Introduction

Back-analysis and reconstruction of factors controlling the time-evolution of ongoing large-scale slope deformations is a challenge for engineering geologists and geomorphologists in view of understanding the role of predisposing (time-independent), preparatory (time-dependent) for the onset and triggering (transitory) factors for the acceleration of the Mass Rock Creep process (MRC, Chigira et al. 1992). The MRC can involve entire hillslopes or valley flanks acting on a large space–time scale, in response to a gravitational disequilibrium. The MRC process develops in a continuous and/or impulsive way following three stages of strain rate: primary (or decelerating), secondary (or stationary), and tertiary (or accelerating) (Pánek and Klimeš, 2016; Petley and Allison 1997; Saito 1969).

In addition to the most studied predisposing factors, such as geo-structural and lithological ones (bedding attitude, deformability contrasts, jointing conditions, geological structure—Kellogg et al. 2001; Agliardi et al. 2001; Bianchi Fasani et al. 2014; Alfaro et al. 2019; Petroccia et al. 2020; Vick et al. 2020; Esposito et al. 2021; Di Luzio et al. 2022; Delchiaro et al. 2021a; Discenza et al. 2023) that discriminates the potential occurrence of a Deep-seated Gravitational Slope Deformation (DGSD) and its failure mode, dynamic processes (e.g., groundwater infiltration, river incision) can prepare to slope instability over time, thus accelerating the slope deformations towards a tertiary creep stage and eventually to a generalised failure (Cornelius and Scott 1993; Chigira and Kiho 1994; Marmoni et al. 2023). The tectonic uplift and the consequent river incision can be considered among the most relevant preparatory factors (Tsou et al. 2015; Tu and Deng 2021), whose rates are able to change the elevation above the thalweg causing stress release and providing the kinetic arrangement for MRC process acceleration (Hungr et al. 2014; Della Seta et al. 2017; Delchiaro et al. 2019, 2021b, 2023). Indeed, the interplay among tectonics, river incision, and topography gives rise to topographic stress, a pivotal factor in shaping the size of bedrock landslides at a landscape scale. This stress dynamically impacts the strength of subsurface materials through mechanisms like fracturing and weathering (Li and Moon 2021).

However, the availability of field evidence for constraining morpho-evolutionary models and a multi-stage trend of slope-scale deformations driven by viscous processes, from their onset to the ultimate acceleration, are rare (Bigot-Cromier et al. 2005; El Bedoui et al. 2009; Glueer et al. 2019, 2020; Rabus et al. 2022) or limited to the ultimate conditions. Consequently, understanding the evolution of DGSDs towards failures is anything but close to being understood. Their understanding faces off with the preservation of gravity-induced landforms, which tend to be masked by earth-surface processes over time.

Other critical features to be constrained are the volume of the deforming mass (Hungr et al. 2001; Esposito et al. 2007) and the persistence time (Li et al. 2022). If a highly favourable kinetic arrangement exists and recurrent forcings hit the prone mass, morphoevolution resulting in several small-size failures prevents the enucleation of deeper processes driven by viscous flow, thus avoiding larger volumes reaching critical values (Troiani et al. 2020).

Recurrent seismicity or strong motions occurring in the near or far field (Lenti et al. 2015) can interact with the deforming slopes or landslide masses, favouring transient acceleration or step-like pulses (Gutiérrez et al. 2008, 2023; Amato et al. 2018; Moro et al. 2007, 2012). Therefore, earthquakes cannot be neglected as a possible cause of acceleration of ongoing viscosity-driven deformations (Pánek and Klimeš 2016).

This research aims to shed light on the role of long-term preparatory factors in MRC deforming slopes in the Lorestan arc of the tectonically active Zagros belt (Iran). In detail, the role of active tectonics and its feedback in surface processes are deeply focused, adding pieces of evidence on the frame of the morpho-structural and tectonic evolution of active belts. A spectacular example of an ongoing DGSD involving a stiff calcareous and marly succession in an exemplary context is presented. The area presently falls where active seismicity recurs and is characterised by a semi-arid (steppe) climate (BSk) according to the Köppen (1900) classification, where the total annual rainfall is 354.1 mm, and the mean temperature is 22.2° (Zepner et al. 2021).

The learning case of the present study is the Siah Kuh DGSD, located in correspondence with a fold periclinal tip, along a slope cut by the course of the Dowairij River, 30 km south of the well-known Seymareh giant landslide (Fig. 1; Delchiaro et al. 2019, 2020, 2022; Rouhi et al. 2022). Although the case study represents an excellent example of coupling between tectonics and landscape evolution, originating the preparatory conditions for an MRC-driven process development, the Siah Kuh context has not yet been described in detail.

This paper reports the results of multidisciplinary research aimed at assessing the long-term to present-day slope evolution processes, including (i) field survey, (ii) satellite interferometry, (iii) geomorphometric analysis, and (iv) Optically Stimulated Luminescence (OSL) absolute dating of geomorphic markers. One aspect of such an approach is to reconstruct the long-term incision rate of the drainage network, that in a tectonically active area is directly caused by the tectonic uplift. Specifically, the evolution of the Dowairij River drainage system is examined as the river incision induces stress release at the slope base, likely starting deformational processes. Additionally, this approach aims at assessing the present-day surface displacement in order to detect active landsliding areas in the frame of the entire deforming slope, thus enabling the identification of various deformation styles across the slope.

Regional setting of the Siah Kuh anticline

The study area is located along the periclinal tip of the Siah Kuh fold in the simply folded belt of Zagros Mountains (Fig. 1). These NW–SE-trending mountains are part of the Alpine–Himalayan collisional belt which formed due to the convergence of the Arabian and the Eurasian plates in the Late Cretaceous– Cenozoic (Talbot and Alavi 1996; Stampfli and Borel 2002). Most of the folds within the Zagros fold-thrust belt, especially in the Lorestan zone, are interpreted as detachment folds (e.g., Alavi 1980, 1994; Mercier et al. 1997). The 10- to 12-km thick stratigraphic succession of the Lorestan arc consists of Palaeozoic–Cenozoic sediments. Based on O’Brien (1950), this succession is divided into five mechanical units, which include the basement group, the lower mobile group, the competent group, the upper mobile group, and the passive group. In the Lorestan arc, the competent units that control the fold style are the Sarvak and Ilam Formations (Bangestan Group) and the Asmari Formation. Furthermore, these formations are the topography-forming units in the region (e.g., Casini et al. 2011, 2018).

Fig. 1
figure 1

Geological map of the study area. Geological data, including bedding, are extracted from the NIOC at a scale of 1:100,000 (Llewellyn 1973; Setudehnia and Perry 1967) and integrated by field surveys. MFF: mountain front fault. The coordinate reference system is WGS84. The stress indicators are reported by the world stress map (Heidbach et al. 2018). The climate chart for the locality of Chaşfah (401 m a.s.l.), close to the Siah landslide system, for the period 1990–2019 is also shown in comparison with the average monthly rainfall and temperature for the period October 2014–March 2019 (Zepner et al. 2021)

Full size image

The Zagros active fold-thrust belt is one of the largest areas on the Earth where generally continuous seismicity associated with blind thrust faults at depth is not expressed by coseismic surface rupture in the whole belt (Berberian 1995). However, some large earthquakes can be responsible for up-to-ground faulting due to the presence of several ductile layers (e.g., Berberian 1995; McQuarrie 2004; Hatzfeld et al. 2010; Karasözen et al. 2019; Yaghoubi et al. 2021; Guo et al. 2022). Coseismic uplift and folding of Quaternary deposits and surfaces have been produced due to seismic slip along the basement faults (e.g., Yeats 1986; Berberian 1995; Nissen et al. 2007; Basilici et al. 2020).

Earthquakes comparable in terms of magnitude, depth, and fault mechanism, combined with preserved sedimentary cover rocks, have led to a deformation model encompassing basement shortening in conjunction with—but decoupled from— the folding of sedimentary cover rocks (e.g. Stocklin 1968; Falcon 1969; Jackson 1980; Jackson and Fitch 1981; Ni and Barazangi 1986; Berberian 1995; McQuarrie 2004). The GPS-derived velocities demonstrate that shortening is not distributed homogeneously along or across the belt. In the northwest of Zagros shortening is accommodated chiefly along the Mountain Frontal Fault (MFF), where most of the major earthquakes occur, and along the Zagros front (Fig. 2, Hessami et al. 2006).

Fig. 2
figure 2

a Map view of the Instrumental seismicity (U.S. Geological Survey 2020) over the period 1980–2023 in terms of magnitude and depth; b swath profile (with a width of 2 km) across the MFF with c the projection of earthquakes within a buffer of 25 km from the swath midline; d stem plot of earthquake datetimes with the reported SAR (synthetic aperture radar) interferometry InSAR data temporal range to place the following analyses in time. The coordinate reference system is WGS84

Full size image

Most of the large earthquakes in the Zagros occur within the Simply Folded Belt characterised by large open anticlines (Talebian and Jackson 2004). Most focal mechanism solutions of earthquakes in the Zagros suggest the presence of active reverse faults (e.g., Nowroozi 1972; Jackson and McKenzie 1984; Gillard and Wyss 1995). The determined focal depths in Zagros suggest NE-dipping reverse faults in the basement (e.g., Jackson and Fitch 1981; Jackson and McKenzie 1984; Berberian 1995). Hence, the active MFF is considered to be a segmented major seismogenic blind thrust in the basement of Zagros with important structural, topographic, geomorphic, and seismotectonic features (Berberian 1995). The noticeable subsidence of the Zagros Foredeep and the Dezful Embayment, along with the thickening of the Gachsaran evaporites and the Aghajari-Bakhtiari synorogenic molasse, provides evidence of relative motion along the MFF and the Dezful Embayment fault since the Early Miocene (Berberian 1995). As a consequence of MFF activity, the southwestern edge of the Simply Folded Belt is being uplifted due to vertical motion along the MFF. The current seismotectonic frame of the Zagros shows that this active region accounts for the large deformation rates linked to the oblique Arabia–Eurasia convergence (e.g., Palano et al. 2018).

The Siah Kuh anticline lies upstream of the MFF. The MFF is a NW- to SE-trending blind basement thrust (a major topographic front) that causes up to 5 km of step in the structural elevation (e.g., Falcon 1961; Emami et al. 2010; Tavani et al. 2020). The 40-km-long Siah Kuh anticline, which is one of the frontal anticlines of the Simply Folded Belt, is an asymmetric fold characterised by a steep forelimb and a moderate to gently dipping backlimb. This anticline is part of a series of sigmoidal anticlines that are located in the south termination of the Lorestan arc (between Lorestan in the North and Dezful Embayment in the South). The sigmoidal shape of the Siah Kuh anticline was achieved in the Pliocene due to vertical axis rotations of an already existing anticline (Casini et al. 2018). The Sarvak and Ilam Formations form the core, and the Gurpi and Padbeh Formations flank the core of the anticline and separate the Bangestan Group from the Asmari Formation which forms a topographic relief encircling the anticline (Casini et al. 2018). This anticline is considered as a fault-propagation fold which is cut by the Balarud lineament.

Materials and methods

To reconstruct the long-term to present-day evolution of the Siah Kuh DGSD, it was necessary to tune an integrated methodology. In particular, satellite interferometry and OSL dating of geomorphic markers along the Dowairij River were implemented in order to quantify the landscape evolution rates at different scales. Indeed, satellite interferometry is focused on a decennial scale and can be used to detect active landsliding areas and further analyse their short-term histories and trends (e.g., Smith et al. 2004; Lan et al. 2022; Yao et al. 2022), while geomorphic markers cover a time window of more than 105 years (Delchiaro et al. 2019, 2023). Since the interferometry scale can be considered as a static “photograph” of the slope at present in comparison with the longer histories obtained by OSL dating of geomorphic markers, satellite interferometry results were used for discriminating the present landslide mechanisms in the Siah Kuh slope. Conversely, OSL dating of geomorphic markers aimed at reconstructing the long-term evolution of the drainage network across the regional tectonic features and at defining its preparatory role over the deformation history of the Siah Kuh slope.

Remote and field surveys

Remote surveying was first carried out using Google Earth satellite images (2018 Landsat Imagery) to preliminary identify gravity-induced landforms related to the Siah Kuh landslide system, and satellite Synthetic Aperture Radar (SAR) interferometry data to verify the presence of ground displacement in association with the detected gravity-induced landforms. The initial reference for creating geological cross-sections of the landslide system was provided by geological maps from the National Iranian Oil Company (NIOC) at a scale of 1:100,000 (Llewellyn 1973; Setudehnia and Perry 1967). However, the accuracy and detail of the final map and sections were enhanced through field surveys, which allowed for cross-checking and refinement of the mapping process. Additionally, remote sensing techniques were used, taking advantage of the sparse vegetation in the landscape to further improve the accuracy of the mapping.

Satellite SAR interferometry was applied using a STACK approach that uses a multiple set of images with the same acquisition geometry, to create interferograms related to a given area (Ferretti et al. 2000, 2001; Berardino et al. 2002; Hanssen 2005; Bert 2006). Among the various techniques, the Persistent Scatterer Interferometry (Ferretti et al. 2000, 2001) was used, based on the observation of time- and space-coherent pixels, the so-called “persistent scatterers”, represented by anthropogenic and natural structures like buildings, antennas, and exposed rocks. The SARscape software tool was used to process 147 Sentinel-1 SAR images for the ascending orbit in the time frame between 17 October 2014 and 31 March 2019, as well as 132 Sentinel-1 SAR images for the descending orbit in the time frame between 06 October 2014 and 26 March 2019. A coherence threshold of 0.75 has been used to obtain an improvement in the signal–noise ratio while guaranteeing at the same time a good persistent scatterers density. Since the displacement recorded by the SAR sensor is the unidimensional average velocity along the satellite line-of-sight (LOS) projection, we resolved the actual tridimensional movement of the velocity vectors along the horizontal (east–west, Vh) and vertical (up–down, Vv) directions (Raspini et al. 2017; Tofani et al. 2013; Notti et al. 2014) using the GIS-based PS-Toolbox, a SAR interferometry post-processing software developed by NHAZCA S.r.l. The vectorial decomposition plugin needs two distinct layers for both geometries and their respective orbit parameters (i.e., heading and incidence angles) as input data.

The remote analysis led to the first detection of gravity-induced landforms and geomorphic markers. Then, a field cross-check was also performed. Specifically, during the field survey, we mapped the most significant active landforms and checked the relict ones extracted by the semi-automatic procedure described in detail in the “Geomorphometric analysis” subsection. Furthermore, the sampling of two strath terraces to be dated with the OSL method (Murray and Olley 2002; Wintle and Murray 2006; and references therein) was carried out. The details are reported in the “OSL dating method” subsection.

Geomorphometric analysis

Stream longitudinal profiles retain important characteristics, including non-lithological knickpoints that signify variations in rock uplift rates, reflecting changes in base level (e.g., Berlin and Anderson 2007; Delchiaro et al. 2019, 2021c; Moumeni et al. 2024; Wang et al. 2024). In this regard, the plano-altimetric distribution of the main geomorphic markers along river profiles allowed to constrain the Dowairij River valley evolution. In order to preliminary map the geomorphic markers of the river valley evolution, the semi-automatic detection of terrace treads was performed. The ALOS PALSAR Digital Elevation Model (DEM) with 12.5 m resolution (Rosenqvist et al. 2007) was used in the analysis. The drainage network was extracted and analysed using the MATLAB©-based TopoToolbox (Schwanghart and Scherler 2014) and the Topographic Analysis Kit (TAK) by Forte and Whipple (2019). The flow accumulation threshold was set according to that proposed for the fluvial domain (106 m2) by Montgomery and Foufoula-Georgiu (1993). The TopoToolbox function ‘STREAMobj2SWATHobj” was used to create the swath profile of Dowairij River with a width of 1000 m. Then, the TAK function “ProjectOntoSwath” was used to project points (identified geomorphic markers) on the river swath profile. Specifically, the identified geomorphic markers are flattish surfaces associated with treads of strath terraces cross-checked in the field.

The semi-automatic extraction of terrace treads was performed using the TopoToolbox (Schwanghart and Scherler 2014) and following the procedure described by Delchiaro et al. (2024). This methodology draws inspiration from the approach outlined for identifying marine terraces as described in Jara-Muñoz et al. (2016, 2019). The semi-automatic detection of low-relief and gently inclined smooth areas is based on Surface Classification Models (SCM) developed by Bowles and Cowgill (2012) that use a combination of topographic slope (SLP) and roughness (RG), the latter being defined as the standard deviation of the slope (Frankel and Dolan 2007).

$${\text{SCM}}= \left(\frac{\mathrm{SLP }<\mathrm{ SLPr}}{{\text{SLPr}}}*0.5\right)+\left(\frac{\mathrm{RG }<\mathrm{ RGr}}{{\text{RGr}}}*0.5\right)$$
(1)

where SLPr and RGr are defined respectively as slope and roughness cut-offs.

The distribution of slope and roughness were truncated, respectively, at 5° and the 0.9 quantile of RG. Since false-positive classifications may usually occur at valley bottoms, a 25-m buffer distance from the drainage network was considered to filter out such erroneous classifications. To extract the elevations above the thalweg, the resulting SCM patches were intersected with the values of a Relative Elevation Model (REM), also known as a Height Above River (HAR) raster (Dilts et al. 2010) computed using TopoToolbox (Schwanghart and Scherler 2014). Finally, the obtained terrace treads with relative elevation values were distinguished by upstream and downstream of the MFF, in order to identify possible flights of terraces.

OSL dating method

OSL sampling is a very delicate and quite complex technique; therefore, it is necessary to prevent the sample from being exposed to light because the luminescence signal could be reduced or even reset. In choosing the most suitable site to sample, of course, levels were identified with original sedimentary structures, avoiding bioturbations and post-depositional alterations. Once the site for sampling was identified, it was important to carefully clean off the slope and prepare, according to the consistency or cementation of the material, the equipment necessary for taking the sample, without it being exposed to light. Furthermore, to minimise the effects of cosmic radiation and thereby avoid the risk of rejuvenated ages, the samples were taken at least 1 m below the topographic surface (or below eventual erosional surfaces identified within the deposits). The deposits, mainly characterised by fine-grained loose sediments (size < 2 mm) were sampled by a hammer to insert a metal tube horizontally into a vertical face, which must be isolated from light and humidity immediately after collection. To maximise the uniformity of the natural radioactivity of the burial period, the tube was inserted into zones of homogeneous sediment at least 30 cm wide and thick. From the same level where it was sampled, an additional 500–800 g of sediment was extracted to evaluate natural radioactivity (if the annual dose rate measurement is not performed in situ), for the mineralogical and grain size analysis, as well as to determine the moisture content. The OSL dating was performed by the LABER OSL Laboratory, in Waterville, OH, USA. Quartz was extracted for the equivalent dose (De) measurements. In the OSL laboratory, the sample was treated first with 10% HCl and 30% H2O2 to remove organic materials and carbonates, respectively. After grain-size separation, the fraction of 90–125 µm size was relatively abundant, so this fraction was chosen for De determination. The grains were treated with HF acid (40%) for approximately 40 min to remove the alpha-dosed surface, followed by 10% HCl acid to remove fluoride precipitates. Luminescence measurements were performed using an automated Risø TL/OSL-20 reader. Stimulation was carried out by a blue LED (λ = 470 ± 20 nm) stimulation source for 40 s at 130 °C. Irradiation was carried out using a 90Sr/90Y beta source built into the reader. The OSL signal was detected by a 9235QA photomultiplier tube through a U-340 filter with a 7.5 mm thickness. For De determination, the single-aliquot regenerative dose (SAR) protocol (Murray and Olley 2002; Wintle and Murray 2006) was adopted. The preheating temperature was chosen to be 260 °C for 10 s and then cut heat was 220 °C for 10 s. The final De is the average of the De of all aliquots, and the final De error is the standard error of the De distribution. For each sample, at least 12 aliquots were measured for De determination. The De was measured using SAR on quartz, and the aliquots that passed criteria checks were used for the final De calculation. Recycling ratios were between 0.90 and 1.1, and recuperation was relatively small. The cosmic ray dose rate was estimated for each sample as a function of depth, altitude, and geomagnetic latitude. The concentration of U, Th, and K was measured by neutral activation analysis (NAA). The elemental concentrations were then converted into the annual dose rate, considering the water content (lab-measured) effect. The final OSL age is then De/Dose rate.

Results

Remote and field surveys

The fluvial and gravity-induced landforms are shown in the geomorphological map of Fig. 3. All the fluvial and runoff landforms are carved into bedrock and can be distinguished into a main bedrock channel corresponding to the Dowairij River and active gully tributary streams. It can be noted also that the main channel direction changes from W–E to NNW–SSE extending from the backlimb to the forelimb of the fold. In Fig. 3a, b, and c, a focus of the satellite image on the main landforms detected by a remote survey is shown, while the landforms checked during the fieldwork are shown in Fig. 3d, e, and f. The study revealed four distinct zones characterised by varying gravitational stress. Zone A (Fig. 3a), situated downstream of the fold axis in the fold’s forelimb, is marked by ongoing deformation landforms like tension cracks within the Asmari Formation. Zone B (Fig. 3b and e), positioned upstream of the fold axis in the fold’s backlimb, showcases a system of rototranslative scars involving the contact between the Pabdeh-Gurpi and Asmari Formations. Additionally, zone C, located in the fold’s forelimb, displays compressional bulges primarily developed in the Pabdeh-Gurpi Formation (Fig. 3c and f). Furthermore, a minor rototranslative landslide within the deformation system was observed along the river gorge, characterising zone D (as shown in Fig. 3d).

Fig. 3
figure 3

Geomorphological map of Siah Kuh slope and surrounding areas. The most significant landforms and features are reported on the Google Earth satellite image (source—Google Earth, Image© 2023 CNES/Airbus and Image© 2023 Maxar Technologies). Google Earth satellite images (a–c) and pictures taken during the field survey performed on August 2019 which represent different typologies of gravity-induced landforms: a tension cracks, b scars, c bulging, d auxiliary landslide along the river gorge, e scars, and f bulges

Full size image

Satellite SAR interferometry

The maps of ground displacements derived from advanced satellite SAR interferometry in both ascending and descending geometry as well as the vertical and horizontal components are shown in Fig. 4, along with some time series related to the areas described in the previous section. Specifically, the previously identified zone A is characterised by the maximum displacement with values around − 10 mm year−1 in ascending geometry, and around 3 mm year−1 in descending geometry. The vertical and horizontal components are, respectively, around − 6 and − 10 mm year−1. Zone B shows minor displacements, around − 4 mm year−1 in ascending geometry and around 3 mm year−1 in descending geometry. While the vertical and horizontal components are similar, ranging from − 2 to − 4 mm year−1. Zone C, on the other hand, is characterised by an approximately stable signal, except for some PSs with shifts around 2 mm year−1 in both geometries. The vertical and horizontal components confirm the same stable or slightly positive ground displacement rates. Finally, zone D shows minor shifts, around − 2 mm year−1 in both geometries. The vertical component is characterised by values of − 4 mm year−1 that is higher than the horizontal one that records a stable signal. From the reported time series, it can be observed that all displacements are characterised by a constant rate without acceleration. In this regard, as shown in Fig. 2d, the interferometric analysis period has no seismic events with a magnitude greater than 5, nor any intense rainfall event that could have influenced the time series rates. The recorded displacement rates demonstrate reliability in the context of the slow-to-fast MRC process, particularly when compared to measurements obtained in case studies with comparable mass volume and structural settings (Maffei et al. 2005; Lenti et al. 2015).

Fig. 4
figure 4

Surface velocity maps in ascending and descending geometries (a–b) and up–down and east–west velocity components (c–d) obtained by using a coherence threshold of 0.75 on persistent scatterers. Ground displacement time series related to zone A (e and f), zone B (g and h), and zone C (i and l), respectively, in ascending and descending geometry. The coordinate reference system is WGS84 UTM 38N

Full size image

The obtained displacement rates are not useful for back-and-fore scenarios as several limits affect the reliability of the interferometric technique to detect MRC evolution, especially, when close to the failure stage (accelerating creep). The main limitation is related to the difficulty of transposing a decennial observation on an extremely long process, especially in the case of catching the phase of a multi-stage creep process spanning within a large time window. A high possibility of confusing short-term temporary change of the displacement trend with the transition between the different displacement trends that characterise the 3 different phases of the creep, indeed, exists. Nonetheless, the good fit of the satellite interferometric evidence with the slope accessory landforms, which testify to an ongoing MRC process, is encouraging.

Figure 5a and b shows the geological cross-sections related to the deforming slope from which the layers dipping downward with a dip angle of around 10° are observed. It is also inferred that the deformation is closely related to the evolution of the river network, whose vertical incision of about 600–500 m generated the deformation that migrated retrogressively by about 1800–2000 m. As shown in the swath profiles of Fig. 5, the NE sector is characterised by pronounced rototraslative kinematics with larger displacements with prevalent horizontal component in the slope zone (zone A) and opposite though smaller displacements in the distal zone (zones C and D). While the SW sector (zone B) is characterised by smaller displacements with similar vertical and horizontal components that are concentrated where the range in elevation of the swath profile is greater, suggesting lateral-spreading kinematics to be defined.

Fig. 5
figure 5

a and b Geological cross-sections of the Siah Kuh deforming slopes (colours refer to geological legend reported in Fig. 1). Swath profiles with a width of 500 m in ascending and descending geometries (c–f) and with up–down and east–west velocity components (g and h) along different sectors of the landslide system along which the persistent scatterers are projected. The swath profiles and geological cross-sections are reported in Fig. 4 as traces 1 and 2

Full size image

Analysis of geomorphic markers

The application of the SCM equations led to the preliminary identification of several flat, poorly rough surfaces that can be associated with river terraces. Therefore, the field survey allowed us to check their presence and to define their main features, as well as to sample their deposits for OSL dating.

The best geomorphic markers preserved in the study area are represented by two suites of strath terraces (Fig. 6). One suite is located entirely upstream of the MFF while the other is preserved downstream.

Fig. 6
figure 6

Geomorphic markers upstream and downstream the MFF dated with the OSL method. a Overall view of the tip of the Siah Kuh fold upstream of the MFF and the associated sampled strath terrace level at 325 m a.s.l. (b, c, and d). e and f Strath terrace-level sampled downstream of the MFF. The location of the photos a, b, c, and d are reported in Fig. 3, while the location of e and f is in correspondence of the S2 sampling point reported in Fig. 7

Full size image

They testify to the mutual evolution between the footwall and the hanging wall of the MFF (Fig. 7). The hierarchy of the different orders was based on the distribution of elevation values above the thalweg of terrace treads upstream and downstream of the MFF, respectively. Specifically, as shown in the histograms in Fig. 8, relative elevation ranges were identified for each order, the delineation of which was done by natural breaks. In each case, the classification covered terrace areas within 150 m above the thalweg.

Fig. 7
figure 7

Map of the study area in which the ordered flat surfaces and OSL sample location are shown. The coordinate reference system is WGS84 UTM 38N

Full size image
Fig. 8
figure 8

Histograms of the relative elevation values associated with the flat surfaces of the downstream (a) and upstream (b) suite of strath terrace treads. c and d Elevation offset between the two suites for all terrace orders

Full size image

In the downstream suite, 4 terrace orders were identified, whose relative height ranges from 0 to 20.5 m for the fourth order, from 20.5 to 62.5 m for the third order, and from 62.5 to 71.5 m for the second order and is greater than 71.5 m for the first order. Four terrace orders were also identified in the upstream suite, whose relative height ranges from 0 to 30 m for the first order, from 30 to 78.5 m for the second order, and from 78.5 to 89.5 m for the third order and is greater than 89.5 m for the fourth order. Comparing the histograms of the two suites, it can be seen that each order step occurs at lower elevations in the downstream suite and that the ranges of values are wider for the upstream suite. For example, the transition between third and second order in the downstream suite occurs at 62.5 m while in the upstream suite at 78.5, thus 16 m greater. The range of the third order in the downstream suite is 42 m (20.5 to 62.5 m), while upstream is 48.5 (30 to 78.5 m).

As reported in Figs. 6 and 9, samples for OSL dating were collected from the fourth-order level of the downstream suite and from the third-order level of the upstream suite. In Table 1, the results of OSL dating are reported. The fourth order strath terrace downstream of the MFF dated at 4.77 ± 0.3 ka, while the third order upstream of the MFF dated at 9.54 ± 0.60 ka.

Fig. 9
figure 9

Projection along the swath longitudinal profile of the Dowairij River (with a width of 1000 m) of the geomorphic markers. It is noteworthy that in correspondence with the landslide system, terrace treads are not preserved

Full size image
Table 1 OSL ages obtained for the geomorphic markers recognised in the Dowairij River valley. Detailed information and photos of sampling sites are available in the text and in Fig. 6
Full size table

The traditional model for the formation of strath terraces involves the gradual widening of the river valley floor to create a broad, flat, bedrock surface into which the river later incises, leaving behind a strath (Burbank and Anderson 2012). In this regard, the ratio between the thickness of the eroded bedrock below the strath surface (∼27 m and 2 m, for samples S1 and S2, respectively) and the time elapsed since the beginning of the incision process (9.54 ± 0.60 ka and 4.77 ± 0.3 ka, for samples S1 and S2, respectively) allow for estimation of an incision rate (corresponding to an uplift rate) of 2.8 ± 0.2 mm year−1 and 0.42 ± 0.03 mm year−1.

As can be observed from the swath longitudinal profile of Dowairij River (Fig. 9), the topography is generally steep at high elevations along the river course, and a major break in slope is noted downstream of the Siah Kuh fold in the proximity of the MFF. Along the swath profile of the river, the maximum relief (the difference between the maximum and minimum elevation) generally characterised the overall reach upstream in comparison with the downstream prosecution. In detail, the upstream sector shows the maximum elevation gaps in correspondence with the intersection of the Siah Kuh fold. In this perspective, the elevation gap is proportional to the erosional stage of the examined sector, and the morphology of the minimum curve is an indicator of the activity of the process that generated it.

The projection of the geomorphic markers along the river profile highlights what is already evident in the histogram distribution of the relative elevation values of the two suites. Indeed, the downstream suite is characterised by lower elevation ranges between tread orders than the upstream suite that shows a more scattered distribution. Furthermore, the fourth-order treads in the upstream reach crop out less frequently than the downstream ones. It is also possible to observe at the intersection of the MFF with the longitudinal profile of the river a variable offset between the relative elevation of the terrace orders. As shown in Fig. 8, an increasing offset from the fourth to the first terrace order is recorded, from 0 to about 50 m of dislocation.

Discussions

Siah Kuh landslide system

Transient, tectonically active mountain landscapes are widely affected by landslides in response to gravity-induced disequilibrium on hillslopes. In this regard, the Siah Kuh landslide system demonstrated to be an excellent example where the coupling between tectonic deformation and river incision originated the predisposing and preparatory conditions for the evolution of an MRC process. Indeed, the InSAR analysis period has no seismic events with a magnitude greater than 5, nor any intense rainfall events that could have influenced the time series rates, and as shown from the InSAR time series, no changes in the rate trends have been observed.

More in detail, remote and field surveys allowed us to identify four different zones characterised by different gravity-induced stress fields. The whereabouts of every ground displacement were well-mirrored by the spatial location of the most rapidly moving radar targets. In detail, satellite SAR Interferometry was applied in order to confirm the above-described evidence and, in particular, to quantify the ground displacement rate, as proof of a steady state (i.e., secondary) rate, which progressed evolving from lateral spreading to rototranslative mechanism, respectively, located in the forelimb and the backlimb of the fold.

These two failure modes can be interpreted as two different evolutionary stages of the same process, transposed spatially through the fold axis. In fact, the rototranslative mechanism could represent an advanced phase of the deformation evolution while the lateral spreading is an earlier phase. A similar diachronous process evolution has been observed in a landslide system in central Sardinia characterised by plateaus with a prominent limestone scarp overlying extremely altered metamorphites (Demurtas et al. 2021a, b). In detail, different evolutionary stages from lateral spreading to rock avalanches passing through sackung type have been identified (Demurtas et al. 2021a, b).

In our case study, the diachronous evolution of the process is likely due to the different positions along the axis of the fold. Hence, in the axial zone where uplift and fluvial incision are greater the relief is higher, while in the tip zone, the relief is lower. However, the entire landslide system is located upstream of the MFF, whose uplift rate was determined through OSL dating of two strath terraces located upstream and downstream of the fault. The resulting uplift rate of 2.8 ± 0.2 mm year−1 upstream of the MFF in comparison with that recorded downstream of 0.42 ± 0.03 mm year−1 is in agreement with the local uplift rates obtained by Oveisi et al. (2007, 2009) for the Mand and Gisakan folds, ranging between 0.8 and 3.3 mm year−1 in the Fars region. Indeed, these morphostructures, as well as the Siah Kuh anticline, are associated with the MFF, appearing to account for up to 50% of the geodetically derived shortening rates (Oveisi et al. 2007, 2009). The Kirkuk of the Kurdistan region recorded an average regional rock uplift rate in the range of 0.54–0.83 mm year−1 (Le Garzic et al. 2019; Tozer et al. 2019) and higher rates of 0.7 and 1.35 mm year−1, respectively, in the Safin and Sarta anticlines close to the MFF boundary (Zebari et al. 2021). In this area, the cumulated shortening rates deriving from the vertical uplift rates sum up to less than half of the geodetically obtained shortening rates in the belt (Zebari et al. 2020, 2021). However, since this study does not cover the entire Lorestan belt, presenting data only on the SW tip termination of the Siah Kuh fold, as well as due to the unavailability of comprehensive geodetic studies with high resolution, further research is welcome for assessing the absorbing capacity of the MFF along the Lorestan arc in comparison with the entire Zagros belt.

Mass rock creep conditioning factors

The MRC is one of the main processes which induce the damaging of rock masses in transient tectonically active mountain landscapes. Although the limit of theories is represented by the difficulty of accurately estimating the starting time of the process as well as of discriminating the distinct phases, reconstructing the morpho-evolutionary history of the river valley slope is key in shedding light on the evolution of the gravity-induced deformation as well as into the evaluation of many other interrelated variables. In this sense, Quaternary landscape evolution modelling of slope-to-valley floor systems plays a fundamental role as a tool for chronologically constraining the creep evolution of entire slopes (Bozzano et al. 2016; Della Seta et al. 2017; Troiani et al. 2020; Delchiaro et al. 2019, 2022), the nonlinearity of the time-displacements, as well as the accurate history of the past erosion rate evolution and its link with the deforming slope.

A comparison among the morpho-evolutionary models related to three case studies (Siah Kuh, Seymareh, and Loumar landslides) in the Lorestan region of Zagros Mountains is reported in Fig. 10, in relation to other similar cases that occurred worldwide. The Seymareh rock avalanche, detached ∼10 ka from the northeastern flank of the Kabir Kuh fold (Zagros Mts., Iran), is recognised worldwide as the largest rock slope failure (44 Km3) ever recorded on the exposed Earth surface (Watson and Wright 1969; Roberts and Evans 2013; Shoaei 2014; Delchiaro et al. 2019, 2020, 2022; Rouhi et al. 2019, 2022). The Loumar rockslide affected the NE slope of the Gavar anticline, and its evolution is strictly related to the vertical and lateral growth of the anticline and to the evolution of the Seymareh River drainage system. The failure that occurred at 5 ka was not catastrophic, the rock mass did not fragment and moved as a block slide retaining its integrity during the motion (Delchiaro et al. 2023).

Fig. 10
figure 10

Illustrated sketch of predisposing and preparatory factors driving the evolution of MRC of the three case studies (Seymareh, Siah Kuh, and Loumar landslides) from Zagros belt. Examples of case studies described in the text are also reported for the three landslide types. The critical volume for Siah Kuh and Loumar cases was computed by multiplying the deformed mass by 0.2 km taken as the average thickness of the collapsed Asmari Fm

Full size image

In all these three cases, the morpho-evolutionary study was decisive both for reconstructing the slope-valley system evolution before the failure and for identifying the main predisposing and preparatory factors that controlled slope failures. Based on a detailed case study comparison, it is possible to distinguish predisposing and preparatory factors, among which, for the predisposing ones as follows: (1) the overlay of stratigraphic units with significantly different rheological behaviour, (2) a bedding moderately dipping downslope, (3) the fold style, (4) the position of the deforming mass in the morphostructure, and (5) the presence of pervasive joint sets or non-systematic planes. The main preparatory factors are (1) local relief, (2) incision rate, (3) involved volumes, and (4) persistence time of the deforming volume under quasi-static incision stages.

Although earthquakes or other triggering events cannot be neglected as a possible cause of the acceleration of ongoing deformations (e.g., Keefer 1994; Malamud et al. 2004; Owen et al. 2008; Parker et al. 2011; Hovius et al. 2011; Li et al. 2014), we chose to focus just our interpretations on predisposing and preparatory factors. Indeed, since triggers can occur when the creep state of the slope has already advanced, their effectiveness in causing slope failure is strongly dependent on the creep deformation stage, which controls stress states and the degree of evolution of the slope damage. At this aim, further analysis should be evaluated through stress–strain numerical modelling accounting for a detailed morpho-evolutionary model of the valley-slope system (Della Seta et al. 2017; Marmoni et al. 2023).

Time-independent predisposing factors

The stratigraphy of the source rock mass, also described in detail by Roberts and Evans (2013) for the Seymareh case, accounts for different rheological behaviours, which could have induced localization of strain rates within the slope and the evolution of a shear zone leading to failure. Although quantification of the rheological properties is lacking for the lithological units of the folds affected by slope failures (Gavar, Kabir-kuh, and Siah folds), some inferences can be made according to previous studies on rock masses affected by MRC (Apuani et al. 2007; Bozzano et al. 2012; Bretschneider et al. 2013; Della Seta et al. 2017; Marmoni et al. 2023). More particularly, the time-dependent viscoplastic behaviour, more typical of clayey and marly deposits, which have lower viscosity values, can justify time-dependent strains. In fact, a stiffness contrast exists between the upper member of the Pabdeh Formation and the overlying Asmari Formation. The bedding is moderately dipping downslope (around 15° for Siah case and around 20° for Seymareh and Loumar ones), leading to a lower vertical thickness of (and consequently a reduced lateral confining effect by) the epicontinental and continental units, such as the Gachsaran and Agha Jari Formation. Moreover, the low dip angle of the strata also reduces the vertical thickness of the Asmari Formation caprock (Delchiaro et al. 2019) and consequently the incision necessary for kinetic freedom of the landslide, relative to a more steeply dipping sequence, which would require a greater depth of the incision. Overlays of stratigraphic units with significantly different rheological behaviour have been noteworthy recorded as a predisposing factor in the literature. For instance, in Meseta Belgrano (Argentina), where Pánek et al. (2020, 2023) recorded one of the largest concentrations of large slope failures (including rock avalanches and rockslides) with volume > 108 m3 in Patagonia, the rock mass strength and the strata dip play a key role in predisposing both the kinematics. Indeed, stratigraphy is composed of mainly sub-horizontally bedded sedimentary and volcanic rock with contrasting litho-technical rheology. Other examples come from the large rock-slope failure inventory (including rock avalanches and rockslides) for the Kyrgyz Range by Strom and Korup (2006) and for Central Asia (Strom and Abdrakhmatov 2018). As also discussed by Strom (2021), the type of rock involved in the large-scale slope failure could be informative about the initial rock slope failure, as in the case of planar rockslide that is common in sedimentary or metamorphic rocks with well-pronounced bedding or schistosity and monoclinal structures. Another exemplary case confirming the role of stratigraphic and geo-structural setting in rock sliding initiation is given by “La Clapiere” rockslide in the Western European Alps (e.g., Julian and Anthony 1996; Cappa et al. 2004; Guglielmi et al. 2005; El Bedoui et al. 2009), where the slope is underlain by a metamorphic basement including gneiss and migmatite whose foliation is parallel to the adjacent valley but locally toppled with a dip of around 40°. In addition, many other active slides, such as the Downie Slide (British Columbia, Donati et al. 2020), the Aknes rockslide (Norway, Ganerød et al. 2008), or the Campo Vallemaggia rockslide (Switzerland, Bonzanigo et al. 2007) display along surfaces parallel to bedding, foliation or schistosity located in correspondence with structural material heterogeneities. Also, for spreads, it has been demonstrated the predisposing role of contrasting lithologies as in the cases in the Betic Cordillera (Azañón et al. 2005; Alfaro et al. 2019). Specifically, it has been reported by Azañón et al. (2005) that in the Guadix basin, large-scale rotational slides (500–1500 m3) with spreading kinematics develop on an almost horizontal succession composed of conglomerate over impermeable clay lithologies. Alfaro et al. (2019) reported an impressive case of lateral spreading located in the Sierra de Aitana, originating along the northern limb of faulted Sierra de Aitana anticline, whose stratigraphy is characterised by 20° dipping strata composed of Eocene limestone overlaid on Eocene marls.

Time-dependent preparatory factors

As it regards the preparatory conditions due to MRC process, the volume of rock mass involved in the viscous deformation, as well as the time span available for the creep evolution, can be regarded as the most important discriminating factor in the evolution of the phenomenon towards a paroxysmal failure (such as a rock avalanche), or a medium-slow movement (such as a rockslide), or slow movement (such as lateral spreading). Moreover, the relief inherited by the deepening of the river valley adjacent to the deforming slope plays a crucial role in developing such a kind of instabilities as directly proportional to the potential gravitational energy involved in the slope, as testified in the case of the MRC which involved the Mt. Conero anticline (Italy) up to its generalised collapse (Troiani et al. 2020; Marmoni et al. 2023).

More in detail, for the here considered case study, the persistence time effective for the creep process, as the time window, constrained by the kinetic freedom of the slope to downslope mass movement up to when the movement can be actually recorded, is in the order 104–105 years for Siah and Loumar cases, while the Seymareh case is characterised by the longest time span. Regarding the deformed volume, Seymareh is characterised by around 44 km3, in comparison with 1.6 and 0.6 km3 for Siah and Loumar cases, respectively. Finally, although relief energy values range from 400 to 600 m for Loumar and Siah to 2000 m for Seymareh, the river incision rates are quite similar among the cases. The rate of the Seymareh case is indirectly determined by using the rates obtained with numerical simulation of a fold growth in Zagros Mts. (Collignon et al. 2016), while the rate from the Siah case is directly determined from OSL dating of fluvio-lacustrine sediments (Delchiaro et al. 2023). This could be explained by an older and more complex tectonic history associated with the Kabir Kuh fold.

In a broader view, it has been demonstrated that morphoevolution plays an important and dynamic role, often difficult to be constrained, in dynamically preparing the slope to failure. River incision and slope undercutting in response to base level changes due to climate (Azañón et al. 2005), tectonics (Strom and Korup 2006; Strom and Abdrakhmatov 2018; Alfaro et al. 2019), glacio-isostatic uplift (Pánek et al. 2020, 2023), or debuttressing during deglaciation (e.g., Julian and Anthony 1996; Cappa et al. 2004; Guglielmi et al. 2005; El Bedoui et al. 2009) are the most frequent geomorphological processes kinetically and tensionally releasing slopes isolating a rock volume that starts to deform until failure.

As a final remark, we can conclude that the deforming rock mass volume, together with the relief energy, is interrelated with the evolution time (Li et al. 2022) so it is possible to hypothesise a close correlation between the paroxysmal failure mechanism (e.g., Seymareh rock avalanche) and the achievement of a critical volume (Esposito et al. 2007) with respect to more progressive deformations such as viscous sliding (e.g., Loumar case study) involving considerably smaller volumes of rock masses.

This finding also appears in line with the energetic interpretation of the deformation process which evolves into almost paroxysmal failure, since the combined contribution of inertia (mass) associated with a strain rate contributes to the quantification of the kinetic energy connected to the creep deformation and, as a consequence, to the quantification of the potential energy release in the ultimate phase of failure.

Conclusions

This research provided valuable insights for the quantification of the combined role of time-dependent rock mass deformations and landscape evolution rates in favouring massive rock slopes to failure. The main aim of this study was to evaluate the present-day slope deformation stage in the frame of a complex landscape evolution by taking into account tectonic activity, inferring erosion rates, and monitoring ground displacement rates. By employing advanced techniques such as SAR interferometry analyses, geomorphometry, and OSL dating, we successfully assessed ground displacement rates associated with the gravitative and river incision rates, proving evidence of ongoing deformation process, that are genetically interlinked to the uplift evolution of the Siah Kuh fold, located upstream of the MFF.

From a comparison among the morpho-evolutionary models related to three case studies (Siah Kuh, Seymareh, and Loumar landslides) in the Lorestan region of Zagros Mountains, we then obtain a comprehensive understanding of the main static predisposing and dynamic preparatory factors and mechanisms governing large rock slope deformation.

The following are the main static predisposing factors results:

(1) the overlay of stratigraphic units with significantly different rheological behaviour (Asmari and Pabdeh Formations), (2) a bedding moderately dipping downslope (from 8 to 25°), (3) the presence of pervasive joint sets or non-systematic planes, (4) the folding style (asymmetric folds, such as fault propagation folds, or fault bend folds), and (5) the position of the deforming mass in the anticline morpho-structure.

The following are the main preparatory factors results:

(1) Local relief (from 400 up to 2000 m), (2) Incision rate (from 0.8 to 2.8 ± 0.2 mm year−1), (3) Involved volumes (from 0.6 up to 44 km3), and (4) Persistence time of the deforming volume under quasi-static incision stages (from 104 to 105 years).

As a final remark, we can argue that the deforming rock mass volume, together with relief energy, is interrelated with the evolution time so that it is possible to hypothesise a close correlation between the achievement of a critical volume and an ultimate and paroxysmal failure (e.g., Seymareh rock avalanche) in comparison with progressive failure (e.g., Loumar case study) involving considerably smaller volumes of rock masses.