We employ a pull- and push-type apparatus with a computer-controlled moving wall, a horizontal (slope = 0∘) film-faced plywood basal board, and 12 mm security-grade side glass walls (Fig. 2) (Gottron, 2018). We use only granular materials with brittle behaviour to simulate upper-crustal deformation. The overall model dimensions are 40 × 60 × 3 cm (width, length, and height, respectively). Experiments that incorporate extension are carried out in two stages. For the initial extensional stage, removable basal sheets are attached to the moving wall, following a technique employed in previous analogue modelling studies (e.g. Panien et al., 2006; Granado et al., 2017; Jara et al., 2018; Yagupsky et al., 2008; Likerman et al., 2013; López et al., 2022). The basal plastic sheets are 0.4 mm thick each, and their borders are cut perpendicular to the moving wall (Fig. 2c, left panel). Up to three basal sheets can be stacked and attached independently to the moving wall (Fig. 2c, right panel). The wall is moved backward (pull) at a constant speed of 1 mm min-1 for 3 cm, pulling the basal sheets with it to create an extensional basin. The differential relative movement between the uppermost basal sheet(s) and the apparatus basal board or the underlying basal sheet(s) creates a velocity discontinuity where normal faults are initiated. During the extensional stage (3 cm of backward displacement), the wall motion is paused every 1 cm to fill the interior of the depocentres to simulate syn-tectonic sedimentation (e.g. Panien et al., 2006; Jara et al., 2018). The filling of the grabens is done by sifting microbeads from a height of ∼ 15 cm with a smaller sifting container in order to avoid sedimentation in the rift shoulders. At the end of extensional basin formation, the basal sheet is detached from the moving wall and fixed to the basal board to prevent its motion in subsequent extensional stages.

Since we evaluate the control that multiple extensional basins have on the evolution of orogenic wedges, we systematically vary the number of basins that are incorporated into the models. The number of basins created in the models is defined by how many basal sheets are employed, and how many times the extensional basin procedure previously described is repeated (Fig. 2d). In all models, the final extensional stage is followed by a shortening stage, imposed by the forward movement of the wall (push). Both stages of deformation are carried out at a velocity of 1 mm min-1. Environmental conditions in the laboratory are controlled and kept within a range of 22 ± 1 ∘C for temperature and 45 ± 5 % for relative humidity for all experiments. Deformation is monitored using four digital single-lens reflex (DSLR) cameras that are simultaneously triggered, and pictures are taken every 30 s. Surface and lateral strain, as well as topography, is then quantified using a high-resolution particle image velocimetry (PIV) and digital photogrammetry monitoring system (LaVision DaVis^®). We calculate the incremental strain at each time increment, and progressive addition of these values is considered as cumulative values.

We use dry quartz sand to simulate crustal materials and microbeads to simulate weaker detachment layers at the base of the sand pack. Sedimentary infill is also carried out with microbeads, coloured for visualisation purposes, to simulate less competent sedimentary rocks (e.g. Panien et al., 2005). By using an approximate proportion of 0.1 g of dry powder pigment per 1000 g of microbeads, the mix only differs in colour but not in mechanical properties in comparison to the uncoloured microbeads used for the detachment layer. Both sand and microbeads obey the Mohr–Coulomb criterion (Coulomb, 1773; Krantz, 1991; Vermeer, 1990; Teixell and Koyi, 2003), where a strain-rate-independent rheological behaviour can be assumed. We performed shear box tests to obtain values for angle of internal friction and cohesion. Our values are similar to previously published results for the same materials (Klinkmüller et al., 2016). The employed quartz sand, prepared by Schlingmeier Quarzsand GmbH (Type G12T), has a bulk density of 1.53 g cm-3, grain size range between 0.1 and 0.4 mm, average grain size of 0.24 mm, an angle of internal peak friction of 33∘, angle of reactivation of 30∘, and a cohesion of 58 Pa. The microbeads, manufactured by Kuhmichel Abrasiv GmbH, have a bulk density of 1.51 g cm-3, grain size range between 100 and 200 µm, average grain size of 174 µm, an angle of internal peak friction of 28∘, angle of reactivation of 24∘, and a cohesion of 28 Pa. The microbead–plywood board interface friction angle is 19∘. Microbeads used for the basal layer and quartz sand are sifted using a purpose-built sifting device based on Maillot (2013), from a fixed height of ∼ 40 cm from the base of the apparatus. Analogue experiments must be scaled in terms of length, time, and forces to appropriately reproduce natural scenarios. We follow Hubbert (1937) and Ramberg (1981) to scale stress in materials with brittle behaviour, a standard approach in analogue modelling studies, following the relationship given by σ∗=ρ∗⋅g∗⋅l∗, where σ, ρ, g, and l are stress, density, gravity, and length, respectively. Asterisks indicate they are scaling factors, which are calculated as the ratio of model (m) to nature (n): 1σ∗=σmσn=ρmρn⋅gmgn⋅lmln. Our experiments are performed in a normal gravity field; therefore gm/gn=g*=1. We obtain our density scaling ratio by considering our quartz sand bulk density of ∼ 1530 kg m-3 and crustal rock densities of ∼ 2700 kg m-3 to obtain ρm/ρn=ρ*=0.567, and we fix a length relation of 1 cm = 5 km in order to obtain a length scaling ratio lm/ln=l*=2 × 10-6, which is appropriate for properly modelling scaled material strength using quartz sand (Ritter et al., 2016). The stress ratio is then (Eq. 2) 2σ∗=1.13×10-6, meaning that a cohesion of 58 Pa for quartz sand in the model corresponds to ∼ 51 MPa in nature, and 28 Pa for microbeads corresponds to ∼ 25 MPa in nature, falling within the range of ∼ 20–110 MPa calculated with uniaxial compressive strength tests on natural rocks (Jaeger and Cook, 1976; Twiss and Moores, 1992; Handin, 1969). The angle of internal friction is dimensionless; therefore, our model materials represent an appropriate analogue as 33∘ for quartz sand and 28∘ for microbeads fall within the range derived from experimental rock tests (Jaeger and Cook, 1976). Our models are not dynamically scaled since the rheology of granular materials is strain-rate-independent. General scaling parameters are summarised in Table 1.

Analogue modelling has inherent limitations, mainly related to the difficulty of reproducing complex rheologies and simulating realistic temperature evolution of rocks, therefore representing a simplification of nature (Koyi, 1997). In addition to this simplification, we have identified a number of limitations in our models. Sidewall friction in a push-type apparatus like ours is inevitable, and we therefore treat the internal side of the glass walls with a water-repellent product (Rain X^®) as it reduces the sand–glass interface friction (Cubas et al., 2013). Since the side contact area is relatively small due to the thin layering, the unwanted side effect can be considered negligible for this series of experiments. This is supported by a consideration of the ratio between the top surface area and the side surface area in our experiments, which is 0.075 and thus below the value of 0.1, which Souloumiac et al. (2012) determined as the upper limit for negligible sidewall effects. The apparatus was designed to prevent material from sliding through the gap between the moving wall and the basal board during shortening. However, a very thin layer of microbeads still passes beneath the wall at early stages of deformation. As soon as crustal rupture is achieved, the entire thrust sheet is translated by the moving wall, and no extra material slides through the gap. Although we recognise that the deformation imposed in our experiments strictly creates shear zones, we refer for ease of comparison to natural examples to areas of material failure in this study as faults and/or thrusts.

In terms of orientation of principal stresses in the upper crust, basin inversion can be generally described as a change from an extensional setting, where σ1 is vertical, and σ3 is horizontal, to a compressional setting with σ1 horizontal and σ3 vertical. Assuming a Mohr–Coulomb criterion of failure for upper-crustal rocks, the dip angle at which normal and reverse faults should form can be calculated when the cohesion and angle-of-internal-friction values are known. Considering a rock analogue to the quartz sand employed in our experiments with an angle of internal friction of 33∘, extension would result in the formation of normal faults at ∼ 66∘ from the horizontal, while compression would lead to new thrusts forming at 33∘ (Fig. 3a). The resulting unfavourable orientation of normal faults to reactivate under compressive stress has been the subject of numerous studies (i.e. Jaeger and Cook, 1976; Cooper and Williams, 1989; Bonini et al., 2012; Zwaan et al., 2022), and it is the main reason why complete fault reactivation remains a challenging process to recreate by means of analogue modelling that employ materials with a simplified rheology in relation to nature. Analogous set-ups in numerical models have successfully simulated the reactivation of misoriented extensional faults by compressive stresses (e.g. Granado and Ruh, 2019). Despite the limitations of analogue materials, certain conditions such as obliquity of fault strike to σ1 (Del Ventisette et al., 2006), weaker basin infill (Panien et al., 2005), horizontal stress direction rotation (Ritter et al., 2016; Bonini et al., 2012), and type of post-rift sediments (Granado et al., 2017) have been tested in laboratory experiments and demonstrated to promote basin inversion and/or partial fault reactivation.

In this study we present various models with a different arrangement of internal mechanical heterogeneities prior to the onset of shortening: a reference model with a homogeneous crust and models with one, two, and three extensional basins (Fig. 2e). Considering the materials' density, cohesion, and angle of internal friction, Mohr diagrams representing the state of stress at the base of different sections of the models at the onset of shortening are plotted in Fig. 3b and c. With this theoretical framework (Fig. 3) as a basis, the following preliminary hypotheses can be drawn:

Under conditions where layers have the same thickness and similar densities, the lower angle of internal friction of the microbeads will make the model sedimentary infill more prone to failure with the onset of shortening. Thrusts are expected to propagate through the weaker extensional basins (Fig. 3b–c).

During shortening, the extensional basins are incorporated in a thrust wedge, which has heterogeneous material properties caused by the mixing of the microbeads of the basin fill with quartz sand. Critical taper theory (Dahlen, 1984, 1990) predicts a lower taper angle of 6.2∘ (unsoftened) to 7.2∘ (softened) for a pure quartz sand wedge overlying a thin microbead basal layer. Theoretical dip values of forward-verging thrusts are ∼ 20∘, and steeper back thrusts dip at ∼ 37∘ (Dahlen, 1990). The critical taper angle (which equals surface slope for our horizontal base experiments) for a pure microbead wedge is 7.9∘ (unsoftened) to 9.7∘ (softened). For our wedges, surface slopes between 6.2 and 9.7∘ can thus be argued to be valid values, where we would expect a tendency towards the higher values within the lower taper angle due to the mixing of microbeads into the sand (Buiter, 2012).

As thrust sheets are emplaced over the extensional basins during shortening the resulting strength increase caused by the increase in σ3 may be significant enough to keep the system stresses in a state such that the localisation of new thrusts would be determined by not only the internal rheological heterogeneities but also the need of the wedge to reach a critical taper value. New frontal thrusts then develop to decrease the taper angle.

Experiment REFNB (reference experiment, no basal sheets) consists of a 5 mm thick microbead layer overlain by a 25 mm thick cover of quartz sand, with interbedded coloured horizontal layers that act as passive markers. No basal plastic sheets are employed in this model. Results of experiment REFNB show a typical piggyback deformation style, with thrust sheets developing in sequence and being gradually transported towards the foreland. Shortening is accommodated first by a pop-up structure consisting of a forward and back thrust stemming from the basal detachment layer 2.3 cm from the moving wall (T1 in Fig. 4a). Activity in the back thrust decreases, and strain is then only accommodated for by the frontal thrust T1 until a new pop-up structure forms, 12.9 cm from the moving wall (T2 in Fig. 4a). The entire thrust sheet is translated towards the foreland until a third pop-up structure develops, after 16 cm of displacement (T3 in Fig. 4a). After 20 cm of displacement, the measured wedge taper angle is 8.5∘ (Fig. 4a).

Experiment REFWB (reference experiment, with basal sheets) was designed to create the exact same initial conditions as experiment REFNB, with the only addition of including plastic basal sheets fixed to the apparatus base, with vertical sub-millimetric steps – perpendicular to the shortening direction – located 9.5, 19.5, and 24.5 cm from the initial position of the moving wall (Fig. 4b). The plastic sheets are used in subsequent experiments to create extensional basins and remain in a passive role during the shortening phase. We test here their potential influence during shortening. Experiment REFWB develops a pop-up structure almost immediately after the initiation of the wall movement (T1 in Fig. 4b), as observed in experiment REFNB. However, in REFWB the pop-up structure forms 7 cm from the moving wall, which is closer to the edge of the first basal sheet step, meaning that its associated basal velocity discontinuity could play a role in determining the location of the first thrust sheet in the accretionary wedge evolution. After approximately 7.5 cm of displacement, the initial pop-up structure is progressively abandoned, and a new frontal thrust forms towards the foreland, stemming from the basal detachment layer 14 cm from the moving wall (T2 in Fig. 4b). In this case, the location of the frontal thrust and its conjugate back thrust is not aligned with one of the basal sheet steps (Fig. 4b). Most of the shortening is accommodated by the second frontal thrust until the system reaches ∼ 17 cm of displacement. At this point, a new pop-up structure develops, stemming from the basal detachment layer 20.5 cm from the moving wall (T3 in Fig. 4b). The new frontal thrust localises most of the wedge deformation until the final stages of the experiment (Fig. 4b). After 20 cm of displacement, the measured wedge taper angle is 9.3∘ (Fig. 4b)

The overall evolution of the accretionary wedge in REFWB is similar to experiment REFNB. Both experiments were run twice, showing little variability in terms of pop-up structure formation and thrust localisation. The velocity discontinuities formed by the sub-millimetric geometrical steps of the basal sheets only influence the model in early stages, defining the onset of the first thrusts. However, the fact that the sheet edges are not aligned with the subsequent floor thrusts allows us to assume that their effect in the evolution of the modelled accretionary wedge is negligible for the purpose of our study.

Experiment 1B consists of two stages. An initial stage of 3 cm extension in which an extensional basin with border faults located between 24.5 and 27.5 cm from the wall (measured along the base) is created, followed by a stage of 20 cm of compressional displacement. Shortly after the initiation of shortening, frontal and back thrusts develop approximately 5 cm from the moving wall (T1 in Fig. 5). During the first 6 mm of displacement, before the onset of the first frontal thrust, the distal extensional basin shows internal horizontal compression, reaching peak values of up to 16 % of cumulative strain (7 % on average) but without fault reactivation or the formation of new thrusts. No further internal deformation of the basin is observed between 0.6 and 4 cm shortening.

The initial thrust sheet is transported towards the foreland, and after 5 cm of total displacement, strain accommodation is partitioned, and stress builds up again in the extensional basin. Subsequently, the initial forward thrust is progressively abandoned, and strain localisation is transferred entirely near the pre-existing extensional basin (T2 in Fig. 5). A new forward thrust forms along the limit between the faulted blocks in the foreland side and the weak microbead depocentre filling, without reactivating the extensional border faults. The extensional faults on the hinterland side are cross-cut, and the thrust sheet propagates towards the foreland, above the pre-rift sediments (Fig. 5; 8 cm displacement). With ongoing deformation, no further deformation is observed beyond the extensional basin, and the wedge grows in height through the development of three new back thrusts that form progressively towards the location of the wall. After 20 cm of displacement, a wedge taper angle cannot be accurately estimated, as the wedge is in transition to a critical taper, and two distinctive wedge fronts can be established at 27.7∘ (determined by the angle of repose of the microbeads) and 3.7∘ in the more planar, higher section (Fig. 5).

The 3 cm extensional phase for experiment 2B leads to the formation of a distal basin with border faults located between 24.5 and 27.5 cm from the wall (measured along the base) and a central basin with border faults located between 19.5 and 22.5 cm from the wall. Initial compressional deformation is accommodated by the development of a pop-up structure, forming 14.3 cm from the wall, aligned with the pre-existing extensional basin located closer to the hinterland (T1 in Fig. 6). The frontal thrust cross-cuts the border faults on the hinterland side and progressively transports the basin fill over the pre-rift sand pack, towards the foreland. After 2.3 cm of displacement, a new forward thrust develops in the proximity of the wall (T2 in Fig. 6). As observed in experiment 1B, the early development of a thrust sheet located relatively farther away from the wall than in the reference models causes the wedge to grow in a subcritical, unstable manner (Fig. 6; 4 cm displacement). Internal deformation of the wedge is therefore controlled by a new forward thrust forming 2.4 cm from the moving wall (T2). This out-of-sequence thrust accommodates shortening for another 2.5 cm of displacement, when the frontal thrust T1, located along the weak graben fill, is reactivated (Fig. 6; 8 cm displacement). PIV analysis at 8 cm of displacement indicates no normal fault reactivation, but the frontal thrust does show a strong deflection of its dip angle, from 20∘ at the base to 37∘ when it interacts with the extensional basin (Fig. 6; 8 cm displacement). Coevally, a series of back thrusts become active on the hinterland side of the thrust sheet as the wedge continues to advance. At later stages, the frontal thrust located along the central basin and its conjugate back thrust decrease their activity, and strain localisation is progressively transferred to the distal basin (Fig. 6; 12 cm displacement). Before the onset of a new pop-up structure in the distal basin, cumulative strain within the syn-rift strata reached values of up to 24 %. After 12 cm of total displacement, shortening in the foreland is accommodated first by a back thrust aligned with the extensional basin border faults, later replaced by the formation of a new forward thrust (T3 in Fig. 6). At 16 cm of displacement, the wedge continues to grow, with most of the strain localised along the two forward thrusts and steeply dipping back thrusts (Fig. 6; 16 cm displacement). The final stages of deformation are characterised by a new pop-up structure forming in the foreland (T4 in Fig. 6). The wedge taper angle after 20 cm of shortening is 9.7∘ (Fig. 6).

Experiment 3B is carried out following the same procedure as experiments 1B and 2B, but now creating three extensional basins before the onset of shortening. The extensional phase for this experiment results in the first basin located between 28.5 and 31.5 cm from the moving wall, the second basin between 19.5 and 22.5 cm from the moving wall, and the third and last basin located between 9.5 and 12.5 cm from the moving wall (all measured along the base). Early compressional stages are characterised by the formation of a forward thrust, with its location being controlled by the weak basin fill of the basin, cross-cutting the extensional border faults located on the side of the moving wall (T1 in Fig. 7). This forward thrust is followed by the formation of a back thrust, both stemming from the basal detachment layer 5.8 cm from the moving wall. Close inspection of PIV data shows that the interior of the proximal extensional basin localises up to 20 % of strain before the onset of the pop-up structure. After 6.5 cm of displacement, strain is partitioned and progressively transferred to the central extensional basin, where a new forward thrust develops, aligned with the interface between the rifted blocks and the weak sedimentary infill, as occurred in earlier stages of this experiment (T2 in Fig. 7). No clear reactivation of border faults can be detected in this process, although the high angle (49∘) of the associated back thrust could be associated with the partial reactivation of the hinterland border faults. The central basin shows horizontal compression and a partial uplift of 2.2 mm prior to crustal failure. Cumulative strain is calculated to reach values of up to 26 % within the basin before the onset of new thrusts.

Strain localisation along this central basin is short-lived, and a new forward thrust develops 16.2 cm from the moving wall, with its localisation being determined by the existence of the distal basin (T3 in Fig. 7). The new frontal thrust becomes the most active structure at this stage of the accretionary wedge evolution. Two new back thrusts form and show a noticeable deflection in their dipping angle when interacting with the uplifted central basin. After 16 cm of displacement, the entire sedimentary fill of the first basin is transported past the frontal thrust tip point. At this stage, internal deformation of the wedge is accommodated by the frontal thrust T3 and its conjugate back thrust, as well as by a forward thrust stemming from the base of the wall and aligned with the location of the weak remobilised sediments in the interior of the wedge (Fig. 7; 16 cm displacement). After 20 cm of shortening, strain is localised only along the frontal thrust and a back thrust that transects the entire wedge, with a variable dip angle of 39∘ in its lower section and 60∘ in its upper section (Fig. 7; 20 cm displacement). The measured wedge taper angle is averaged to 10.7∘, although two wedge fronts with different slopes can alternatively be measured as 28 and 5.5∘ (Fig. 7).

Certain features can be observed across all experiments with pre-existing extensional basins. Independent of the position of the graben (proximal, central, or distal), detailed analysis shows that normal fault reactivation sensu stricto is not the main mechanism for basin inversion during shortening stages. The change from extension to shortening is instead characterised by a partial inversion that can be outlined in three stages:

Horizontal stresses are transferred across the quartz sand (analogous to stronger crust), resulting in the shortening of the microbeads (analogous to weak sedimentary infill) and an associated upward bulging. Cumulative strain localising in the interior of the basins prior to the development of forward thrusts ranges from 5 %–26 %. When multiple extensional basins are present, the proximal basin absorbs most of the stress transferred through the quartz sand as it is closest to the wedge.

EXP1B shows the largest thrust spacing among the experiments, with the first pop-up structure developing in the vicinity of the moving wall (Fig. 5; 2 cm displacement), followed by the transfer of horizontal stresses onto the pre-existing extensional basin, located ∼ 17 cm beyond the basal detachment of the initial frontal thrust (Fig. 5; 8 cm displacement). This experiment shows the maximum amount of strain accumulation in the extensional basin prior to its failure by shortening and represents the distal extensional basin with the earliest evidence of a certain degree of inversion. The emplacement of frontal thrusts differs from the reference experiments as in EXP1B they are conditioned by the location of the extensional basin (cf. Figs. 4 and 5). The distal extensional basin acts as an efficient localiser that takes up stress, and no thrusts develop beyond its position throughout the entire experiment.

EXP2B is the only experiment of this series that developed out-of-sequence frontal thrusts. The first frontal thrust localised along the central extensional basin, before strain was accommodated by a pop-up structure in the vicinity of the wall (Fig. 6; 4 cm displacement). Thrust spacing closely resembles that of the reference experiments, although they follow a different chronological order. Contrary to what is observed in the other experiments, the distal graben normal faults that are cross-cut are those located at the foreland side of the basin, resulting in an oppositely verging uplift of the sedimentary infill (Fig. 6; 16 cm displacement)

EXP3B evolution is characterised by in-sequence thrusts that localise along each of the extensional basins (Fig. 7). The main difference with EXP1B and EXP2B is that the inversion of the central basin is more complex: early stages of deformation produce partial uplift, but it is short-lived as prograding sediments from the wedge result in an overburden that makes the basin rheologically stronger, and deformation is soon transferred to the distal basin (Fig. 7; 12 cm displacement). As in EXP1B, no further shortening structures are observed beyond the position of the distal basin (Fig. 7; 20 cm displacement).

Fault dip angles in the analogue experiments average 26–32∘ for the forward thrusts and 31–46∘ for the back thrusts. The forward thrusts are thus somewhat shallower than the theoretical dip angle of 33∘ for reverse faults in undeformed sand, whereas the back thrusts are mainly steeper than this value. For mature wedges, the theoretical dip angles become 20∘ for forward thrusts and ∼ 37∘ for back thrusts (Dahlen, 1990; Sect. 3.2 herein). The measured forward thrusts are thus steeper than the theoretical value, whereas the back thrusts could be argued to fit fairly well. However, in our opinion the thrust dip measurements have limited validity as they are a manifestation of internal wedge deformation during its transition to become critical.

When comparing the experiments in which basins are incorporated in the thrust wedge to the reference experiments without basins, we find a perceptible effect of the pre-existing extensional basins on the final taper angle of the wedge: the taper angle increases. The reference experiments with a homogeneous quartz sand layer had measured taper angles of 8.5 and 9.3∘ after 20 cm of shortening, which are somewhat higher than the predicted values of 6.2–7.2∘. Measured values for experiments 2B and 3B are higher again, at 9.7 and 10.7∘, respectively (Fig. 8). In experiments 1B and 3B the shape of the advancing wedge is irregular, making it harder to draw a best-fit line that determines the taper angle (Fig. 8c and d). In these cases, the protracted localisation of shortening along the distal extensional basin results in greater accumulation of scraped-off and slope materials in the frontal section. As a result, the frontal slope is determined by the angle of repose of the mix of quartz sand and microbeads. When comparing experiments 1B, 2B, and 3B after the same amount of shortening (Fig. 8), experiment 2B displays the lowest wedge taper angle, as it formed a new frontal thrust at a slightly earlier stage of deformation (Fig. 8c–d). However, the overall temporal evolution of the wedge taper angle during shortening is similar for the experiments with a previously extended crust. Results point towards an increase in the taper angle in comparison to wedges with a homogenous crust (cf. Fig. 8a–b and c–d). This agrees with our predictions (Sect. 3.2) as the inclusion of the weaker microbeads of the basin fill is expected to increase the taper angle.

In terms of local structures and transient mechanisms of strain accommodation, we can classify partial inversion in the models into three stages: basin shortening and uplift, thrust localisation and reactivation, and deformation relocalisation due to changes in brittle strength.

Our modelling results can be compared to case studies of fold-and-thrust belts forming over previously extended crust. While aspects such as crustal composition, type of syn-rift infill, shortening direction and its variation over time, and distances or thicknesses may differ, we argue that our findings can be extrapolated from modelling to nature in order to provide first-order insights on the evolution of deformation and basin inversion in fold-and-thrust belts. Here we compare the most relevant aspects of our models with natural examples across multiple length scales.