yamuna River, suggesting that the Proto-Yamuna River flowed through these regions (modified after Valdiya 2002)

Full size image

Integration of our work with available archeological information can be used to reconstruct changes in the course of the Yamuna River in the Haryana Plains (Fig. 22). As suggested earlier by Wilhelmy (1966), the Yamuna flowed along the modern Chautang River, forming the Drishadvati River. It was entering the Haryana Plains near Karnal, as indicated by lineaments. As Pre-Harappan (Vedic Civilization) sites are concentrated along the Jind–Hissar–Rohtak triangle on the Drishadvati course and age of the Vedic Civilizations is 3000–6000 b.c. (Radhakrishna1999), it is likely that the Chautang River course is the oldest course of the Yamuna in the Haryana Plain. Splays form the Yamuna/Drishadvati has created the Karnal Terminal fan.

It can be seen from the soil-geomorphic map (Fig. 3), from the Rohtak–Jind–Hissar triangle of the Pre-Harappan culture, towards northwest a younging sequence fluvial plain (4.5 Ka) and Old Yamuna Plain-II (4.1 Ka) and III (3.6 Ka), suggesting that the Yamuna was shifted northwestward through these plains in three stages. In the second stage, it was following the lineament D-II in the fluvial plain (Fig. 5), when the River Yamuna entered the Yamuna Plains from a still more northerly position. In next stage, the Yamuna occupied Old Yamuna Plain-II. In the last stage (< 4.1 Ka), the major part of the River Yamuna was shifted to the Old Yamuna Plain-I and a small part flowed through Old Yamuna Plains-III, till about 3.0 Ka to form the Old Yamuna Terminal fan. This occupation of the Old Yamuna Plains-III by the remnant Yamuna River was due to initiation of the activity of longitudinal faults with downthrows to the southwest. The Sutlej River started shifting by 3.9 Ka and soon became a tributary of the Indus River, though a minor part of it flowed through the Old Sutlej Plain-II till about 2.75 Ka.

The cause of the systematic shifting of the Yamuna toward northwest is probably tilting of the Haryana Plains toward the northwest. Also, the Sutlej River from which was a tributary of the Ghaggar (Saraswati) River (Fig. 21), started shifting northwestward almost at the same time (~ 1900 b.c.) due to tilting of the Panjab block toward NNW (Singhai et al. 1991) and soon became an independent entity. Thus, major changes in drainage pattern in the Haryana and Panjab states led to complete disruption of the supply of a perennial flow of the two major rivers (Yamuna and Sutlej) into the Ghaggar River (Saraswati River). This caused northward and eastward movement of the Late Indus Valley (Harappan) population. The cause of shifting of rivers was a tectonic one, involving tilt and uplift of Haryana block and tilt of the Panjab block. The recent studies on tracing the Vedic Saraswati river have integrated different proxies such as OSL dating, isotopes, remote sensing, sedimentological and archeological data, field studies, and hydrogeomorphic, neotectonic, and climate studies to ascertain the course of the Vedic Saraswati. Bhadra et al. (2006) analyzed subsurface oozing water from ponds in Jind district of Haryana and concluded that Ghaggar river flows on the paleochannel of Vedic Saraswati in western Haryana. The studies have reported paleochannels of Vedic Saraswati in the same region (western Haryana, Hissar, and Jind districts) where the paleochannel is identified in the present study with the help of GPR. Isotope study by Khonde et al. (2017) showed that Vedic Saraswati discharged into Arabian Sea until 10 Ka. Valdiya (2013) assimilated a number of geophysical, geological, sedimentological, archeologic, geomorphological, tectonic, and geochronological evidence to trace the Vedic Saraswati and attributed disappearance of the river to tectonic and climate changes. Gupta et al. (2004, 2008) also shows many abandoned channels of the river in upper middle reach of Saraswati. Saini and Mujtaba (2010) show geomorphology and lithological architecture of the section across the paleochannel of the river from western Haryana.

After shifting away of the main Yamuna River from the Haryana Plain, the activity of the Markanda Fault produced the Markanda Terminal Fan at about 3.4 Ka and activity of the Markanda, Jind, Karnal, and Rohtak faults at different times produced the Young Chautang terminal Fans I–IV (Table 5) on their downthrown blocks.

In the second phase, different longitudinal faults were developed and some more terminal fans (Markanda Terminal Fan and Chautang Terminal Fans I–III) were formed due to their activity. Similar tilting and later development of faults to form terminal fans have been reported in the Gandak megafan (Pati et al. 2011b). Thus, in response to compression, first, large blocks in flat plains try to accommodate this stress by tilting. Limited numbers of terminal fans (Old Yamuna and Karnal Terminal Fans) do develop due to the activity of some longitudinal faults at this stage. If the compression continues further, then these blocks respond by the development of faults accompanied by the formation of terminal fans.

The third phase of Yamuna River (Y3 course in Fig. 23) coincides with the course of Vedic Saraswati as envisaged earlier (Valdiya 2002) except in the foothill region. These changes in directions of the present Yamuna River in the Haryana plains as worked out above seem to be a good explanation for the presence of pilgrim centers along the modern Saraswati stream, high discharge of this river in the past and even establishment of the Adi Badri (place of origin of the Saraswati River) pilgrim centre.

Fig. 23
figure 23

Evolution of Yamuna and Sutlej rivers in Haryana plains and adjoining areas during the period 3000 b.c. and 2100 b.c. “Y” stands for the River Yamuna and “S” for Sutlej

Full size image

Conclusions

  1. 1.

    Major landforms recognized from the Haryana and adjoining areas are Himalayan Piedmont, fluvial plains (Old Yamuna Plains I–III, Old Sutlej Plains I–II, Old and Young Katha Plains), Eolian Plain, Terminal Fans, Aravalli Hills, and associated Pediments and Piedmont.

  2. 2.

    Based on the degree of soil development and OSL ages, individual landforms were further subdivided into soil-geomorphic units, e.g., Piedmont into Oldest Piedmont, Old Piedmont-I and II, and Young Piedmont. All 25 soil-geomorphic units were identified. Based on the degree of soil development and OSL ages, these soil-geomorphic units were grouped into six members (QIMS-I to VI) (Quaternary Indus Morphostratigraphic Sequence) of a morphostratigraphic sequence: QIMS-VI 9.86–5.38 Ka, QIMS-V 5.38–4.45 Ka, QIMS-IV 4.45–3.60 Ka, QIMS-III 3.60–2.91 Ka, QIMS-II < 2.91–1.52 Ka, and QIMS-I < 1.52 Ka.

  3. 3.

    There is in general an increase in the degree of soil development from member QIMS-I to QIMS-VI soils.

  4. 4.

    Nine sub-parallel, NW–SE trending longitudinal faults, i.e., Ambala Faults-I and II, Markanda Fault, Dhund Fault, Karnal Fault, Patiala Fault, Jind Fault, Rohtak Fault, and Hissar Fault, were identified using drainage, paleo-drainage, and DEMs. In DEMs, these faults (except the Hissar Fault) show apparent downthrown sides to the south. These are probable thrust splays off the Himalayan Frontal Thrust or independent to it developed due to compression. The Hissar Fault, which is a normal fault, exhibits downthrown side toward the north.

  5. 5.

    A DEM with a high vertical exaggeration of 1200, prepared by ERDAS software, indicates NE–SW trending lineaments, which are interpreted to indicate the direction of past flow of the Yamuna through these plains.

  6. 6.

    Two types of terminal fans have been identified from the area: terminal fan formed by the involvement of the whole discharge of the stream and splay terminal fans. These are formed across a fault on downthrown, low-lying block by splays from the river. Also, the classic Markanda terminal fan is a splay terminal fan.

  7. 7.

    Terminal fans formed by streams with distributary pattern are characteristic of regions with a semi-arid climate and are absent in regions with wetter climate.

  8. 8.

    Using GPR studies, all faults except the Jind Fault have been confirmed and found to consist of a set of a number of faults. Presence of paleochannels has been identified along the suspected course of the Lost Saraswati River, which will intensify the present research in future.

  9. 9.

    Major upheavals in the Proto-history of NW India that took place at about 3000 b.c. were probably caused by tectonic activity and climatic change to drier conditions. Our work indicates that the second major upheaval at about 2000 b.c. leading to the breakdown of the drainage system of the region was a tectonic one.