Relief rejuvenation of geologically ancient areas often occurs in response to intraplate deformation. In these long-lived areas, the reactivation of pre-existing structures is a common mechanism for strain accommodation, which is often favoured over the generation of new structures (e.g. Butler et al., 1997). Thus, previous discontinuities can be expected to localise the relief rejuvenation associated with recent or active tectonics, which can therefore be analysed to delve into the kinematics of recent deformation.

The relief of the Variscan basement outcropping in the western and central Iberian Peninsula is characterised by roughly W–E or WSW–ENE topographic highs (1000 to 2500 m a.s.l.) separated by Cenozoic sedimentary basins (Fig. 1a). On the basis of both numerical and analogue models, these topographic domains have mainly been interpreted as the expression of lithospheric folding (Cloethingh et al., 2002; Fernández-Lozano et al., 2012) due to the superposition of successive intraplate deformation phases from the Late Cretaceous–Early Cenozoic onwards. They have been mainly associated with plate boundary activity both to the north (Pyrenean–Cantabrian orogen; e.g. Roest and Srivastava, 1991; Teixell et al., 2018) and to the south (Gibraltar Arc; e.g. Spakman and Wortel, 2004; Balanyá et al., 2012) of Iberia (Fig. 1a). In the upper crust, such intraplate deformation is often accommodated by the reactivation of pre-existing faults, which largely control the spatial distribution and orientation of the topographic features (De Vicente et al., 2007; Martín González, 2009; Brum da Silveira et al., 2009; Vázquez-Vílchez et al., 2015).

The southernmost topographic high (900–1300 m a.s.l.) within Variscan Iberia is Sierra Morena (Fig. 1a and b). It is a WSE–ENE relief, approximately 600 km long and 80 km wide, and WSW-ENE relief, the southern limit of which coincides with a regional linear escarpment separating the foreland (i.e. the Variscan Iberian Massif) from the Cenozoic Guadalquivir foreland basin of the Betic Chain (northern branch of the Gibraltar Arc). Sierra Morena as a whole is interpreted as a forebulge, a positive relief that develops, paired with foreland basin subsidence, due to lithospheric flexure in front of major fold-and-thrust belts (e.g. Flemings and Jordan, 1990; DeCelles, 2012, and references therein). In this tectonic context, the Sierra Morena forebulge uplift is linked to the lithospheric flexure of the Iberian Massif under the Betics orogenic wedge load, enhanced by the buckling of the Iberian lithosphere (Cloething et al., 2002; García-Castellanos et al., 2002). Concerning this folding, a NW–SE-oriented compression, connected to the convergence between Africa and Eurasia, is invoked for the neotectonic intraplate deformation of Iberia (Herraiz et al., 2000; Cloething et al., 2002).

This interpretation of the overall Sierra Morena relief as a flexural forebulge folded by NW–SE-oriented buckling is mainly based on the numerical modelling of the lithosphere rheological behaviour (Cloething et al., 2002; García-Castellanos et al., 2002). Nevertheless, little work has been done to address the geomorphological imprint of this tectonic model in Sierra Morena.

In this regard, previous studies propose that the tectonic setting for the current deformation of the western Betic fold-and-thrust belt (FTB) includes the westward migration of the Gibraltar Arc hinterland (Balanyá et al., 2012; Gutscher et al., 2002). This would imply that the tectonic mechanism responsible for the Miocene accretion of the Betics (e.g. Faccenna et al., 2004; Balanyá et al., 2007) is still active. Additionally, analytical and analogue models (Díaz-Azpiroz et al., 2014; Barcos et al., 2015, 2016) provide a N99∘ E–N109∘ E trending horizontal velocity vector to explain the Neogene to recent dextral transpressional kinematics observed in segments of the western Betic FTB (see also García et al., 2016; Jiménez-Bonilla et al., 2015). This velocity vector orientation fits well with such a westward motion of the hinterland relative to the FTB.

Irrespective of the current tectonic setting, Neogene, or even younger, activity has been reported in specific areas of Sierra Morena. Thus, in the Guadiamar drainage basin (westernmost Sierra Morena; GB in Fig. 1c), river migration due to Holocene regional tilting has been described (Salvany, 2004). Additionally, seismicity consistent with reactivation of pre-existing fractures has been pointed out in the Sierra Albarrana area (SA in Fig. 1c; Herraiz et al., 1996), and the upper Miocene infill of the Guadalquivir foreland basin appears locally intersected by faults (Expósito et al., 2016; Matas et al., 2010a).

In this study, we assume that this recent tectonic activity, linked to intraplate tectonics, must have rejuvenated the Sierra Morena relief. Therefore, its geomorphological features must be an expression, among other factors, of its current tectonic setting. Furthermore, we expect that the Sierra Morena relief rejuvenation pattern is significantly conditioned by the reactivation of those previous structures that are favourably oriented.

In this regard, here we analyse the relief rejuvenation of an approximately 130 km long segment of Sierra Morena (Fig. 1b), which is potentially related to the recent reactivation of pre-existing structures in the Betic foreland. To this end, we first analyse the space distribution of key geomorphic features that would characterise relief rejuvenation, i.e. preserved relicts of Miocene landscape surfaces, patches of shallow-water marine Miocene deposits, incised meanders, active knickpoints, perched terraces, stream gaps, and piracy. Second, we localise the discrete tectonic structures that accommodate differential uplift and relief segmentation within the forebulge. We therefore carried out qualitative landform interpretations and quantitative analyses focusing on tributary drainage networks located on the Guadalquivir river right bank, i.e. those streams draining the Betic forebulge southward. The kinematics of these reactivated structures will be used in this study to explore the recent, even current, tectonic scenario compatible with the observed reactivation pattern. Finally, we will discuss previous geophysical data (Herraiz et al., 2000; De Vicente et al., 2007) and relief modelling (García-Castellanos et al., 2002), as viewed from the newly obtained geomorphological and structural data.

Our results prove that the characterisation of relief rejuvenation patterns together with the kinematic analysis of the main relief-controlling structures provides information about both the role of reactivated structures in the relief rejuvenation modes and the tectonic setting in which rejuvenation occurs.

In this study, we analyse an approximately 130 km long segment of the western and central Sierra Morena, including its boundary with the Guadalquivir river valley (Fig. 1). This topographic limit coincides with the contact between two geological units, i.e. the Betic foreland to the north, constituted by the Iberian Massif, and the Betic foreland basin to the south.

The Iberian Massif crops out in western and central Iberia, also constituting the buried basement of the main Cenozoic basins (Fig. 1a). This geological domain belongs to the Variscan chain, a late Paleozoic orogenic belt, which extends from Morocco to central Europe. Regarding its distinctive tectonic features, the southernmost Iberian Massif has been divided into three different zones, from SW to NE, namely the south Portuguese Zone (SPZ), the Ossa-Morena zone (OMZ), and the central Iberian zone (CIZ). Most of our study area belongs to the SPZ and the OMZ, with the CIZ only being represented in its NE limit (Fig. 1b).

In our study area, the SPZ is composed of pre-orogenic Devonian to Carboniferous metadetritics, profusely intruded by igneous rocks, which are locally covered by unconformable post-orogenic Permian sequences (e.g. Oliveira, 1990; Simancas, 1983). In its northern limit, a narrow band of slates with interbedded tuffites and quartzites crops out (Matas et al., 2010b). The OMZ and the CIZ are mainly formed by Neoproterozoic to Carboniferous metadetritic rocks, although lower Cambrian carbonate formations are well represented. Volcanics are frequently interbedded in the Neoproterozoic rocks of both zones and in the lower Cambrian to Ordovician sequences of the OMZ. Plutonic rocks mainly crop out related to the SPZ/OMZ and OMZ/CIZ boundaries (Matas et al., 2010a, b).

The current regional limit between the OMZ and the SPZ is located along the transpressive, from the top to the SW, southern Iberian shear zone (Fig. 1b). It is interpreted as part of a complex, long-lived Variscan suture, including amphibolites of oceanic affinity and high-pressure/low-temperature rocks (e.g. Crespo-Blanc and Orozco, 1988; Díaz-Azpiroz and Fernández, 2005; Fonseca and Ribeiro, 1993; Simancas et al., 2003; Pérez-Cáceres et al., 2015, and references therein). In our study area, this boundary is hidden to the SE under the infill of the Permian Viar basin (Figs. 1b and 2a), a NW–SE elongated outcrop, interpreted as a late Variscan intramontane basin with an infill made up of lower Permian detrital and volcanic rocks (Simancas, 1983; García-Navarro and Sierra, 1998). According to previous interpretations, the basin would have been inverted from the end of the early Permian to the Middle Triassic (García-Navarro and Fernández, 2004). This inversion would have produced both an open N150∘ E-oriented syncline and the NW–SE Viar reverse fault system (VFS) along its NE limit. This Permian sedimentary basin coincides with the current River Viar drainage basin (Fig. 1b) downstream of the VFS.

In our study area, the Iberian Massif records a complex Variscan deformation characterised by mainly Devonian to lower Carboniferous, NW–SE trending, folds and thrusts (e.g. Expósito et al., 2002; Mantero et al., 2011; Simancas et al., 2003). It should be noted that pervasive fractures of diverse orientation and kinematics have been reported to cut the main Variscan structures in our study area. Many of them have been attributed to the late stages of the Variscan deformation (e.g. García-Navarro and Fernández, 2004), although fractures with Cenozoic activity have been often reported close to the Iberian Massif/Guadalquivir foreland basin limit (e.g. Matas et al., 2010a; Expósito et al., 2016; Vázquez-Vílchez et al., 2015).

During the Mesozoic tectonic evolution of southern Iberia, the Iberian Massif relief was attenuated towards the SE by the formation of the rifted south Iberian paleomargin, separated from the continental Alboran Domain margin by the Tethys Ocean (Dercourt et al., 1986; Faccenna et al., 2004). The westward migration of the Alboran Domain led to the closure of this alpine oceanic basin in the early Miocene, giving rise to the formation of the Betic Chain on the southern Iberian paleomargin (Balanyá and García-Dueñas, 1987). In the middle Miocene, the Betic orogenic load provoked the lithospheric flexure of the Betic foreland (Van der Beek and Cloetingh, 1992), thus creating the Betic foreland basin (i.e. the Guadalquivir foreland basin) coupled with the forebulge just to the north of the Betic mountain front (García-Castellanos et al., 2002). The Guadalquivir foreland basin remained as a marine basin connected to the Atlantic Ocean from the middle Miocene to the beginning of the Pliocene when the asynchronous westward uplift of the basin led to the coast migration to the SW (e.g. Fernández et al., 1998; González-Delgado et al., 2004; Sierro et al., 1996). During the Quaternary, the Guadalquivir river drainage system evolved, and the upper Miocene shallowing-upward marine sediments were partially covered by alluvial deposits.

In our study area, the oldest infill of the Guadalquivir foreland basin corresponds to Tortonian calcirudites and calcarenites, which unconformably overlie the pre-Mesozoic rocks of the basement. On top of them, Messinian basinal deposits are represented by marly clays. Quaternary deposits consist of alluvial conglomerates, sands, silts, and clays related to the Guadalquivir river floodplain and to an extensive strath terrace system (Moral et al., 2013).

As previously mentioned, the southern boundary of the Iberian Massif basin also separates two distinctive relief domains, i.e. Sierra Morena to the north and the Guadalquivir river alluvial plain to the south, which is set on the most low-lying areas of the Guadalquivir foreland basin.

The main summits of Sierra Morena are 900 to 1300 m a.s.l. (above sea level) and 900–1000 m a.s.l. in the study zone. Traverse topographic profiles across Sierra Morena present a remarkable asymmetry. The relief of the main mountain ridges related to the southern Iberian sub-plateau and the Guadalquivir valley is about 300–400 and 800–900 m a.s.l., respectively. Furthermore, the northern boundary of Sierra Morena is a gently dipping transitional piedmont, whereas the southern boundary coincides with a well-defined escarpment of low sinuosity.

The NW–SE orientation of the main streams draining the southern slope of Sierra Morena is determined in their upper reaches by the Variscan orogenic grain (Fig. 1b and c). Nevertheless, the Variscan residual relief gradually disappears toward the limit with the Guadalquivir river basin and is substituted by an erosional relict surface of plausible middle–late Miocene age (Yanes et al., 2019).

At the foot of Sierra Morena, the Guadalquivir river valley is a WSW–ENE-oriented depression limited to the east and south by the Betic mountain front and to the west by the Atlantic Ocean. In our study area, it is characterised by an overall low-lying relief, with altitudes ranging between 30 and 200 m a.s.l. in the Neogene sediments and between 10 and 100 m a.s.l. in the Quaternary alluvium. Nevertheless, its easternmost sector exhibits a moderately rugged topography, with elevations that reach up to 800 m a.s.l. Furthermore, the Guadalquivir river is in a markedly asymmetric position, located at the northern edge of both its valley floor and alluvial deposits. Terraces that developed on the wider left-hand slope of the Guadalquivir river valley are gently tilted northwards (Moral et al., 2013).

As stated above, the characterisation of the Sierra Morena relief rejuvenation pattern and the identification of the structures localising such rejuvenation are crucial for our study objective. First, we have searched for signs of relief rejuvenation by applying both geomorphic qualitative methods and geomorphic indices. The purpose of this was to detect the key relief features of relief rejuvenation and the potential involved structures (mainly faults). We then proceeded to use field-based structural analysis methods to determine the kinematic regime of these faults by means of different kinematic indicators such as slickenlines or brittle S-C fabrics.

Our analysis focused on six main drainage basins showing signs of relief rejuvenation. They are located on the right slope of the Guadalquivir river valley, i.e. those draining southern Sierra Morena (Fig. 1b). Additionally, some minor sub-basins have also been included (see also Fig. 2a). For this, we used the 12 m resolution TanDEM-X digital elevation model (DEM; Wessel, 2016).

In terms of the qualitative approach, we focused our attention on those drainage network features indicative of relief rejuvenation (stream piracy, wind or water gaps, incised meanders, etc.) and on the geometry and distribution of a Miocene erosional relict surface that often occupies interfluves. Our qualitative analyses are based on the landscape morphometry, which has been characterised by means of local relief maps (difference between maximum and minimum elevation by using neighbourhood statistics; 15×15 pixels) and slope maps. We also applied several geomorphic indices, often used in low-rate tectonic regions (e.g.  Silva et al., 2003; Perez-Peña et al., 2009), to better determine the recent tectonic activity associated with the main mountain fronts. After that, we analysed the kinematics of those structures with a position and geometry that are compatible with the observed modes of relief rejuvenation.

Finally, we tested the relationship between those structures associated with the main rejuvenated mountain fronts, the distribution of knickpoints on river profiles in χ space (Perron and Royden, 2013; Schwanghart et al., 2021), and the drainage divide network migration (Scherler and Schwanghart, 2020; Schwanghart and Scherler, 2020). This analysis enables us to delve further into the relationship between the analysed faults and the current drainage dynamics. It also sheds light on the relative relief rejuvenation among the drainage basins linked to the fault activity in Sierra Morena.

As described in previous sections, the topographic distribution of Sierra Morena is regionally defined by a NNW–SSE segmentation that is primarily localised in the Sierra Morena southern slope break, but that also extends northwards, often conditioning the orientation of the secondary stream networks (e.g. Fig. 2a). Thus, this segmentation is defined not only by the topographic step at the southern limit of Sierra Morena but also by the tilting and displacement, to the north, of both the Miocene erosional paleo-surface and Miocene marine deposits (e.g. Figs. 2b and 4b, d).

We have found that this NNW–SSE segmentation is spatially related to faults that trend from SW–NE to WNW–ESE, with the former being the best represented (Fig. 6). Most of them exhibit steep surfaces that dip with either a N or S component. Their slickensides indicate slip senses that vary from the dominant normal dip slip to the dominant right-lateral slip, with the oblique, right-lateral normal faults being well represented. Although the normal slip component is dominant, a reverse component also occurs.

Some of these faults involving Paleozoic rocks seem to have been formed by the reactivation of previous structures. For example, in igneous rock outcrops of both the SPZ and the Permian Viar basin, these faults are evenly distributed with a spacing of metric order, thus pointing to the reactivation of previous joints. In the same way, slickensides indicating normal to dextral faults have been observed along the WSW–ENE Embalse del Cala fault (Figs. 2a and 3c), which is a Variscan, left-lateral structure located west of the Permian Viar basin (García-Navarro and Férnandez, 2004).

As mentioned above, this fault system is particularly developed in the Sierra Morena escarpment, defining its southern mountain front. The presence of both relict surface remnants and shallow-water Miocene rocks along the top of the escarpment enables a minimum vertical throw of the Paleozoic basement to be established. This vertical throw seems to have been greater in the east of our study area (up to 480 m north of Córdoba to the south of S6; Fig. 4b–d), gradually decreasing toward the west (≈50 m west of the Permian Viar basin to the south of S1; Fig. 2a). In the eastern sector, the vertical throw is clearly exhibited by the abrupt W–E to SW–NE escarpment located to the NW of Córdoba. At its foot, escarpment-derived colluvium of debris facies can be observed, both pre- and post-dating the Medina Azahara archaeological site (Fig. 4a), which was abandoned and vandalised soon after 1036 (López-Cuervo, 1983). In the northernmost archaeological site, the most recent colluvium (ca. 2 m thick) seals part of the ancient buildings and incorporates scarce fragments of material from Medina Azahara buildings, such as ceramic tiles and wall bricks.

In addition to this regional NNW–SSE relief segmentation, a significant WSW–ENE relief segmentation also occurs across the Permian Viar basin (Fig. 2), i.e. roughly parallel to the overall Sierra Morena escarpment trace. In the NE limit of the Permian Viar basin, this segmentation is markedly localised along the VFS, which consists of mainly NW–SE trending faults. They present moderate to steep northeastward dips, and their slickensides indicate reverse left-lateral slip (Fig. 6d).

In contrast to the NE boundary of the Permian Viar basin, relief rejuvenation in its SW boundary does not seem to be conditioned by NW–SE-oriented faults but is rather significantly distributed in diverse structures. Here, the SPZ rocks are unconformably overlain by gently northeastward Permian rocks (SW limb of the Viar synform), and signs of relief rejuvenation seem to be mostly related to the WSW–ENE fault system. The different relief rejuvenation modes observed to the SW and NE of the Permian Viar basin explain that, although HI yields similar values for both boundaries, Smf values are higher in the SW (i.e. the SPZ/Viar basin boundary). Here, differential uplift does not seem to be accommodated by a discrete structure such as the VFS but rather mainly accommodated by distributed structures of limited extent.

The geomorphological and structural dataset in previous sections suggests that the relief rejuvenation and segmentation observed in our studied area is related to two main fault systems that operated coupled with the regional flexural uplift of the forebulge area. The first one consists mainly of WSW–ENE-oriented faults, which are highly localised along the similarly oriented southern Sierra Morena mountain front. The second one is the NW–SE-oriented VFS, which limited the Permian Viar basin to the NE (Figs. 1b and 2a).

All the analysed basins are crossed in their lower reaches by one of these fault-bounded mountain fronts (Fig. 7), which are either related to the Sierra Morena slope rupture (S1, S3, S4, S5, and S6) or to the VFS (S2). Thus, the drainage network located upstream of faults may have recorded their activity through knickpoint migration.

In this regard, the distribution of the knickpoints upstream of the fault-related main mountain fronts is analysed in this section, and the distribution is compared with that predicted by our knickpoint pattern models (Schwanghart et al., 2021). This analysis will enable the comparison of recent tectonic activity along different sectors of Sierra Morena and enable the deviation between observed and predicted knickpoints to be explored in terms of divide network reorganisation.

To perform χ transformations of longitudinal river profiles, we first calculated the m/n ratio for each basin in Eq. (8) that minimises the variability in elevation values for similar values of χ (Fig. 8). The obtained values indicate that the larger catchments (S1, S2, S5, and S6) are consistent with the widely used reference value m/n=0.45, while the smaller ones (S3 and S4) show lower values. Since all considered catchments are similar concerning their fluvial erosion characteristics, a reference value m/n=0.45 was used for all the catchments in order to compare χ values and thus the locations of knickpoints across catchments.

Figure 9 shows the position on both the mainstream and tributary longitudinal profiles for basins S1 to S6. Horizontal coordinates are displayed for both the regular distance (left side) and χ space (right side). Coordinates (0, 0) correspond to the intersection between the mainstream and the fault-related mountain front. Artificial dams and tributary inlets into reservoirs are also projected along profiles. Finally, covariance estimates (ρ) for knickpoint patterns on a spatial covariate (χ transformed distance from the outlet) are reported with and without artificial dams.

The longitudinal profiles of main streams, except for S4, are partially conditioned by the presence of dams, with this being of particular note in S1, which exhibits five dams distributed over 100 km. Despite these dam-related interruptions, the S3 and S4 longitudinal profiles seem to display a slightly concave-upward shape, whereas those corresponding to S2 and S6 show a conspicuous convex-upward geometry.

The knickpoint distribution along profiles presents certain differences from one drainage basin to another. In S2, S5, and S6, most of the knickpoints are related to the middle and lower values of both x (Fig. 9c, i, and k) and χ (Fig. 8d, j, and l). Conversely, knickpoints are evenly distributed along S1 plots (Fig. 9a and b). Finally, the scarce knickpoints of S3 and S4 are located in middle and high values of the horizontal coordinates (Fig. 9e–h).

The χ space profiles offer a better perspective for discussing the knickpoint pattern dependence. Indeed, the covariance estimate (both with and without artificial dams) highlights how the majority of the higher knickpoints are in the lower part of S1, S2, S5, and S6, whereas in the upper part, several minor knickpoints form secondary peaks in the covariance trend. Knickpoints are particularly clustered in S2, where most of them are distributed in a very narrow strip corresponding to the lower part of the drainage basin, with χ values around 1000–1700 m and elevations around 200 m above the outlet point. Furthermore, S6 knickpoints are clearly grouped in the lower part of the basin, though to a lesser extent than in S2, with χ values between 2000 and 3500 m and elevations between 200 and 400 m above the outlet point.

From the comparison of the covariance computed with and without artificial dams in the dataset, it can be observed that the dependence of the knickpoint pattern is variably affected by the occurrence of artificial dams. In this regard, such a dependence is strong in S1, S3, S5, and S6, as can be inferred by the shifting of the intensity peaks when covariate values obtained with and without dams are compared. However, the dependence of the knickpoint pattern to artificial dams does not seem to be significant in S2, thus only slightly raising the intensity peak of the plot. S4 is not at all biased, since no knickpoints were automatically extracted that coincide with the only artificial dam of this drainage basin. In this regard, the plano-altimetric configuration of some dams cannot exclude the pre-existence of a knickpoint favouring the dam location. Examples of this are the artificial dam of S2 or the highest one of S6, whereby the location on the χ plot fits with the values range of the knickpoints.

As can be observed in Fig. 9, only S2 and, to some extent, S6 networks show a strong consistency between the probability of having knickpoints and the actual location of a robust number of knickpoints. Both S3 and S4 show peaks of density values above the average. However, there is considerable uncertainty surrounding these results because of the few knickpoints located in the peaks of the χ value. Instead, the intensity peaks of S1 and S5 are highly distributed and lower than the average. In detail, S1 shows values slightly higher in both the lower and in the upper parts of the basins, while S5 is affected by greater uncertainty and does not show any outstanding density peak of expected knickpoints. The notable difference between the actual and expected knickpoint patterns, especially in S1 and S5, will be analysed in Sect. 7.

Our results show that the combination of geomorphological and structural analyses of ancient areas can be a powerful tool to not only explore the role of reactivated structures on relief rejuvenation but also to delve into the tectonic setting responsible for such a reactivation. This combined analysis can be exported to any long-lived region in order to compare the proposed tectonic setting with the observed geomorphological features. In our case, we have interpreted, in tectonic terms, the main geomorphological traits of Sierra Morena, which is a flexural relief (i.e. the forebulge) generated in the Betic foreland due to the load exerted by the orogenic pile.

The analysed drainage network of Sierra Morena shows a wide range of relief rejuvenation signs such as (i) preserved relicts of a Miocene peneplain, shaped at the foreland basin base level and now located several hundreds of metres above it (this peneplain is, in places, stepped and/or tilted), (ii) incised meanders, fault-related streams deflection, and stream piracy, and (iii) patches of shallow-water marine Miocene deposits (now at different altitudes up to 400 m a.s.l). These elements, taken together, give rise to a topography of flat-lying interfluves, often constituted by either patches of Miocene rocks or remnants of the Miocene peneplain, deeply incised by the drainage network.

The qualitative geomorphological characterisation together with the use of geomorphic indices evidence the relationship between mountain fronts and Quaternary activity of both newly formed and reactivated faults. The main faults controlling relief rejuvenation belong to two different groups. The first is mainly composed of WSW–ENE, right-lateral normal faults responsible for the NNW–SSE relief segmentation of Sierra Morena and the abrupt escarpment between this range and the Guadalquivir foreland basin. The other fault group is localised on the Viar fault system (VFS), previously interpreted as a late Variscan left-lateral reverse fault system, and is responsible for WSW–ENE relief segmentation in the western sector of the study area.

Knickpoint pattern modelling, based on χ space and performed on six drainage basins, enables us to define the relationship between the activity of the analysed structures and the current geomorphological features of Sierra Morena in a more precise manner. The main stems of these selected basins are crosscut by fault segments, which are related to either the Sierra Morena southern escarpment or the reactivated VFS. The planimetric and altimetric knickpoints clustering on χ profiles confirms that the faults located in the Sierra Morena southern escarpment have experienced recent, although somehow diachronous, tectonic activity. Our analysis also evidences differences in relative uplift along this escarpment. Moreover, divide migration controls the deviation observed between modelled and observed knickpoint distribution, as indicated by the computed values of the divide asymmetry index (DAI).

On the whole, our results deviate from those expected in a typical flexural forebulge. In addition to flexural-related deformation, the kinematic features of faults related to the current Sierra Morena relief point to the Quaternary, strongly partitioned, transpressional deformation of the Betic foreland. This is compatible with the intraplate propagation of the Betic fold and thrust deformation.