40Ar/39Ar muscovite geochronology of Arperos Basin inversion in Southern Mexico: new insights into cretaceous shortening initiation in southernmost North America

ABSTRACT During Cretaceous–Eocene time, North America was involved in a succession of shortening events that produced major orogenic belts. Previous work permitted the characterization of the kinematics and dynamics of some of these orogenic events along various transects. However, the timing of shortening is not well constrained everywhere. This has led to misinterpretation of the tectonic history of some sectors of the North American Pacific margin and to an incomplete understanding of the subduction dynamics. Mexico is one of those sectors in which the shortening history is not completely defined. We present structural and 40Ar/39Ar isotopic data from major shortening structures in southern Mexico, documenting two episodes of shortening: a first one at ~ 118–112 Ma, not considered in previous works, and a second one at ~ 104–90 Ma, which was previously interpreted as the record of shortening initiation. Integrated with previous data, our 40Ar/39Ar results indicate that, in Mexico, the shortening history was developed in two main stages: 1) a Late Aptian–Early Albian stage of localized shortening, during which the oceanic Arperos Basin was closed and the Guerrero terrane arc accreted to nuclear Mexico forming a ~ 80 km-wide suture belt, and 2) a latest Albian–Eocene stage, in which shortening was propagated from the North American trench to the plate interior, producing the development of a critically tapered wedge. Whether these two stages represent two superposed orogenic cycles or two steps within the same orogenic cycle remains to be established. In any case, the Late Aptian–Early Albian stage of deformation, which was not fully recognized in most previous works, is a key for our understanding of the causes that triggered shortening in southern North America after ~ 100 m.y. of extension associated with Pangea break-up GRAPHICAL ABSTRACT


Introduction
Between Cretaceous and Eocene time, the tectonic evolution of western North America was punctuated by several regional-scale shortening events (e.g. the Sevier and Laramide orogenies in the USA and the Mexican orogeny in Mexico; Yonkee and Weil 2015;Fitz-Díaz et al. 2018;Hildebrand and Whalen 2021).These events followed one another in close succession, giving rise to major orogenic belts (Figure 1) and extensive foreland basin systems, which shaped the North American Pacific margin (Dickinson and Lawton 2001), fundamentally changed the environmental conditions, and influenced the floral and faunal distribution and diversification (Suarez et al. 2014).In addition, these orogenic events favoured the formation of extensive ore deposits (e.g. the Laramide orogenic gold trend; Goldfarb et al. 2008;Izaguirre et al. 2017), and were instrumental in the formation of hydrocarbon reservoirs in foreland clastic wedges (e.g. in the circum-Gulf-of-Mexico area; Lawton et al. 2009;Cossey et al. 2020).Based on this premise, reconstructing the kinematics and dynamics of these shortening events is vital for our understanding of the tectonic evolution of North America and the origin of its biodiversity, as well as for a more successful exploration of natural resources.In the last decades, the effort of many authors who studied various transects across the North American Pacific margin, from Alaska to Mexico, permitted the regional characterization of the kinematics and, in some cases, the dynamics of some of these orogenic events (Bird 1998;DeCelles and Coogan 2006;Fitz-Díaz et al. 2018;Hildebrand and Whalen 2021).However, the timing of shortening is not yet exhaustively constrained everywhere.Since some of these shortening events occurred very close in time in the same area, they are often confused, and the structures associated with one event are erroneously assigned to another.According to previous works, these shortening events are likely the result of different tectonic processes that include accretion of juvenile terranes (Coney et al. 1980;Dickinson 2009), variation in convergence rates, and changes in subduction direction (Bird 1998;Saleeby 2003;English and Johnston 2004).Only the precise age definition of the shortening structures permits an exhaustive idea of the timing, duration, and magnitude of these tectonic processes along the North American margin and improves our understanding of the subduction dynamics.
The Triassic-Lower Cretaceous stratigraphic record of Mexico is dominantly composed of fluvial to marine successions formed during rifting of western equatorial Pangea (Goldhammer 1999;Martini and Ortega-Gutiérrez, 2018).These successions were successively folded and sheared, forming a regional-scale foldthrust belt with a present-day impressive topographic expression in the Sierra Madre Oriental (Fitz-Díaz et al. 2018; Figure 1).Previous works along the most iconic cross-sections of the fold-thrust belt (Suter 1987;Carrillo-Martínez 1989;Chávez-Cabello et al. 2011;Cuellar-Cardenas et al. 2012;Fitz-Díaz et al. 2018;Cid-Villegas et al. 2022) permitted major progress in the understanding of its architecture and, in some cases, its dynamics, especially in northern and central Mexico.However, there are still some points of disagreement regarding the Mesozoic-Cenozoic shortening history of Mexico (Fitz-Díaz et al. 2018;Hildebrand and Whalen 2021).Some major points that deserve to be clarified are: 1) what is the age of shortening initiation in Mexico; 2) whether shortening initiation was coeval along the Mexican Pacific margin or not; and 3) whether shortening structures in Mexico are the manifestation of one single orogenic event or the superposition of different and successive orogenic events associated with different driving forces.
In this work, we present new 40 Ar/ 39 Ar ages from major shortening structures in the Santo Tomás area of southern Mexico (Figure 1).Our data provide the evidence that rocks in the study area experienced two superposed shortening events and document a major late Early Cretaceous folding and thrusting event that was not previously recognized in southern Mexico.The integration of our new ages with previous data indicates that this late Early Cretaceous event is the oldest, regional-scale shortening event that occurred in Mexico after a ~ 100 Ma-long period of extension associated with Pangea break-up.For that reason, this event is crucial to understand the cause of shortening initiation along the Pacific margin of southern North America.

Geological background
The Cretaceous-Eocene shortening structures in Mexico Guzman and de Cserna (1963) are the first authors that analysed the significance of Upper Cretaceous-Eocene shortening structures of Mexico in a regional tectonic context.These authors proposed that these structures are the expression of a major orogenic event named the Hidalgoan orogeny, which was interpreted as the southern extent of the Laramide orogeny in the USA (Figure 1).This idea generated traction within the Mexican geological community and was adopted by most authors during the following decades (Campa and Coney 1983;Suter 1984;Carrillo-Martínez 1989).A few years later, Campa (1985) introduced the name Mexican fold belt to designate the Upper Cretaceous-Eocene shortening structures.By the end of the '80, the reconstruction of the shortening history was improved by stratigraphic and structural studies of the Guerrero terrane suture belt in central Mexico (Figure 1).The suture belt is composed of the inverted Arperos Basin, which is tectonically overlain by the Guerrero arc (Chiodi et al. 1988; Lapierre  Yonkee and Weil (2015) and Fitz-Díaz et al. (2018.)et al. 1992;Ortíz-Hernandez et al. 1992; Figure 2a).Most authors (Cabral-Cano et al. 2000;Martini et al. 2011;Boschman et al. 2017;Centeno-García 2017) interpreted the Arperos Basin as a Late Jurassic-Early Cretaceous back-arc trough developed between the Mexican continental mainland and the Late Jurassic-Early Cretaceous, west-facing, extensional Guerrero arc, presently exposed in the Guerrero terrane of western Mexico (Figure 1,b).The maximum width of the Arperos Basin prior to its collapse and inversion is unknown.Considering that oceanic spreading within the Arperos Basin lasted for a maximum of ∼18 Ma (e.g.Martini et al. 2014), and taking into account that modern back-arc spreading rates range from 95 (e.g.Lau Basin; Taylor and Martinez 2003) to 11 mm/yr (Mariana basin; Deschamps and Fujiwara 2003), the amount of oceanic crust formed in the Arperos Basin could have ranged from ∼ 1700 to ∼ 200 km.Chiodi et al. (1988) and Quintero-Legorreta (1992) recognized that the Arperos Basin and Guerrero arc in central Mexico display two superposed sets of folds, whereas the unconformably overlying Upper Albian, neritic limestone of the La Perlita Formation shows only one set of open folds.Based on these observations, these authors suggested that an Early Cretaceous shortening event produced the closure of the Arperos Basin and accretion of Guerrero arc to the Mexican mainland prior to the Late Cretaceous-Eocene Laramide orogenic event.Some authors (Tardy et al. 1994;Dickinson and Lawton 2001) accepted this scenario and proposed that the various shortening pulses that affected Mexico during Mesozoic and Cenozoic time are likely the manifestation of variation in convergence rates and changes in subduction directions along the North American trench (Martini et al. 2016).On the other hand, other authors that studied the Arperos Basin in southern Mexico (Salinas-Prieto et al. 2000;Talavera-Mendoza et al. 2007) did not recognize the field and stratigraphic evidence of Early Cretaceous inversion for the Arperos Basin and supported the idea that shortening structures in Mexico are the expression of a Late Cretaceous-Eocene, single orogenic event.
In the last few years, the importance of a pre-Late Cretaceous shortening event has been resumed by the seminal paper of Fitz-Díaz et al. (2018).In their review, Fitz-Díaz et al. (2018) recognized that the oldest manifestation of shortening in Mexico is represented by the closure of the Arperos Basin in Albian time, which was followed by the development of a critically tapered wedge that extends from the present-day Pacific margin to the Gulf of Mexico.According to Fitz-Díaz et al. (2014), this wedge was developed during episodic pulses of deformation between at ~ 93-80 Ma, ~ 75-64 Ma, and ~ 55-43 Ma.These authors grouped the Guerrero terrane suture belt and the critically tapered wedge into a single orogenic belt named the Mexican orogen (Figure 1).They also recommended that the Mexican orogen should be considered as a separate structural entity, distinct from its northern counterpart in the USA, because it has structural characteristics and potential tectonic drivers that are different from the Laramide orogen.

The Arperos Basin
The Arperos Basin crops out discontinuously in Mexico, forming a ~ NNW-trending belt subparallel to the North American Pacific margin (Figure 3a).The type area of the Arperos Basin is the Sierra de Guanajuato in central Mexico (Tardy et al. 1994; Figure 3b).Other deepmarine successions that have been correlated with the Arperos Basin crop out in the Zacatecas area, ~ 200 km to the north (Ortega-Flores et al. 2016;Figure 3c), and between Santo Tomás and Arcelia, in southern Mexico (Freydier et al. 1996;Martini et al. 2014; Figure 3d).In these areas, the Arperos Basin is arranged in a tectonic pile of thrust sheets that are overthrust by the Guerrero terrane arc with top-to-the-SE to -NE transport directions (Elías-Herrera et al. 2000;Fitz-Díaz et al. 2008;Martini et al. 2013).In all these areas, the Arperos Basin displays an asymmetric depositional architecture, with volcaniclastic metaturbidites sourced by the Guerrero terrane arc on its western side and siliciclastic to calcareous metaturbidites derived from the Mexican mainland on the eastern side (Martini et al. 2011(Martini et al. , 2014;;Ortega-Flores et al. 2016).Sedimentation within the basin was accompanied by magmatic activity, which was initially dominated by peraluminous daciticrhyolitic volcanic and subvolcanic rocks hosting volcanogenic massive sulphide deposits and, in a later stage, intraplate-like and middle ocean ridge basalts (Ortíz-Hernandez et al. 1992;Freydier et al. 1996;Martini et al. 2014).Closure of the Arperos Basin is solidly bracketed in the Sierra de Guanajuato between ~ 119 Ma, the age of the youngest group of detrital zircon grains in volcaniclastic metaturbidites (Martini et al. 2011), and Late Aptian time (~ 108 Ma; Gradstein et al. 2012), the palaeontologic age of the La Perlita Formation limestone, which unconformably overlies the inverted basin (Chiodi et al. 1988).After basin inversion, a second shortening event produced a set of folds and top-to-the-NE to -SW shear zones, which were developed on previously deformed successions and the overlying Upper Aptian limestone as well.
Like in the Sierra de Guanajuato, the Arperos Basin successions in southern Mexico were affected by two main shortening events, the oldest of which has been related with basin closure and Guerrero terrane accretion (Salinas-Prieto et al. 2000;Talavera-Mendoza et al. 2007;Martini et al. 2014)  40 Ar/ 39 Ar whole-rock ages between ~ 105 and ~93 Ma from metabasaltic rocks and extended the age of the Arperos Basin into Cenomanian time.Based on these data, the Arperos Basin inversion in southern Mexico has been considered to have occurred in Turonian time or later (Mendoza and Suástegui 2000;Fitz-Díaz et al. 2008).It is, however, important to point out that some authors (Cabral-Cano et al. 2000;Martini et al. 2014) highlighted the importance of Late Cretaceous tectono-thermal events, which locally reset, partly to totally, K/Ar and 40 Ar/ 39 Ar dates obtained from Lower Cretaceous rocks.In the Arcelia area, Cabral-Cano et al. (2000) reported K/Ar dates between ~ 94 and ~ 83 Ma for metavolcanic rocks interbedded with metasedimentary deposits that contain a Lower Cretaceous fauna.In the Santo Tomás area, Martini et al. (2014) documented a rhyolitic dike with a U-Pb zircon age of ~ 130 Ma cutting a metabasalt that returns an 40 Ar/ 39 Ar date of ~ 101 Ma.Based on these considerations, the Albian-Cenomanian depositional age for the youngest stratigraphic part of the Arperos Basin remains controversial.
In the Zacatecas area, the Arperos Basin also shows a complex deformational history, characterized by at least two main shortening events (Ortega-Flores et al. 2016).However, the lack of a well-defined chronostratigraphic framework fails to constrain the age of Arperos Basin inversion in that area.
More to the north, the Arperos Basin is interpreted to continue in the Peninsular Ranges of Baja California (Figure 3e).According to some authors (Fitz-Díaz et al.

2018
), the Early Cretaceous Alisitos arc and its adjacent deep-marine basin (Busby et al. 1998;Johnson et al. 1999) are likely the equivalent of the Guerrero terrane arc and Arperos Basin in central and southern Mexico.U-Pb zircon ages of deformed metasedimentary rocks and syn-to post-tectonic intrusive bodies in the Peninsular Ranges indicate that basin inversion and arc accretion were largely completed between ~ 112 and ~ 108 Ma (Johnson et al. 1999;Alsleben et al. 2008), although shortening may have locally continued until ~ 100 Ma (Alsleben et al. 2014).

Methods
With ~1.9x10-17 moles).Samples were analysed over a year after irradiation, and the amount of 37 Ar, an interfering isotope that is possibly produced during the irradiation, was not meaningful; therefore, the correction for 37 Ar was not considered in the age calculation.GA1550 biotite (98.79 Ma; Renne et al. 1998) was used as the flux monitor.J values and errors for samples were determined by polynomial fit to replicate analyses of standards at measured positions along the length of the irradiation capsule.After irradiation, the best single crystals were selected for loading and placed in individual pits of a Cu sample holder.Laser power is expressed as a % of maximum output power of a Merchantek MIR 10-30 CO2 laser (a facetted lens is used to diffuse the beam).
All errors are reported at 1σ level.Details of the geochronological results are given in Supplemental File A.

Field and petrographic observations
The Santo Tomás area is located in the state of Mexico, southern Mexico, ~ 200 km to the south of the Sierra de Guanajuato and ~ 10 km to the west of Valle de Bravo (Figures 3d and 4a).In this area, the Arperos Basin mostly consists of calcareous and siliciclastic metaturbidites derived from the Mexican continental mainland (Martini et al. 2014).Metavolcano-sedimentary successions of the Guerrero arc are tectonically emplaced over the Arperos Basin metaturbidites (Figure 4a-b).As recognized also by previous authors (Fitz-Díaz et al. 2008;Martini et al. 2014), successions of the Arperos Basin and Guerrero arc in the Santo Tomás area display two superposed sets of ductile to brittle-ductile shortening structures.Moreover, Fitz-Díaz et al. (2008) documented that these structures are cut by inverse brittle faults developed during a third shortening event, the age of which is unknown.In this study, we will focus only on the ductile and brittle-ductile structures.D1 structures are millimetre-to decimetre-scale, isoclinal F1 folds with NW-striking axes and decametre-to kilometre-scale, T1 thrust faults that display a top-to-the-NE direction of tectonic transport (Figure 4a,b).The Santa Barbara thrust fault is probably the most iconic structure associated with the D1 shortening event (Figure 4).This regional-scale fault has been described in detail by Fitz-Díaz et al. (2008) and juxtaposes rocks of the Guerrero arc with the Arperos Basin successions (Figure 4).Another kilometre-scale thrust fault representative of the D1 shortening event is the Ojo de Agua thrust fault, which is exposed along the dirt road that connects the villages of San Nicolás and Ojo de Agua (Figure 4a,b).
The Ojo de Agua thrust fault developed at the contact between continentally-derived metaturbidites of the Arperos Basin and an overlying succession of metaturbidites and metaconglomerate (Figure 4a,b), interpreted as syn-tectonic deposits associated with the earliest stage of Arperos Basin inversion, prior to accretion of the Guerrero arc (Martini et al. 2014).This major thrust fault has a thickness of ~ 70 m and is composed of mylonites developed in metasandstone and phyllite (Figure 5a).Mylonites display an S1 foliation, which is penetrative at the submillimeter-scale and is expressed by aligned muscovite grains ± chlorite ± oxide grains (Figure 5b).Like the other T1 thrust faults in the Santo Tomás area, submillimeter-to centimetre-scale σporphyroclasts observed on XZ-planes of the finite strain ellipsoid indicate a top-to-the-NE direction of tectonic transport across the Ojo de Agua thrust fault (Figure 5c).
In the Santo Tomás area, D2 structures are superposed on F1 folds and T1 thrust faults.D2 structures are NW-striking, millimetre-to decametre-scale, open to isoclinal F2 folds and metre-to kilometre-scale T2 thrust faults, which display a top-to-the-SW direction of tectonic transport (Figure 4a,b).A representative outcrop of D2 structures is located at the southern side of the Santo Tomás dam (Figure 4a).There, metre-to decametre-scale T2 thrust faults and associated metre-to millimetre-scale F2 folds are ubiquitously superposed on a ~ 100 m-thick, top-to-the-NE T1 thrust fault (Figure 5d), which bounds two different thrust sheets of the inverted Arperos Basin.In this outcrop, the S1 foliation is expressed by aligned muscovite ± chlorite ± oxide grains and is ubiquitously superposed by an S2 crenulation cleavage, which is penetrative at the submillimetric scale and is defined by aligned, very fine flakes of muscovite and oxide seams (Figure 5e).Submillimeter-to centimetre-scale σ-porphiroclasts observed on XZ-planes of the finite strain ellipsoid indicate a main top-to-the-NE direction of tectonic transport for D2 structures.

Muscovite 40 Ar/ 39 Ar geochronology
We dated muscovite grains from structures representative of the D1 and D2 shortening events.Sample MAM3 is a low greenschist facies mylonite from a decametrescale exposure of the T1 Ojo de Agua thrust fault (Figures 4a and 5a).No macro-and microscopic D2 structure is clearly observed at the sampled outcrop.Under the microscope, sample MAM3 displays a single population of muscovite grains that are aligned along the S1 foliation and vary in size from a few millimetres to a few microns (Figure 5b).We selected 21 inclusion-free, unweathered, 250-400 μm-sized muscovite grains for single-crystal total fusion 40 Ar/ 39 Ar analyses.Singlecrystal analyses yield apparent ages ranging between 119.1 ± 6.3 and 86.9 ± 4.1 Ma (Figure 6a and Supplemental File A).Seventeen single-crystal total fusion analyses define a coherent group of dates that yields a weighted mean age of 111.7 ± 1.1 Ma (75.1% confidence, Mean Square Weighted Deviation = 1.2; Figure 6a).Two grains yield younger ages of ~ 102 and ~87 Ma, and two other grains return slightly older ages of ~ 118 and ~ 114 Ma.
Sample MAM1 is a phyllite collected at the southern side of the Santo Tomás dam (Figures 4a and 5d), from a decametre-scale T2 thrust fault superimposed on a -~ 100 m-long T1 thrust fault.Along the T2 fault, the S1 foliation is intensely folded, with overturned folds indicating a top-to-the-SW sense of tectonic transport (Figure 5d).Under the microscope, sample MAM1 shows two main foliations.Foliation S1 is continuous schistosity defined by aligned muscovite grains that vary in size from a few millimetres to a few microns (Figure 5e).Foliation S2 is a crenulation cleavage, which is expressed dominantly by concentration of very fine oxide grains and flakes of muscovite that are only a few tens of microns in size.We selected 38 inclusion-free, unweathered, 250-400 μm-sized muscovite grains for singlecrystal total fusion 40 Ar/ 39 Ar analyses.Given their size, the selected muscovite grains are associated with foliation S1.The analysed muscovite grains return a continuous spectrum of dates between 89.9 ± 1.6 and 103.7 ± 1.0 Ma (Fig. 6B).The obtained dates do not form a coherent group of ages that overlap with in the error.

The age of shortening in the Santo Tomás Area
Most of the analysed muscovite grains from sample MAM3, which was collected from the T1 Ojo de Agua thrust fault, define a coherent group of dates with a weighted mean age of ~ 112 Ma.Considering that the analysed muscovite grains are grown along the S1 mylonitic foliation, we interpret this age as an approximation of the time of top-to-the-NE tectonic transport along the T1 Ojo de Agua thrust fault and, in general, of D1 shortening in the Santo Tomás area.Previous authors pointed out that successions of the inverted Arperos Basin in southern Mexico locally experienced a Late Cretaceous tectono-thermal event, which produced Ar loss and partial to total resetting of K/Ar and 40 Ar/ 39 Ar ages (Cabral-Cano et al. 2000;Martini et al. 2014).Therefore, adopting a conservative position, we interpret the ~ 112 Ma age as the minimum age of shortening in the Santo Tomás area.Considering that the youngest group of detrital zircon grains in deformed metaturbidites of the Santo Tomás area is of ~ 118 Ma, the age of D1 shortening is bracketed to the ~ 118-112 Ma time interval.Two muscovite grains return younger ages of ~ 102 and ~ 87 Ma, which we tentatively interpreted as the result of partial or total Ar loss.
Data from sample MAM1, which was collected from a T1 thrust fault severely overprinted by T2 thrust faults and D2 folds, are more difficult to interpret.Like for sample MAM3, the analysed muscovite grains are from the S1 foliation.However, differing from sample MAM3, dates obtained for sample MAM1 do not form a coherent age group; instead, they define a continuous spectrum of ages between ~ 104 and ~ 90 Ma, covering a whole age range of ~ 14 Ma.Field and petrographic data indicate that, at the MAM1 outcrop, T2 thrust faults and associated F2 folds are ubiquitously superimposed on the S1 foliation from the submillimeter-to decametre-scale.According to this consideration, we preliminarily interpret the obtained range of dates as the result of various degree of Ar loss experienced by the analysed D1 muscovite grains, which was caused either by mechanical deformation of muscovite grains or heat associated with the D2 shortening event.This interpretation is supported by the fact that the range of dates returned by sample MAM1 overlaps with the ~ 105-83 Ma range of partly to totally reset K/Ar and 40 Ar/ 39 Ar dates obtained from the Lower Cretaceous successions of the Arperos Basin (Delgado-Argote et al. 1992;Cabral-Cano et al. 2000;Elías-Herrera et al. 2000;Martini et al. 2014).Based on this interpretation, we suggest that the ~ 104-90 Ma range of dates returned by sample MAM1 roughly approximates the time interval of the D2 ening event in the study area.
In synthesis, the shortening history of the Arperos Basin in the Santo Tomás area comprises the D1 event, which took place between Late Aptian and Early Albian time (~ 118-112 Ma), followed by the successive D2 event, which occurred sometimes between latest Albian and early Late Cretaceous time (~ 104-90 Ma).

The timing of arperos basin inversion and guerrero terrane accretion in Mexico
As observed also by previous authors (Fitz-Díaz et al. 2008;Martini et al. 2014), the Arperos Basin and Guerrero arc successions in the Santo Tomás area are presently juxtaposed and arranged in a tectonic pile, which consists of stacked thrust sheets bounded by major T1 thrust faults.This indicates that basin inversion and arc accretion occurred during the D1 shortening event, the age of which is constrained by our data between ~ 118 and ~ 112 Ma in the Santo Tomás area.Together with isotopic and biostratographic data presented in previous works, our results demonstrate that closure of the Arperos Basin and accretion of Guerrero terrane to the Mexican mainland produced a late Early Cretaceous, regional-scale suture belt, which extends from southern Mexico to the Peninsular Ranges of Baja California with a main NNW trend (Figure 7).The timing of basin inversion and arc accretion is bracketed to the ~ 118-112 Ma time interval in the Santo Tomás area of southern Mexico (this study; Figure 7), between ~ 119 and ~ 108 Ma at the Sierra de Guanajuato in central Mexico (Chiodi et al. 1988;Quintero-Legorreta 1992;Martini et al. 2011), and to the ~ 112-108 Ma time interval in the Peninsular Ranges (Johnson et al. 1999;Alsleben et al. 2008), where deformation may have locally continued until ~ 100 Ma (Alsleben et al. 2014).These age intervals overlap at ~ 112 Ma (Figure 7), precluding a major diachrony in the closure of the Arperos Basin.More precise chronological data, as well as a more systematic sampling of the suture belt, are required to clarify whether the Guerrero terrane accretion occurred at the same time everywhere or there are some slight differences in the timing of accretion across the Pacific margin.
The ages documented for the Arperos Basin closure and Guerrero terrane accretion are the oldest ones reported in Mexico for the shortening stage that followed the ~ 100 Ma-long period of extension associated with Pangea break-up.Thus, available data provide the definitive evidence that Late Aptian-Early Albian closure of the Arperos Basin has been a major tectonic event that marked the initiation of regional-scale shortening in Mexico.Previous models that failed to consider this Late Aptian-Early Albian shortening event must be revised (Salinas-Prieto et al. 2000;Talavera-Mendoza et al. 2007;Hildebrand and Whalen 2021).The Late Cretaceous age proposed by previous authors (Salinas-Prieto et al. 2000;Talavera-Mendoza et al. 2007;Hildebrand and Whalen 2021) for accretion of the Guerrero terrane is based on K/ Ar and 40 Ar/ 39 Ar dates that turned out to be reset, as well as on the age of major structures that were thought to be representative of the D1 shortening event, while they were rather developed during the D2 one.One iconic example that best represents such a confusion between structures associated with the D1 and D2 shortening events is the Teloloapan thrust fault in southern Mexico (Figure 4d).This fault has been erroneously interpreted as the Guerrero terrane suture with the Mexican mainland (e.g.Campa and Coney 1983;Mendoza and Suástegui 2000).As reported by Talavera- Mendoza et al. (2007), fossil fauna contained in syn-tectonic turbidites associated with the Teloloapan thrust-fault bracket the activity of this structure to Turonian time (~ 94-90 Ma; Gradstein et al. 2012).However, our data indicate that Arperos Basin inversion and Guerrero terrane accretion took place ~ 15 m.y. earlier and show that the age of displacement along the Teloloapan thrust fault coincides with the age range of the D2 shortening event.
In summary, it appears that, due to the lack of reliable isotopic data and a detailed study of the regional structural architecture, and given the proximity in age of the two different shortening events, structures of one event have been erroneously assigned to another.Unfortunately, this led previous workers to miss a significant part of the shortening history of southernmost North America, in particular the onset of shortening, which is vital to understand the geodynamic cause that produced the deformation of the North American Pacific margin.

Why did the Arperos Basin accommodate most of deformation during shortening initiation?
Available data document that the onset of shortening along the Pacific margin of Mexico is marked by collapse and closure of the oceanic Arperos Basin, culminating with accretion of the Guerrero and Alisitos terranes to the Mexican mainland (Figure 8a,b).At present, the kinematics of the Arperos Basin closure is not clearly understood.Some previous workers have proposed that the basin was closed by west-dipping subduction beneath the Guerrero terrane (e.g.Tardy et al. 1994;Martini et al. 2014); some others have suggested that subduction took place at both sides of the basin, beneath the Guerrero terrane and the Mexican mainland (Dickinson and Lawton 2001).In Figure 8b, we tentatively adopt the scenario in the Arperos Basin was closed by westdipping subduction.We consider that this scenario better explains the lack of arc magmatism in the Mexican mainland, as well as the structural architecture of the Guerrero terrane suture belt, with the Guerrero terrane arc tectonically emplaced over the Arperos Basin with a top-to-the-NE transport direction.Despite the uncertainty on the kinematics of basin closure, available data indicate that, during shortening initiation, the deformation was well localized along the Arperos Basin, forming a ~ 80 km-wide suture belt that runs from southern to northern Mexico with a NNW trend (Figures 7 and 8a,b).The Guerrero and Alisitos terranes interior, as well as the Mexican mainland interior, were not involved in this deformation event, as suggested by the absence of Early Cretaceous shortening structures and regionalscale angular unconformity to the west and east of the suture belt.We propose that such a marked localization of deformation during shortening initiation was controlled by the occurrence of a main zone of weakness in the lithosphere.At the moment of shortening initiation, in Late Aptian or Early Albian time, the Arperos Basin was a major zone of attenuated lithospheric thickness and, considering that the oceanic floor within the basin was formed just a few m.y.before shortening was initiated, was probably also a zone of high heat flow.Based on these considerations, when shortening began, the Arperos Basin likely represented the weakest part of the lithosphere that bounded to the east the subducting Farallon plate (Figure 8b).For that reason, deformation was mostly accommodated by the Arperos Basin during shortening initiation.This scenario highlights the fundamental control that pre-existent zones of lithospheric weakness play over the evolution of orogenic belts.In addition, considering that the Arperos Basin was dominantly composed of thinly bedded turbidite and radiolarite deposits, whereas the adjacent continental  Chiodi et al. 1988;Quintero-Legorreta 1992;Martini et al. 2011), and ~ 112-108 Ma in the Peninsular Ranges (Johnson et al. 1999;Alsleben et al. 2008), where deformation may have locally continued until ~ 100 Ma (Alsleben et al. 2014).These age intervals overlap at ~ 112 Ma.We preliminarily take this age as a first approximation of the Guerrero terrane accretion age.Baja California is restored to its Cretaceous palaeogeographic position according to Schaaf et al. (2000).
mainland and Guerrero terrane interior were dominated by carbonate platforms and reefal limestone, the lateral variation in lithology and mechanical properties may have also contributed to such a localization of deformation.
By the time the Arperos Basin was closed, and the Guerrero and Alisitos terranes were amalgamated to the Mexican mainland, first-order heterogeneities in lithospheric thickness and heat flow were probably largely reduced.Relatively rapid lithospheric thickening across the Arperos Basin during the development of the Guerrero suture belt is well indicated by the fact that Aptian deep-marine successions made up of turbidite and radiolarite deposits are uncoformably overlain by the Late shallow-marine limestone of La Perlita Formation (Chiodi et al. 1988;Quintero-Legorreta 1992).Cooling of the Arperos Basin area and progressive thermal rehomogenization of the southern North American plate after basin closure is also a plausible process, although it is difficult to estimate the time required for it to occur.The heat flow of modern back-arc basins of the western Pacific Ocean mostly varies in the range of ~ 90-380 m W m −2 (e.g. the Japan, Ryuku, Shikoku, and Mariana basins; Sychev and Sharaskin 1984 and references therein), whereas recent orogenic belts that formed during the past ~ 30-35 m.y. by progressive closure of an oceanic basin (e.g. the Alpine-Himalayan belts) are associated with a heat flow that ranges between ~ 40 and ~ 80 m W m −2 (Rybach et al. 1997;Speranza et al. 2013;Basilici et al. 2019).These data suggest that, ~ 30-35 m.y. after the closure of an oceanic basin, the lithospheric heat flow decreases by more than a half.Based on these data, we tentatively propose that major thickness and thermal heterogeneities in the lithosphere to the east of the Farallon plate could have been largely reduced ~ 15 m.y. after the Arperos Basin was closed, favouring the propagation of shortening to a more extensive area and permitting the development of the ~ 650 km-wide, critically tapered wedge between latest Albian and Eocene time (Figure 8c).Whether the Late Aptian-Early Albian Guerrero-Alisitos suture belt and the latest Albian-Eocene critically tapered wedge are two different and superposed orogenic belts as proposed by Hildebrand and Whalen (2021) or two different phases of the same orogenic cycle as proposed by Fitz-Díaz et al. (2018) remains to be established.In any case, these two belts are the manifestation of a complex shortening evolution, which consists of two main stages: 1) an early stage of ~ 10 m.y. or less that involved subduction of a back-arc oceanic plate, and in which deformation was localized along a pre-existent zone of lithospheric weakness, and 2) a subsequent stage of ~ 60 m.y., in which shortening was propagated from the North American trench to the plate interior, producing the development of a critically tapered wedge.The second stage of the deformation evolution was described in detail by Fitz-Díaz et al. (2014), who documented that shortening propagation occurred in at least three pulses at ~ 93-80 Ma, ~ 75-64 Ma, and ~ 55-43 Ma.On the other hand, the early stage of deformation was not fully recognized in most previous works.This stage is a key for our understanding of the causes that triggered shortening in southern North America after a ~ 100 m.y.period of extension.

Conclusions
The integration of our new 40 Ar/ 39 Ar ages with previous isotopic and biostratigraphic data indicates that the Mexican shortening history that followed extension associated with Pangea break-up was complex and punctuated by several deformation events.In this work, we group these events into two main stages.Stage 1 took place between Late Aptian and Early Albian time, marking the beginning of shortening.During this stage, deformation was localized across the Arperos Basin, which was a major zone lithospheric weakness.Shortening during stage 1 produced the synchronous closure of the Arperos Basin and accretion of the Guerrero terrane arc to the Mexican mainland.Considering that this stage marked the change from extensional to compressional tectonics in southern North America, reconstructing the structural evolution of Guerrero terrane suture belt surely represents a key issue to understand the process that controlled the dynamic of subduction and triggered shortening in Mexico.Stage 2 took place between latest Albian and Eocene time.Considering that Arperos Basin closure was largely completed by the beginning of stage 2, firstorder heterogeneities in thickness and heat flow were largely reduced, leading the lithosphere to the east of the Farallon plate to propagate shortening from the trench to the continental interior.The result of this stage of deformation is a ~ 650 km-wide fold-thrust belt, which extends from the modern trench to the vicinity of the Gulf of Mexico and overprints the Guerrero terrane suture belt.These two stages may represent two different and superposed orogenic event or two different phases within the evolution of the same orogenic cycle.
(2) D1 is associated with closure of the Arperos Basin and accretion of the Guerrero terrane.
(3) Inversion of the Arperos Basin marks the onset of shortening in southern North America.(4) During shortening initiation, deformation was localized at the Arperos Basin because was the weakest part of the lithosphere.

Figure 1 .
Figure 1.Schematic structural map of North America, showing the location and extent of major Mesozoic and Cenozoic orogenic belts and tectono-stratigraphic terranes.Adapted from .Yonkee and Weil (2015) andFitz-Díaz et al. (2018.)

Figure 4 .
Figure 4. A. Detailed geological map of the Santo Tomás area, showing the major shortening structures exposed in the region and the location of samples analysed in this work (modified from Martini et al. 2014).B. Geological section across the Santo Tomás area, in which is presented the architecture of the Guerrero terrane suture belt exposed in the study area.

Figure 5 .
Figure 5. A. Photograph showing a detail of the T1 Ojo de Agua thrust fault, from which it has been extracted sample MAM3.This outcrop is composed of mylonitic metasandstone and phyllite from the Ojo de Agua assemblage.B. Photomicrographs showing a detail of the S1 foliation in MAM3.S1 dominantly expressed by muscovite with a grain size that spans from a few millimetres to a few hundreds of microns.C. Photomicrographs showing a σ-porphyroclast from the Ojo de Agua thrust fault mylonite.D. Photograph showing F2 folds and T2 thrust faults exposed along the southern side of the Santo Tomás dam.Sample MAM1 has been extracted from this outcrop.E. Photomicrographs showing a detail of the S1 and S2 foliations of sample MAM1.

Figure 6 .
Figure6.40 Ar/39 Ar results obtained from sample MAM3, representative of the T1 Ojo de Agua thrust fault, and sample MAM1, representative of a T2 thrust fault superimposed to the S1 foliation.Single-crystal total fusion 40 Ar/ 39 Ar analyses are represented in an Age Spectra Plot.Analyses are ordered by volume of 39 Ar.Analyses used for calculation of the weighted mean age are dark grey boxes.

Figure 7 .
Figure 7. Schematic map of localization and extent of the Guerrero and Alisitos terranes suture belts, showing the age range of shortening for each area.Available data that shortening occurred at ~ 118-112 Ma in the Santo Tomás area of southern Mexico (this study), ~ 119-108 Ma at Guanajuato in central Mexico (.Chiodi et al. 1988; Quintero-Legorreta 1992; Martini et al. 2011), and ~ 112-108 Ma in the Peninsular Ranges(Johnson et al. 1999;Alsleben et al. 2008), where deformation may have locally continued until ~ 100 Ma(Alsleben et al. 2014).These age intervals overlap at ~ 112 Ma.We preliminarily take this age as a first approximation of the Guerrero terrane accretion age.Baja California is restored to its Cretaceous palaeogeographic position according toSchaaf et al. (2000).

Figure 8 .
Figure 8. Three-step tectonic evolution of Mexico between Early Cretaceous and Palaeogene time. A. During Early Cretaceous time, Mexico was the site of extension, which produced the development of several continental to marine basins.A large part of extension seems to have been accommodated by the Arperos Basin, which, in its most mature stage was partly floored by oceanic crust.B. Between Late Aptian and Early Albian time, shortening was initiated in Mexico, producing the inversion of the Arperos Basin and the accretion of the Guerrero and Alisitos terranes to the continental mainland.The result of this deformation stage is the development of a ~ 80 km-wide suture belt, which runs from southern to northern Mexico and is well localized at the boundary between the terranes and the Mexican mainland.C. By the end of Early Cretaceous time or the beginning of Late Cretaceous time, a second stage of shortening produced the development of a ~ 650 km-wide critically tapered wedge, which extends from the modern trench to the vicinity of the Gulf of Mexico, overprinting the older suture belt formed during the previous stage.
Boschman et al. (2017)tion across western southern Mexico, showing the location of the Guerrero terrane suture belt and its relationship with other stratigraphic units.B.Interpretative cross section across the Arperos Basin previous to its closure and development of the Guerrero terrane suture belt.Adapted after .Martini et al. (2014).The concept of the Guerrero plate is taken fromBoschman et al. (2017).
. However, the age of the Arperos Basin closure and terrane accretion in southern Mexico is poorly defined.This is mainly because the age of the youngest rocks within the Arperos Basin is controversial.Radiolarians (Contreras-Rodríguez et al. 1990), ammonites (Cantú-Chapa 1968), and other marine invertebrates (Israde-Alcantara and Martinez-Alvarado 1986) indicate a Berriasian-Aptian age for the basin.The age of the youngest detrital zircon group in metaturbidites from the Arperos Basin is ~ 118 Ma (Martini et al. 2014), which agrees with the age of the fossil fauna.However, Delgado-Argote et al. (1992) and Elías-Herrera et al. (2000) obtained