Introduction
The Baza paleolake, in the eastern part of the Guadix-Baza Basin (S Spain) (Fig. 1) is the largest, southernmost paleolake in Europe and holds an invaluable sedimentary archive of the Plio-Pleistocene period. The rich and extensive depositional infill of the Basin also holds the key to understanding the former gateway between the Atlantic Ocean and Mediterranean Sea (the Betic Corridor) before the early Pliocene connection at Gibraltar (e.g., Soria et al., Reference Soria, Viseras and Fernández1998), providing an unprecedented paleoclimatic and paleogeographic record of the region throughout the Cenozoic. Unfortunately, the precise chronology of the basin-fill sediments is still under debate, and in many cases it lacks numerical age dating, in particular for the Baza Formation.
(A, B) Geographical and geological setting of the study area. (C) Detailed geologic map of the eastern portion of the Guadix–Baza Basin. Stratigraphic sections: BdP = Barranco del Paso; BL = Barranco León; FN1 = Fuent Nueva-1; FN3- Fuente Nueva-3; FTC = Fuentecica; MM = Mina Mojona; OD = Orce.
The Baza Formation (Vera et al., Reference Vera, Fernández, López-Garrido and Rodríguez-Fernández1985), the main visible rock unit that fills the eastern part of the basin, includes lacustrine–palustrine carbonates and alluvial sediments, and is known for hosting numerous Plio-Pleistocene paleontological sites (see overview in Maldonado-Garrido et al., Reference Maldonado-Garrido, Piñero and Agustí2017) with micromammal and/or large land mammal assemblages that have been well known and studied for years. Overall, the biozones that are present in the Baza Formation are thought to span the Pliocene–Pleistocene boundary (e.g., Agusti, Reference Agusti1986; Agustí et al., Reference Agustí, Cabrera, Garcés, Krijgsman, Oms and Parés2001, Reference Agustí, Lozano-Fernández, Oms, Piñero, Furió, Blain, López-García and Martínez-Navarro2015). Last, but not least, Guadix-Baza is the only basin in Western Europe that records early human presence in a continuous fluvio-lacustrine environment (Toro-Moyano et al., Reference Toro-Moyano, Martínez-Navarro, Agustí, Souday, Bermúdez de Castro, Martinón-Torres and Fajardo2013).
Besides the abundant studies, literature, and significance of the Plio-Pleistocene deposits, the chronological constraints of the sediments that constitute the Baza Formation have been under debate for years, and therefore the age of Mammal Neogene fauna biozones (MN biozones), which mostly are based on magnetochronology, is still a controversial issue. This paper aims to correct this situation.
Since the 1980s, paleomagnetism has been the first line of inquiry to establish chronologies in the basin (Oms et al., Reference Oms, Garcés, Parés, Agustí, Anadón and Julià1994, Reference Oms, Dinarès and Parés1996, Reference Oms, Dinares-Turell, Agustı and Parés1999. Gibert i Clols et al. (Reference Gibert i Clols, Arribas, Teran, Palomar and Gibert i Clols1992), based on unpublished paleomagnetic results, suggested the presence of the Olduvai Subchron in the section of Barranco de Orce. A number of paleomagnetic studies have ensued (Agusti et al., Reference Agustí, Oms, Garcés and Parés1997, Reference Agusti, Oms and Pares1999; Garcés et al., Reference Garcés, Agustí and Parés1997; Oms, Reference Oms1998; Oms et al., Reference Oms, Agustí, Gabàs and Anadón2000; Gibert et al., Reference Gibert, Ortí and Rosell2007; Scott et al., Reference Scott, Gibert and Gibert2007), whereas Gibert et al. (Reference Gibert, Scott and Deino2024) recently have reported the presence of the Jaramillo Subchron in a new section (OD) near de village of Orce (Fig. 1C), providing a new timeline for the upper part of the basin sedimentary infill. Still, the precise age assignment of such a plethora of magnetostratigraphies may be flawed, mostly due to three circumstances.
First, the lacustrine–palustrine sediments that constitute the vast body of the Baza Formation, while amenable, hold two critical limitations. The natural remanent magnetization (NRM) intensities are often weak or very weak (habitually in the range of 10−6 A/m). While treatable with modern SQUID magnetometers, these intensity values are a serious challenge upon progressive demagnetization. Furthermore, the majority of units, horizons, and beds that have been studied, typically reveal a severe secondary paleomagnetic component (e.g., Oms et al., Reference Oms, Agustí, Gabàs and Anadón2000). While typically removed via alternating field or progressive thermal demagnetization, the intensity that is left is very often too feeble and/or unstable to produce reproducible magnetization directions. On top of all this, bedding of the basin-fill sediments, except for a few places, is almost flat, rendering stability tests such as fold (or simply tilt) non-viable.
Second, in addition to the abovementioned paleomagnetic challenges, the fluvio-lacustrine Plio-Pleistocene deposits have been lacking independent age control for very a long time, probably partly due to the somewhat overall unfavorable sedimentary and geochemical context for the application of numerical dating methods, although with some exceptions. Initial amino acid racemization (AAR) studies by Torres et al. (Reference Torres, Llamas, Canoria, Garcia-Alonso, Garcia-Cortes and Mansilla1997) and Ortiz et al. (Reference Ortíz, Torres, Llamas, Canoira, Alonso, Morena and Lucini2000) furnished various ages for fossil gastropods collected from various fossil localities of the eastern part of the basin associated to the Baza Formation: 983 ± 58 ka and 1095 ± 55 ka for Venta Micena, 476 ± 24 ka for Cúllar-Baza, and 491 ± 84 ka for Huéscar-1 later on. Electron spin resonance (ESR) dating method was tested on fossil teeth and quartz grains from various archaeo-paleontological sites of the Orce area and provided the first numerical age constraints to Barranco León (< 1.43 ± 0.38 Ma; Toro-Moyano et al., Reference Toro-Moyano, Martínez-Navarro, Agustí, Souday, Bermúdez de Castro, Martinón-Torres and Fajardo2013), Fuente Nueva-3 (1.19 ± 0.21 Ma; Duval et al., Reference Duval, Falguères, Bahain, Grün, Shao, Aubert and Toro-Moyano2012), and Venta Micena (1.37 ± 0.24 Ma; Duval et al., Reference Duval, Falguères, Bahain, Grün, Shao, Aubert and Hellstrom2011), despite a number of methodological challenges (Duval, Reference Duval2008).
The first luminescence dating study in the basin was carried out at the fossil locality of Huéscar-1 (Demuro et al., Reference Demuro, Arnold, Parés and Sala2015) and shed new light on the age of the fossil assemblage (i.e., by yielding much younger chronological constraints (420–570 ka) than suggested by the biostratigraphy/biochronology; Ros-Montoya et al., Reference Ros-Montoya, Palombo, Espigares, Palmqvist and Martínez-Navarro2018), although closely consistent with the initial AAR results by Ortíz et al. (Reference Ortíz, Torres, Llamas, Canoira, Alonso, Morena and Lucini2000). Lastly, Álvarez et al. (Reference Álvarez, Parés, Granger, Duval, Sala and Toro2015) reported the preliminary cosmogenic Al–Be burial age of 1.50 ± 0.31 Ma for quartz grains at Fuente Nueva-3. While this age is in agreement within error of the independent age control available from the ESR method, it nevertheless should be treated with caution, where the age can be considered as a minimum since surface erosion rate and thus post-burial production is unconstrained; but, also considered as a maximum because a prior burial history may have been involved with sediment storage (Álvarez et al., Reference Álvarez, Parés, Granger, Duval, Sala and Toro2015; Fujioka et al., Reference Fujioka, Parés, Granger, Padilla, Pérez, Miguens and Revilla2024). In summary, despite some very encouraging results obtained in the last couple of decades, it remains more crucial than ever to obtain a robust independent age control for the deposits of the Baza Formation and the associated archaeological and paleontological records in order to complete and refine the existing chronostratigraphic framework that is mostly based on the combination of magnetic stratigraphy and biostratigraphy.
Third, although sediments in the eastern part of the basin are virtually undeformed, bedding is slightly off horizontal (e.g., Vera et al., Reference Vera, Fernández, López-Garrido and Rodríguez-Fernández1985), which prevents long-distance stratigraphic correlations that are based on elevation (e.g., Gibert et al., Reference Gibert, Scott and Deino2024).
The vast majority of local magnetostratigraphies in the Orce-Venta Micena area have produced dominantly reverse polarity magnetozones so far. These have been assigned, mostly based on the mammal assemblages and the few numerical age constraints available, to the Matuyama Chron (2.61–0.77 Ma; note that the age of all geomagnetic boundaries is taken from Gradstein and Ogg, Reference Gradstein, Ogg, Gradstein, Ogg, Schmitz and Ogg2020). Exceptions include the sections of Mina Mojona, Rincón de la Losa, and Galera (Garcés et al., Reference Garcés, Agustí and Parés1997; Scott et al., Reference Scott, Gibert and Gibert2007).
The Mina Mojona (MM, Fig. 1) is located about a kilometer north from the lake paleomargin. Scott et al. (Reference Scott, Gibert and Gibert2007) reported a 7-meter-long section of mostly clays with gypsum, most likely corresponding to Lithostratigraphic Unit (LU) 6, which was defined by Santamaría Barragán et al. (Reference Santamaría Barragán, Parés, Colino García, Churruca Clemente, Fujioka, Serrano-Ramos and Toffolo2025) as the lowermost outcropping stratigraphic unit in the area. The Mina Mojona section has normal polarity and was tentatively assigned to the Olduvai Subchron based on biostratigraphic considerations. Since no magnetic reversal was recorded at that section, the complete thickness of this normal magnetozone is unknown.
Farther north, about 3 km from the former lake shoreline, the section known as Rincon de la Losa offers a more complete paleomagnetic record, R → N → R from bottom to top, which is also interpreted as the Olduvai Subchron (Scott et al., Reference Scott, Gibert and Gibert2007), constituting as of today, the only locality where the full expression of the Subchron (1.93–1.78 Ma) has been documented in the eastern sector of the Basin, albeit without the support of numerical ages. Unfortunately, the section is 3 kilometers away from key localities such as Fuente Nueva-3 and Barranco León, and therefore a slight tilt, but relevant over such distance, makes within-basin correlation potentially misleading. A tilt of only one degree, for example, over such distance (3 km) translates into about 50 meters of height difference. For completeness, it should also be mentioned that the normal polarity intervals initially identified in the Barranco de Orce section (Agustí et al., Reference Agustí, Oms, Garcés and Parés1997), located close (East) of the Fuentecica section, have not been confirmed by later studies (Agustí et al., 2007; Scott et al., Reference Scott, Gibert and Gibert2007; and references therein).
Be that as it may, it follows that in the eastern sector of the Baza paleolake, where up to 9 small-mammal Pleistocene biozones have been recognized (e.g., Agustí et al., Reference Agustí, Lozano-Fernández, Oms, Piñero, Furió, Blain, López-García and Martínez-Navarro2015), an unprecedented record of early, middle, and late Villafranchian mammal faunas is found (e.g., Rook and Martínez-Navarro, Reference Rook and Martínez-Navarro2010), and possibly the oldest human presence in Eurasia has been reported (Toro-Moyano et al., Reference Toro-Moyano, Martínez-Navarro, Agustí, Souday, Bermúdez de Castro, Martinón-Torres and Fajardo2013), the chronology of the sedimentary fill is still poorly understood. In order to correct this situation, we carried out a combined electron spin resonance (ESR)–magnetostratigraphic study in a section adjacent to the southern margin of the paleolake, where 45 meters of lake and lakeshore sediments are exposed, and aimed to furnish new chronological control on the deposits of the Baza Formation.
Geological setting
The Guadix-Baza basin, which is one of the largest intramontane basins of the Betic Cordillera, with an area of around 5,000 km2, is located on the boundary between the Internal and External Zones. The studied section is referred to, although not described, in Martín-Suárez (Reference Martín Suárez1988) as “Fuentecilla”. Oms (Reference Oms1998) later logged the section, which was described in Oms et al. (Reference Oms, Agustí, Gabàs and Anadón2000) as “Fuentecica”, and it is found about one km east of the Orce village (Fig. 1C), in the eastern part of the basin. The studied sedimentary rocks belong to the Baza Formation (Vera et al., Reference Vera, Fernández, López-Garrido and Rodríguez-Fernández1985), which is made up from bottom to top, by three informal members (Vera et al., Reference Vera, Fernández, López-Garrido and Rodríguez-Fernández1985; Oms et al., Reference Oms, Agustí, Gabàs and Anadón2000, Reference Oms, Anadón, Agustí and Juliá2011; Anadón et al., Reference Anadón, Julià, Oms, Toro, Agustí and Martínez2003): (1) limestone member (lower member), which is a calcareous member dominated by shallow lacustrine limestones in the lower and middle part, and by clays and marly limestones at the top; (2) red terrigenous member (middle member), which is composed of red mudstones with interbedded sandstones with some minor palustrine mudstones and limestones; and (3) silty-calcareous member (upper member), which includes lacustrine limestones, carbonate silts, and dark mudstones.
The Baza Formation was deposited during a period of exclusively continental sedimentation in an endorheic basin (Fernández et al., Reference Fernández, Bluck and Viseras1993; Viseras and Fernández, Reference Viseras and Fernández1994, Reference Viseras and Fernández1995; Viseras et al., Reference Viseras, Fernández, García-García, Meléndez Hevia and Soria1998). This sedimentary stage was later interrupted in the Middle or Late Pleistocene when a geomorphological inversion of the basin took place and it was captured in its entirety by the drainage network of the Guadalquivir River, becoming an exorheic domain mainly subjected to erosion (Viseras and Fernández, Reference Viseras and Fernández1992; Calvache and Viseras, Reference Calvache and Viseras1997). The youngest lacustrine system furnishes a minimum age for the capture by the Guadalquivir River (e.g., García-Tortosa et al., Reference García-Tortosa, Sanz de Galdeano, Alfaro, Jiménez Espinosa, Jiménez-Millán and Lorite Herrera2008, Reference García-Tortosa, Alfaro, Martin-Rojas, Medina-Cascales and Giralt2024; Demuro et al., Reference Demuro, Arnold, Parés and Sala2015).
Although initially described by Oms (Reference Oms1998) and later by Gibert et al. (Reference Gibert, Scott and Ferràndez-Cañadell2006), the Fuentecica section (FTC) has never been studied from a paleomagnetic point of view. Importantly, a stratigraphic layer with micromammal fossils assigned to micromammal zone MmQ1 has been identified upstream along Fuentecica creek (FTC-5 in Fig. 2A) (Agustí et al., Reference Agustí, Piñero, Lozano-Fernández and Jiménez-Arenas2022; Agustí and Piñero, Reference Agustí and Piñero2023), which enables positioning the assemblage within the biostratigraphic biozonation of the basin, possibly constraining the deposits to the Early Pleistocene, which is an older position than other fossil localities such as Fuente Nueva-3, Barranco León, and Venta Micena (e.g., Oms et al., Reference Oms, Agustí, Gabàs and Anadón2000; Agustí et al., Reference Agustí, Lozano-Fernández, Oms, Piñero, Furió, Blain, López-García and Martínez-Navarro2015).
(A) General context of the Fuentecica section. Fossiliferous location “Fuentecica 5” is also shown (see main text). (B) Aerial view showing the location of the stratigraphic profiles (yellow-dashed lines) in relation to the middle and upper members. (C) Detail of the sampled middle member including location of ESR sample and thin section samples (19FTC-E1 and 24FTC-19LD, respectively). (D) Position of ESR sample. (E) Detail of the sampled portion of the upper member, including location of thin section sample 24FTC-LMST, at the very top of the section.
Methods
Lithostratigraphic description
The stratigraphic profile was logged along the ∼NW–SE creek known as La Fuentecica, the headwaters of which are found on the cliff-making, uppermost lacustrine limestones, farther south. Stratigraphy was logged with the help of a Jacob’s staff (a 1.5-m pole, marked off in 10-cm increments with an attached bubble level), which allows direct measurement of true bed thickness. In addition, centimeter-precision GPS was used to validate measurements and get absolute elevations.
Microfacies study
Two oriented blocks of intact sediment were selected for thin section analysis with the aim of determining their microstructure, mineral composition, and post-depositional degree of alteration. One sample was collected from the uppermost carbonates (24FTC-LMST), and another one from detrital deposits farther down in the section (24FTC-19LD) next to the sediment sample (19FTC-E1), which was collected for ESR dating (Fig. 2).
Samples were processed at the National Center for Human Evolution (CENIEH, Burgos, Spain) following a standard procedure: they were first dried in an oven, then embedded in epoxy resin, mounted on glass slides, and polished to a thickness of 30 µm to obtain micromorphology thin sections. These thin sections were observed using an Olympus BX51 polarizing microscope at magnifications of 25–400×. Descriptions were based on standard literature (Stoops et al., Reference Stoops, Marcelino and Mees2018; Stoops, Reference Stoops2021).
Paleomagnetism
Oriented samples for paleomagnetism were obtained in the field as either small blocks or drilled cylinders. In clay-rich sediments, two or three approximately 10-cm3 cubic samples were obtained at a given horizon. A ceramic knife was used to carve a horizontal surface, and an orientation line was scribed (N–S when possible for simplicity) with the help of a compass and level. When needed the block was impregnated with an aqueous solution of sodium silicate and allowed to dry for further hardening. The knife was then used to scribe along the outer edges of the hardened block to facilitate its extraction. Indurated layers such as the limestones were drilled with a portable, gas-powered modified chainsaw engine. Standard 1-inch, stainless steel drill bits were used to obtain ∼5-cm long cylinders. A standard compass on an orienting fixture equipped with an inclinometer was used to accurately measure the trend and plunge of the drilled cylinder.
Thirty horizons (= paleomagnetic sites) were sampled, producing a total of 75 specimens over two different campaigns. All measurements were made at the Paleomagnetism Laboratory of the Geochronology Facility (CENIEH, Spain). Remanent magnetization was measured with a 3-axis, 4.2-cm-access SQUID magnetometer (755-4K Superconducting Rock Magnetometer, “Helium-free,” 2G Enterprises). The magnetometer has a built in in-line system for AF demagnetization up to 170 mT and it is placed in a large set of Helmholtz coils for reducing the ambient magnetic field. Thermal demagnetization was performed with an ASC Model TD48 furnace (single chamber). Rock magnetic analysis included hysteresis cycles, which were obtained with a 3900 VSM MicroMag (Princeton) and progressive isothermal remanent magnetization (IRM) acquisition using an ASC impulse magnetizer (ASC Scientific). Characteristic remanent magnetization (ChRM) component directions were calculated for all specimens using a based principal component analysis (Kirschvink, Reference Kirschvink1980) software (Ramón et al., Reference Ramón, Pueyo, Oliva-Urcia and Larrasoaña2017), guided by visual inspection of orthogonal demagnetization plots (Zijderveld, Reference Zijderveld1967). ChRM directions were used to calculate the virtual geomagnetic poles (VGP) at each sampling site, and the value of the VGP latitude used to build a local magnetostratigraphy.
ESR dating of optically bleached quartz grains
Sampling
One sediment sample (19FTC-E1) was collected in September 2019 from a thin layer of fine-grained sandstones positioned in the lower part of the local sedimentary sequence (i.e., within the lower member; Fig. 2). Sampling was performed by hammering a PVC tube into the section. In-situ measurement of radioactivity was carried out using a NaI probe connected to a Canberra Inspector 1000 multi-channel analyzer at the exact sampling point.
Preparation
19FTC-E1 was prepared in the laboratory under conditions of limited illumination. 100- to 200-µm size fraction was extracted from wet sieving and the quartz grains were purified following the same procedure as in Parés et al. (Reference Parés, Haddoumi, Duval, Aouraghe, Álvarez-Posada, Pla-Pueyo and Benito-Calvo2023).
ESR dosimetry
19FTC-E1 was dated via the multiple aliquots additive dose (MAAD) method and following the multiple center (MC) approach initially defined by Toyoda et al. (Reference Toyoda, Voinchet, Falguères, Dolo and Laurent2000). The sample was divided into 14 aliquots of 150 mg each. Gamma irradiations were performed at CENIEH using a Gammacell-1000. The following irradiation doses were applied to 12 aliquots (dose rate: 5.96 ± 0.14 Gy/min): 46.6, 111.7, 204.9, 419.1, 698.4, 1024.3, 1676.2, 2793.7, 4656.2, 7449.9, 13968.5, and 23280.8 Gy. One aliquot of each sample was optically bleached by artificial solar light after exposure in a SOL2 (Dr Hönle) simulator for about 1500 h in order to evaluate the non-bleachable component of the ESR signals associated with the aluminum (Al) center. Finally, one aliquot was kept unirradiated (= natural aliquot).
ESR measurements were carried out at CENIEH, with an EMXmicro 6/1Bruker X-band ESR spectrometer coupled to a standard rectangular ER 4102ST cavity. To ensure constant experimental conditions over time, the temperature of the water circulating in the magnet is controlled and stabilized at 18ºC by a water-cooled Thermo Scientific NESLAB ThermoFlex 3500 chiller, and the temperature of the room is kept constant at 20ºC by an air conditioning unit. ESR measurements were performed at low temperature (∼90°K) using a ER4141VT digital temperature control system based on liquid nitrogen cooling. Further details about the setup and its stability over time can be found in Duval and Guilarte Moreno (Reference Duval and Guilarte Moreno2012) and Duval et al. (Reference Duval, Voinchet, Arnold, Parés, Minnella, Guilarte, Demuro, Falguères, Bahain and Despriée2020).
The ESR signals of both the aluminum (Al) and titanium (Ti) centers were acquired in a single spectrum (e.g., Bartz et al., Reference Bartz, Duval, Brill, Zander, King, Rhein, Walk, Stauch, Lehmkuhl and Brückner2020), using the following acquisition parameters: 5 mW microwave power, 2048 points resolution, 40 mT sweep width, 342 mT center field, 100 kHz modulation frequency, 0.1 mT modulation amplitude, 40 ms conversion time, 10 ms time constant, and 1–2 scans. All aliquots were measured 3 times after a ∼120° rotation in the cavity in order to consider angular dependence of the Al and Ti signals due to sample heterogeneity. Measurements were repeated four times over distinct days in order to evaluate the variability of ESR data.
A baseline correction with a cubic function was applied to all spectra using Bruker WINEPR software. The ESR intensity of the Al signal was extracted from peak-to-peak amplitude measurements between the top of the first peak (g = 2.0185) and the bottom of the 16th peak (g = 1.9928) (Toyoda and Falguères, Reference Toyoda and Falguères2003). The ESR intensities of the Ti centers were measured following option D (sensu Duval and Guilarte, Reference Duval and Guilarte2015). An example of an ESR spectrum is given in Figure 4.
For each aliquot, ESR intensities of Al and Ti centers were corrected by the corresponding receiver gain value, number of scans, aliquot mass, and a temperature correction factor (Duval and Guilarte Moreno, Reference Duval and Guilarte Moreno2012). The fitting procedures were carried out with Microcal OriginPro 8.5 software using a Levenberg–Marquardt algorithm by chi-square minimization. Final dose response curves (DRCs) were obtained by using the mean normalized ESR intensities and associated error (1σ, or standard deviation) derived from each repeated measurement. Depending on the ESR signal considered, fitting was performed using either an exponential+linear function (EXP+LIN) function (see equation in Duval, Reference Duval2012) or a single saturating exponential (SSE) function (see equation in Duval et al., Reference Duval, Grün, Falgueres, Bahain and Dolo2009), and data were weighted by the inverse of the squared ESR intensity (1/I2) (Duval, Reference Duval2012), the inverse of the squared experimental error (1/s2).
Dose rate evaluation and ESR age calculation
The dose rate was evaluated from a combination of in-situ and laboratory measurements. The gamma dose rate was derived from a field measurement and calculated with the threshold approach (Duval and Arnold, Reference Duval and Arnold2013). External alpha and beta dose rate components were calculated from the radioelement concentrations of the raw sediment measured by ICP-MS/OES following a four-acid digestion procedure. Dose rate calculations were performed with DRAC (v1.2.) (Durcan et al., Reference Durcan, King and Duller2015) using the dose rate conversion factors from Guérin et al. (Reference Guérin, Mercier and Adamiec2011), alpha attenuation from Brennan et al. (Reference Brennan, Lyons and Phillips1991), beta attenuations from Guérin et al. (Reference Guérin, Mercier, Nathan, Adamiec and Lefrais2012), and etching depth attenuations from Brennan (Reference Brennan2003). Etching-depth values of 10 ± 5 µm were considered following Duval et al. (Reference Duval, Guilarte, Campaña, Arnold, Miguens, Iglesias and González-Sierra2018). Current water content was evaluated in the laboratory by drying the sediment at 50ºC in an oven over three weeks. A long-term water content of 40% (% dry weight) was used for the dose rate evaluation, corresponding to 60% of the saturated water content measured in the laboratory. Internal dose rate was assumed to be 30 ± 10 μGy/a, based on the work from Van den Berghe et al. (Reference Van den Berghe, Van Renterghem and Leenaers2008), and using an alpha efficiency of 0.07 ± 0.01 (Bartz et al., Reference Bartz, Arnold, Spooner, Demuro, Campaña, Rixhon, Brückner and Duval2019). ESR ages were calculated by dividing the DE value by the total dose rate. Errors are 1σ.
Cyclostratigraphy
We used rock magnetism to obtain a high-resolution cyclostratigraphy of the lower part of the section. The time resolution of cyclostratigraphy ranges from 0.02 to 0.40 Ma, reflecting the periodicity of the forcing parameters (i.e., the 16–23 ka for precession indexes, 41 ka obliquity index, and 95–128 ka and 405 ka orbital eccentricity indexes; Laskar et al., Reference Laskar, Robutel, Joutel, Gastineau, Correia and Levrard2004). These Milankovitch cycles are known to result in natural long-term variations in Earth’s climate.
Unoriented hand specimens were collected every 0.25 m along the measured section, for a total 104 samples. The samples were packed in standard paleomagnetic boxes and weighed. Both magnetic susceptibility and anhysteretic remanent magnetization (ARM) measurements were made on each sample. The ARM was designed to isolate the fine-grained magnetite concentration of the mass normalized samples using a 100mT AF field and a biasing of 0.97mT DC field as a function of stratigraphic depth. Remanence measurements were carried out with a 2G SQUID magnetometer housed in a shielded room in the Paleomagnetism Laboratory at Lehigh University, USA.
These quick, nondestructive analyses allowed us to test for orbital frequencies and to estimate the sediment accumulation rate of the stratigraphic section. Given the lithostratigraphic succession recorded at Fuentecica, we limited the cyclostratigraphic study to the lower ∼26 meters of the section corresponding to the siliciclastic deposits of the middle member of the Baza Formation where magnetic susceptibility measurements were most effective at encoding the precessional signal, in contrast to the dominantly carbonate uppermost deposits of the upper member. The log of susceptibility data series was used to equalize the large variation in values of the dataset before analyses.
Time-series analyses of the log of susceptibility were computed using the MATLAB-based software Acycle (Li et al., Reference Li, Hinnov and Kump2019). The following steps were applied: (1) detrending: long-term trends in the susceptibility series were removed using linear detrending, considering the relatively stable log value of susceptibility along the section; (2) interpolation: the susceptibility series was linearly interpolated at a median sampling interval of 0.25 m; (3) spectral analysis: the 2π multi-taper method (MTM; Thomson, Reference Thomson1982) with no zero-padding was applied to evaluate the power distribution of different frequencies and compared with robust first-order autoregressive AR(1) model (Mann and Less, Reference Mann and Lees1996) to test for signal significance; the MTM evolutionary spectrum analysis with a sliding window of 30 m was also conducted to examine the variation in spectral power over depth; (4) depth-to-time conversion: the “correlation coefficient (COCO)” toolbox was used to determine the optimal sedimentation rate, enabling conversion of cyclicity from the depth domain to the time domain; (5) astronomical signal filtering: using the estimated sedimentation rates, the primary frequency peak of the potential 95- to 128-ka short-eccentricity, 41-ka obliquity, and 20- to 24-ka precession oscillation was isolated with a Gaussian bandpass filter to extract the astronomical signal; and (6) age model construction: the identified cycles were used to establish an age-depth model.
Results
Lithostratigraphy
We logged in detail the section of Fuentecica (FTC) as a previous step for the paleomagnetic analyses, obtaining a total of 46 meters of continuous stratigraphic profile (932–979 meters above sea level – m asl; Figs. 2, 3). The lowermost unit is terrigenous and reddish-grayish, with some interbedded fine-grained silty sandstones, and makes a distinctive gentle slope. This unit has a visible thickness of 26 meters and belongs to the red terrigenous member (Vera et al., Reference Vera, Fernández, López-Garrido and Rodríguez-Fernández1985). At a height of 959 meters, a distinctive, cliff-making bed is apparent, which corresponds to the bottom of the calcareous silty-calcareous member (Vera et al., Reference Vera, Fernández, López-Garrido and Rodríguez-Fernández1985), which furnishes a convenient marker bed on a decametric scale for the transition between the middle and upper members of the Baza Formation. The upper limestones include sub-metric marly layers, as well as some organic-rich, dark gray mudstones. The section is capped by a ∼7-meter-thick unit of thin-layered, tabular limestones, from where one sample for thin sectioning was collected (Fig. 3).
Fuentecica stratigraphic column (see Figure 2 for location), with the virtual geomagnetic pole latitude values, and bulk susceptibility; vertical axis indicates m asl. See text for discussion.
Overview of ESR data collected from sample 19FTC-E1. (A) Example of ESR spectrum obtained from the measurement of the natural aliquot. The red arrows indicate how the ESR intensities of the Al and Ti signals have been measured. (B–E) ESR dose response curves (DRCs) obtained for the Al (B, C), and Ti (D, E) signals. These DRCs are displayed over the full dose range (B, D) up to 25 kGy, with a focus of the first experimental points up to 5 kGy (C, E).
Microfacies
Sample 24FTC-LMST from the uppermost carbonate unit is characterized by a poorly sorted coarse component dominated by silt-sized grains of quartz and fragments of foraminifera and gastropod shells embedded in a micritic calcium carbonate groundmass (Fig. 5A). The coarse component also includes larger quartz grains (coarse- to medium-grained sand). The upper portion of the sample features horizontal carbonate fronts of variable thickness, often interrupted by voids (Fig. 5B). Porosity is limited to channels and voids caused by the welding of aggregates. Pedofeatures include coatings of secondary calcium carbonate in pores and very few iron oxide aggregates and hypocoatings (Fig. 5C; Table 1).
Photomicrographs of thin sections from Fuentecica. (A) Foraminifera shell fragments embedded in a micritic calcium carbonate groundmass. (B) Carbonate fronts marked by dark gray horizontal lines. (C) Sparitic calcium carbonate coatings in pores. (D) Horizontal beds of silt showing normal grading and clay crusts. All images in cross-polarized light except for (B). (E) Scan of the thin section of sample 24FTC-LMST showing carbonate fronts (arrows), cross-polarized light. (F) Scan of the thin section of sample 24FTC-19LD showing horizontal bedding, cross-polarized light. Oblique marks are artefacts from polishing.
Micromorphological description of the two samples from Fuentecica. The terminology follows Stoops (Reference Stoops2021).
In comparison, fine-grained sand and silt quartz grains are the main components of the coarse fraction of sample 24FTC-19LD, together with occasional shell fragments. They are closely mixed with rounded pellets of calcium carbonate crystals or entirely embedded in a pale-gray fine-grained fraction with no porosity, showing birefringence fabric typical of calcium carbonate (Table 1). In many places the sand fraction exhibits horizontal fabric and appears organized in beds. Similarly, the silt fraction is characterized by normal grading within a clay groundmass, which creates clay crusts of variable thickness (Fig. 5D).
Paleomagnetism
As noted in the introduction, sediments from the Baza Formation in the area of study are known for the generally low NRM intensities and often, magnetic instability (e.g., Oms et al., Reference Oms, Agustí, Gabàs and Anadón2000). As far as the rock-magnetic properties, progressive IRM acquisition curves reveal a dominance of magnetite as the main ferromagnetic carrier (Fig. 6A), with the typical plateau at around 200 mT in the limestones. Red siltstones show a sharp increase of IRM unit about 200 mT, and a slight increase to 1 T barely reaching saturation, whereas the gray siltstones show a slower acquisition of IRM and do not reach saturation, revealing the presence of high coercivity component, likely hematite. IRM results are consistent with the hysteresis loops, which reveal a dominance of paramagnetism in most samples, and saturation at fields of around 0.3 mT (Fig. 6B).
Rockmagnetic analysis. (A) Isothermal remanent magnetization (IRM) acquisition curves for representative lithologies, up to 1 Tesla. (B) Representative magnetic hysteresis loops from samples of different stratigraphic levels up to a maximum applied field of 1 Tesla.
We used both alternating field and thermal demagnetization procedures, including a hybrid approach. Progressive demagnetization often shows an overprint or viscous magnetization in the samples, which is removed by 10–15 mT or typically by 150–200°C. Most samples are virtually demagnetized by 580–600°C, or the intensity is too low or unstable to continue demagnetization above that level (Fig. 7). Characteristic remanent magnetization (ChRM) directions are dominantly southwards and negative (see Table S1 in SM), or northward and positive.
Representative thermal demagnetization, orthogonal vector diagrams, for samples from Fuentecica section. Closed (open) symbols represent projection of the vector end points onto the horizontal (vertical) plane. Temperatures in degrees Celsius. All plots are in geographic (in situ) coordinates.
We used the corresponding virtual geomagnetic pole latitudes (VGP Lat) of the individual ChRM directions to produce a local magnetic polarity stratigraphy (MPS), expressed in terms of normal or reverse polarity. Boundaries between polarity intervals or magnetozones are constrained to lie between adjacent sampling horizons showing opposite polarity, for convenience taken as lying half-way between the two sampled sites. Such criteria entail uncertainty (e.g., Talling and Burbank, Reference Talling, Burbank, Asissaoiu, McNeil and Hurley1993), which will be addressed later. In addition, we used a cut-off angle of 30° to establish a given magnetozone. The local MPS shows a sequence of three reverse magnetozones (R1, R2, R3) and three normal magnetozones (N1, N2, N3). The lower ∼30 meters of the section has reverse polarity (R1), whereas the top 17 meters of the section holds the three normal magnetozones (Fig. 3).
From the three normal magnetozones, N1 may be reasonably questioned when polarity boundaries and lithostratigraphy are compared. The polarity boundaries, both bottom and top, coincide with visible changes in lithology (Fig. 3). Generally speaking, polarity reversals that are consistent with lithological changes should cast doubts on the age of the magnetization (e.g., Langereis et al., Reference Langereis, Krijgsman, Muttoni and Menning2010), because changes can be produced by constrained remagnetization due to contrasting grain size, porosity, or other physical properties. Therefore, even though we do not observe a significant change in bulk magnetic susceptibility (Fig. 7), such magnetozones will not be considered as a true magnetozones in our interpretation. Consequently, the local magnetic polarity will be based on a succession of four magnetozones, with, from bottom to top, R1 and R2 forming a single interval of reverse polarity, N2, R3, and the uppermost N2 (Fig. 3).
The reliability of the dataset has been tested through the so called reversal test, meant to determine whether two sets of paleomagnetic directions, one with normal polarity and the other with reverse polarity, could have been drawn from distributions with mean directions 180°apart (e.g., McFadden and McElhinny, Reference McFadden and McElhinny1990). Indeed, it tests whether magnetizations are contaminated with another component, because both normal and reverse directions would be shifted towards the direction of the contaminating component, and they would no longer be antipodal. In order to carry out the reversal test (Fig. 8A) the reversed polarity mode has been flipped to the antipode in order to compare the two populations of bootstrapped directions, converting the populations to Cartesian coordinates (Tauxe, Reference Tauxe1998). The intervals containing the 95th percentile of the three components overlap, revealing that the reversed and normal modes cannot be distinguished at the 95% confidence level and therefore they pass the bootstrap reversal test (Tauxe, Reference Tauxe1998).
(A) Reversal test. The reversed polarity mode has been flipped to the antipode in order to compare the two populations of bootstrapped directions, converting the populations to Cartesian coordinates (Tauxe, Reference Tauxe1998). The intervals containing 95% of the three components overlap, revealing that the reversed and normal modes cannot be distinguished at the 95% confidence level and therefore they pass the bootstrap reversal test (Tauxe, Reference Tauxe1998). (B) Magnetostratigraphic jackknife analysis for the studied sections. The plot indicates the relationship between average percent of polarity zones retained and the percentage of sampling sites deleted (Tauxe, Reference Tauxe1998). The slope of the line gives the magnetostratigraphic jackknife parameter J. See text for details.
Further reliability of the obtained local magnetic polarity has been quantified using the jackknife procedure adapted for magnetostratigraphy by Tauxe and Gallet (Reference Tauxe and Gallet1991). The jackknife is a statistical technique that relies on progressive deletion of data points in order to estimate a parameter of interest, in this case the magnetic polarity. The jackknife adapted for magnetostratigraphy successively deletes greater proportions of the data (in our case VGP latitudes) and recounts the number of remaining polarity zones. The process is repeated until 20% of the data have been deleted. The average percentage of the polarity zones remaining after each deletion is plotted versus the percentage of sampling sites deleted (Fig. 8B). The slope of the line generated by this procedure (called J) is directly related to the sensitivity of the sampled stratigraphy on the particular site distribution (Tauxe and Gallet, Reference Tauxe and Gallet1991). The value of the jackknife parameter J is therefore an objective estimate of the reliability of the sampled section. Based on numerous simulations, Tauxe and Gallet (Reference Tauxe and Gallet1991) recommend J values in the range of 0 to −0.5 for a robust magnetostratigraphic dataset, whereas J values below −0.7 would reflect magnetostratigraphy that is not well determined. We applied the jackknife technique (software in Tauxe, Reference Tauxe1998) to Fuentecica and obtained a J value of −0.27, which would indicate that the section has almost recorded the majority of the true number of polarity units.
ESR dating of quartz grains
ESR data
With a variation of the DE estimates < 5% (Table 2), the measurement repeatability achieved for the Al signal falls within the higher range of usual standards (Duval et al., Reference Duval, Arnold, Bahain, Parés, Demuro, Falguères and Shao2024). A minimum S/N ratio of 59.6 was estimated from the natural aliquot. While this relatively low value compared with other samples, which may typically show S/N > 100 (e.g., Duval et al., Reference Duval, Arnold, Bahain, Parés, Demuro, Falguères and Shao2024), is probably related to the nature and composition of the quartz being dated, it is nevertheless sufficient to exclude any influence from the high-frequency background noise on the ESR intensities and associated dose estimates. Fitting performed with the EXPLIN function with data weighting by 1/I2 and 1/s2 return very close DE estimates, with a systematically high goodness-of-fit (adjusted r 2 > 0.99; see also Fig. 3). A difference of about 450 Gy may be observed between the dose values calculated with and without subtracting the residual, unbleachable ESR intensity (Table 2).
ESR data derived from the measurement of the Al and Ti (option D) centers. Bleaching coefficient is expressed as the relative difference between the ESR intensities of the natural and bleached aliquots. Repeatability of the ESR intensities is assessed through the variability (1 relative standard deviation) of the mean ESR intensities obtained after each day of measurements. Similarly, the repeatability of the DE values corresponds to the variability (1 relative standard deviation) of the DE values calculated for each day of measurement.
1 Average S/N value calculated from the natural aliquot (Time 1).
2 Fitting performed with the EXPLIN function with data weighting by 1/I2 and without subtracting the residual, non-bleachable, ESR intensity.
3 Bleaching coefficient is expressed as the relative difference between the ESR intensity of the natural and UV-bleached aliquots.
4 Value obtained after subtraction of the residual ESR intensity.
In contrast, ESR measurements of the Ti signal show a repeatability 16.3% (i.e., below usual standards; Duval et al., Reference Duval, Arnold, Bahain, Parés, Demuro, Falguères and Shao2024). A minimum S/N ratio of 7.8 was calculated for the natural aliquot, showing the relatively high ESR intensities compared to the high-frequency background noise Similar to the Al signal, a high goodness-of-fit is systematically achieved (adjusted r 2 > 0.99), while the DE estimates are virtually not affected by the data-weighting option selected (i.e., they differ by ∼3%; Table 2 and Fig. 4).
ESR intensities of the Al and Ti signals and their respective ratios may also be used for characterization purpose (e.g., Demuro et al., Reference Demuro, Arnold, Duval, Méndez-Quintas, Santonja and Pérez-González2020). For completeness, the ESR intensities of the Ti signals in the natural aliquot also were measured following options A, C, and E, sensu Duval and Guilarte (Reference Duval and Guilarte2015; i.e., in addition to the option D evaluated for dating purpose). Options A and D represent a mixture of Ti–Li and Ti–H lines, while options C and E represent pure Ti–H and Ti–Li contributions (Duval and Guilarte, Reference Duval and Guilarte2015). Interestingly, natural quartz sample 19FTC-E1 is characterized by relatively high Ti signal intensity, or low Al intensity, as illustrated by the relatively small Al/Ti(A) ratio of 5.2 (Table 2). This might indicate a relatively low radiation sensitivity for the Al signal. A C/D ratio of 0.54 was calculated, which is within the upper range of values (0.26–0.55) published earlier by Duval and Guilarte (Reference Duval and Guilarte2015), and about half-way between the extreme values obtained for the Porto Maior (0.30) and Cuesta de la Bajada (0.70) samples (Demuro et al., Reference Demuro, Arnold, Duval, Méndez-Quintas, Santonja and Pérez-González2020). This shows a relatively high contribution of the Ti–H line in option D. However, the direct effect on the dose evaluation remains limited, since all options made either of pure Ti–Li (option E) or a mixture of Ti–Li and Ti–H (options A and D) return close-fitting Ti-2 (data weighting by 1/s2) results, confirming their robustness: 1018 ± 59 Gy, 995 ± 54 Gy, and 1038 ± 88 Gy for options A, D, and E, respectively.
To summarize, ESR data obtained for both Al and Ti signals overall show high robustness and reliability, with regular to limited measurement repeatability, but high goodness-of-fit, and no significant fitting bias observed. A dose estimate of about 970 Gy may be obtained for the Ti signal, while the Al signal returns a maximum possible dose of about 1150 Gy, reduced to about 700 Gy after subtracting the residual component of the signal (Table 2). Consequently, the Al DE value obtained is lower than the Ti dose estimate, which cannot be explained by any kind of relatively incomplete bleaching of the Ti signal, given its much faster bleaching kinetics (e.g., Toyoda et al., Reference Toyoda, Voinchet, Falguères, Dolo and Laurent2000; Duval et al., Reference Duval, Arnold, Guilarte, Demuro, Santonja and Pérez-González2017). Saturation of the Al signal can be reasonably excluded given the dose response curve of 19FTC-E1, and the absence of apparent saturation usually observed in the Al signal at irradiation doses of several tens of ka (e.g., Lin et al., Reference Lin, Yin, Ding, Cui, Chen, Wu and Xu2006; Benzid et al., Reference Benzid and Timar-Gabor2020). Limited thermal stability or radiation sensitivity of the Al signal as an explanation for a potential dose underestimation are perhaps the most plausible option in the first instance (see Duval et al., Reference Duval, Fujioka, Haddoumi, Parés, Aouraghe, Churruca Clemente and Lachner2025, and references therein).
Dose rate evaluation
ICP-MS/OES analyses of the raw sediment returned radioelement concentrations of 1.33 for U, 4.30 for Th, and 0.72%, providing a laboratory gamma dose rate of 490 µGy/a when using the measured water content (Table 3). This value is only 3% lower than the value measured in situ. This excellent agreement between the two values not only illustrates the homogeneity of the radioactivity in the immediate surrounding of the sample around the sample, but also indirectly indicates the absence of any significant disequilibrium in the 238U decay chain of the sediment.
Radioelement concentrations measured from the raw sediment by ICP-MS analyses. Laboratory gamma dose rate was calculated using the dose rate conversion factors from Guérin et al. (Reference Guérin, Mercier and Adamiec2011), the measured water content, and the radioelement concentrations. In-situ gamma dose rate value has been derived from the threshold approach.
ESR age calculations
Initial age calculation performed using a long-term water content of 40%, which corresponds to ∼60% of the saturated water content determined in the laboratory, return 738 ± 99 ka and 1011 ± 72 ka for the Al and Ti signals, respectively. In order to evaluate the effect of water content on the calculated age result, we carried out age sensitivity tests from 7% of water content (= measured content) to 70% (= saturated water content). Age results range from 546 ± 72 ka to 908 ± 121 ka for the Al signal, and from 769 ± 53 ka to 1277 ± 92 ka for the Ti signal (Table 4; Fig. 9).
ESR age sensitivity tests using long-term water content ranging from 7% (% dry weight; measured value) to 70% (saturated water content). Numerical values are found in Table 4.
Detail of ESR age estimates and dose rate components. Errors are 1σ. Note that the DE errors are a combination of the fitting errors (from Table 1) and of the gamma source dose rate (2.3%).
The age difference between the two ESR signals varies from 222 ka to 370 ka. Interestingly, all the Al ESR ages suggest a Middle Pleistocene age (< 0.77 Ma), or a late Early Pleistocene, post-Jaramillo (< 1 Ma) chronology when considering an extreme scenario with high water content (> 40%). These age results are hardly incompatible with the magnetostratigraphic evidence positioning the sample within an interval of reverse polarity, about 25 m below a normal interval, suggesting an age much older than the minimum age constraint of either 0.77 Ma or 1.08 Ma, depending on the preferred correlation of the MPS to the GPTS (see discussion). This may have significant implications from a methodological perspective, because it shows that the Al signal does not always provide a maximum age given its slower bleaching kinetics, as traditionally interpreted in accordance with the MC approach. Instead, there is clear evidence here that the Al signal may sometimes provide underestimated ages. In contrast, Ti ESR ages are about 41% older than the Al results and are by far more compatible with a pre-Jaramillo chronology (> 1.08 Ma) for long-term water content > 40%.
Cyclostratigraphy
The 2π MTM power spectrum analysis of the linearly detrended log of susceptibility series identified significant cycles with thicknesses of 23.5 m, 6.89 m, and 4.54 m, with only the 4.54 m cycle exceeding the 95% robust AR(1) significance level (Fig. 10A). Frequencies corresponding to cycles longer than 27.0 m (below 0.037 1/m) were excluded from consideration because they exceed the sampled section thickness. The evolutionary power spectrum analysis reveals that the 23.5-m cycle is stable throughout the portion of the analyzed section, whereas the 6.89-m and 4.54-m cycles are more prominent in the lower part of the section than in the overlying strata (Fig.10). The COCO analysis resulted in two potential sedimentation rates with significant levels >99.9%: 4.7 cm/ka (95% significant interval: 3.4–6.5 cm/ka) and 18.7 cm/ka (95% significant interval: 14.6–23.4 cm/ka) (Fig. 10).
(A) The MTM spectral analysis, MTM evolutionary spectral analysis, filtering of 125-ka eccentricity, 41-ka obliquity and 24-ka precession and corresponding age-model of the log of susceptibility. Pink squares illustrate the three significant frequencies (cycles) in the susceptibility series. The sliding window of evolutionary power spectrum is 30 m. The time of each cycle is determined by the correlation coefficient (COCO) sedimentation rate estimation. (B) COCO analysis of the log of susceptibility. (C) Correlation coefficient values for various potential sedimentation rates. (D) Null hypothesis (H0) significance levels of sedimentation rates. (E) The number of contributing astronomical parameters involved. The target series corresponds to the La2004 astronomical solution at a middle age of 1.2 Ma. Significance levels were estimated using a Monte Carlo simulation of 500 iterations.
Discussion
The VGP latitudinal sequence at Fuentecica reveals a dominantly reverse polarity (R1–R2) covering about 75% of the total thickness of the deposits, two complete normal (N2) and reverse (R3) polarity magnetozones, and an incomplete uppermost normal magnetozone (N3) (Fig. 3). To interpret such local magnetic polarity stratigraphy, we used the ESR age results as a first approximation for anchoring to the Geomagnetic Polarity Time Scale (GPTS). While the overall section falls in the upper part of the Matuyama Chron (2.61–0.77 Ma), as suggested by the dominantly reverse polarity magnetozone R1, which is consistent with previous magnetostratigraphic studies of the Baza Formation deposits, the main question is the attribution of normal magnetozones N2 and N3 to any known normal Subchrons of the GPTS.
On the presence/absence of Olduvai Subchron at Fuentecica
Various previous magnetostratigraphic studies in the Orce-Venta Micena area have already led to the identification of a normal polarity magnetozone tentatively attributed to the Olduvai (1.93–1.78 Ma) (e.g., Scott et al., Reference Scott, Gibert and Gibert2007), or the Jaramillo Subchron (1.08–1.01 Ma) (e.g., Gibert et al., Reference Gibert, Scott and Deino2024), despite the absence of direct independent age control from numerical dating methods. As stated earlier, the Baza Formation in this area has typically provided a dominance of reverse polarity chrons, as expected for a Lower Pleistocene unit. In fact, the iconic section of Barranco León located a few kilometers east has provided exclusively reverse polarity (Oms et al., Reference Oms, Agustí, Gabàs and Anadón2000), a stratigraphic section that encompasses the middle and upper members of the Baza Formation, like Fuentecica.
So far, one magnetostratigraphic section covers part of the lower member near the study area, and is placed farther north, towards the depocenter of the Baza paleolake section. The paleomagnetic section is known as Rincón de la Losa. Along this 26-meter long stratigraphic section, Scott et al. (Reference Scott, Gibert and Gibert2007) reported a 10-meter-thick normal magnetozone, which is interpreted as the Olduvai Subchron based on lithostratigraphic and biostratigraphic evidence. Similarly, a ∼5-meter-thick section of red clays with gypsum at the locality known as Mina La Mojona, which are visible at the talweg of the Cañada de Velez valley (Fig. 1), also produced normal polarity directions, which Scott et al. (Reference Scott, Gibert and Gibert2007) assigned to the Olduvai Subchron.
The stratigraphic section of Cullar Baza, 20 km to the SSW, offers a record encompassing from the Olduvai Chron to the base of the Brunhes Chron along an 80-meter-thick profile (Gibert et al., Reference Gibert, Ortí and Rosell2007). Unfortunately, a physical correlation between these sections is not possible due to the distance. In fact, evident lateral facies changes and normal faults within the area (e.g., Gibert et al., Reference Gibert, De Galdeano, Alfaro, Scott and Garrido2005, Reference Gibert, Ortí and Rosell2007; Oms, Reference Oms, Agustí, Gabàs and Anadón2000) hinder any solid correlation between the stratigraphic sections.
In order to anchor the Fuentecica section to the GPTS, we have to consider the position of the alleged Olduvai Subchron in Mojona section, about 2.8 kilometers east (Scott et al., Reference Scott, Gibert and Gibert2007). In the La Mojona stratigraphic section (∼5 m) the lowermost normal polarity horizon is at around 925 m asl, whereas the base of N2 at FTC is positioned at around 968 m asl (i.e., about 43 m above; Fig. 11). This significant difference in the elevation between both normal intervals hardly allows interpreting N2 as the Olduvai Subchron. In contrast, the base of the studied Fuentecica section is found at an elevation of 932 m asl, slightly higher than the normal magnetozone identified at La Mojona (Fig. 12). Consequently, we find it quite reasonable to infer that the Fuentecica section post-dates the Olduvai Subchron—an assumption that is coherent with the ESR results obtained here, which systematically return ages younger than 1.8 Ma, independently of the long-term water content used (Table 3 and Fig. 9).
Age-depth model established for the Fuentecica sedimentary sequence. The colored lines correspond to the extrapolated sediment accumulation rate (SAR) values inferred for the bottom and top of the sequence. In the middle member, the SAR value (red line) is calculated from cyclostratigraphy analysis. For the upper member SAR values (green and blue) are computed using the age of Cobb Mountain and the base of Jaramillo Subchron, respectively. Solid and dashed SAR lines represent interpolated and extrapolated values, respectively.
Topographic map (A) and profile (B) with the locations of the main magnetostratigraphic sections in the Orce-Cañada de Vélez area (based on Gibert et al., Reference Gibert, Scott and Deino2024), including Orce D (OD; Gibert et al., Reference Gibert, Scott and Deino2024), Fuentecica (FTC; this study), Barranco León (BL; Oms et al., Reference Oms, Agustí, Gabàs and Anadón2000), Mina Mojona (MM; Scott et al., Reference Scott, Gibert and Gibert2007), and Fuente Nueva-1 (FN1; Scott et al., Reference Scott, Gibert and Gibert2007). For readability, the position of the FN1 section has been projected on the profile (white arrow). Additional key sites: Fuente Nueva-3 (FN3), Venta Micena (VM).
GPTS correlation option 1 for N3 and N2
The option of correlating the incomplete normal magnetozone at the top of the sequence (N3) to the Brunhes Chon (0.77–0.00 Ma) is quite tempting. Such an option would lead to the attribution of the normal magnetozone N2 below to the normal Jaramillo Subchron (1.08–1.01 Ma), a correlation still compatible with the Ti ESR age of 1.039 ± 77 ka (1σ) obtained from deposits stratigraphically located about 25 m below (Table 4). When considering the upper 2σ error range of this result, it would provide a maximum age constraint of 1.183 Ma for the deposits above. Moreover, given the uncertainty on the long-term water content, an older maximum age constraint cannot be reasonably discarded, reinforcing the reliability of this correlation (Fig. 9). However, assigning magnetozone N3 to the Brunhes Chron would make the uppermost layers younger than 0.77 Ma, so within the Middle Pleistocene. We do acknowledge that such correlation is in conflict with the overall chronostratigraphic interpretation of Baza Formation, established through decades of research in the area, as Early Pleistocene. On the other hand, the possibility that the uppermost carbonates may not belong to the Baza Formation but are associated with younger post-capture deposits related to the glacis formation can be reasonably excluded.
Petrographical analysis (section “Microfacie”) shows sedimentary properties that are typical of the Baza Formation. Sample 24FTC-LMST shows foraminifera shell fragments embedded in the carbonate, and the overall texture and grain size are consistent with a low-energy water environment subject to periodic input of silicate clasts, presumably windblown given their scarcity and poor sorting. The absence of bedding and the presence of fossils and carbonate fronts indicative of surface exposure, suggest that the sediment formed in a shallow lacustrine depositional environment. In summary, the uppermost limestones around Fuentecica creek appear to be part of the sedimentary infill and not related to carbonates related to the glacis development that postdates the lacustrine sedimentary infill (e.g., Garcia-Tortosa et al., Reference García-Tortosa, Alfaro, Martin-Rojas, Medina-Cascales and Giralt2024, and references therein).
GPTS correlation option 2 for N3 and N2
An alternative option is to assign the uppermost N3 normal interval to the Jaramillo Subchron (1.08–1.01 Ma) and therefore the N2 to the preceding normal subchron in the GPTS (i.e., Cobb Mountain; 1.22–1.18 Ma). This correlation nicely meshes with the recent study by Gibert et al. (Reference Gibert, Scott and Deino2024) of a section next to the village of Orce, about 1 km WSW of Fuentecica, where the Jaramillo Subchron has been identified at an elevation of 990 m asl (OD section in Gibert et al., Reference Gibert, Scott and Deino2024) (Fig. 12). In the studied section of Fuentecica, the base of Jaramillo is at 978–975 m asl (Fig. 7), just slightly lower than in OD, which would suggest a slight, yet significant over several hundreds of meters, apparent dip to the north of the sedimentary infill. In fact, Santamaría Barragán et al. (Reference Santamaría Barragán, Parés, Colino García, Churruca Clemente, Fujioka, Serrano-Ramos and Toffolo2025) reported an apparent dip of 0.5° to the SW at Barranco León creek. Such minimum dip, if constant, would translate into a height of about 9 meters over one kilometer, which is the distance between the Orce and Fuentecica sections. Therefore, such slight tilt of strata can reasonably explain the difference in elevation of the polarity chrons around the area.
Assigning N3 and N2 to the Jaramillo and Cobb Mountain subchrons, respectively, implies a minimum age of 1.22 Ma to the sediments located farther down in the section, including those dated by ESR. While this minimum age constraint seems inconsistent with the Ti ESR age estimate of 1.039 ± 0.077 Ma, even when considering a 2σ error range, we cannot exclude that this may indirectly support the hypothesis of a higher long-term water content than initially considered (i.e., > 40%; Fig. 9), which is, on the other hand, compatible with micromorphological evidence.
The texture of the sampled siltstones as revealed by the thin section adds important information as far as water content. We observed microstructures that range from massive to intergrain microaggregate with complex packing voids. The coarse fraction is composed of fine-grained sand to silt, sub-angular quartz grains (dominant); coarse- to medium-grained sand, sub-angular quartz grains (few); rounded pellets of calcium carbonate (frequent); and fragmented calcium carbonate shells (common) (Fig. 5). The studied siltstones represent relatively high energy environments, due to the larger grain size and the occurrence of horizontal sand beds. The presence of shells in the bedded sand also may be interpreted as periodic input of shallow lake waters into a stream system. The flow regime varied over time, as indicated by the normal grading of the silt beds alternated with clay crusts. Moreover, in several places the quartz sand and calcium carbonate pellets are cemented by a micritic carbonate matrix, which could be either interpreted as a post-depositional process involving waterlogging or related to a single event of cementation of the sand soon after deposition (if not during deposition itself as part of the mud fraction). Overall, these features suggest that the sediments most likely retained water over an extended period of time. Consequently, the hypothesis of protracted presence of water, as also suggested by the thickness of the lacustrine deposits, seems supported. When considering a long-term water content closer to the saturation level (i.e., > 40%), the resulting ESR ages are increasingly compatible with the minimum age constraint of 1.22 Ma given by the base of N2: with maximum 2σ upper range of 1.28 Ma, 1.37 Ma, and 1.46 Ma calculated for 50%, 60%, and 70% water content, respectively (Fig. 9; Table 4).
We already discussed the lack of reliability of polarity chron N1, but it is worth noting the implications should it be a real magnetozone. In such a case, the succession of three normal polarity chrons would lead to the interpretation of Cobb Mountain, Jaramillo, and Brunhes chrons from bottom to top. As a consequence, the top portion of the stratigraphic section would have had a slower sedimentation rate, as it would terminate into the base of the Brunhes chron. On the other hand, such an interpretation would hardly fit the interpretation of the existing magnetostratigraphic sections in Orce (Gibert et al., Reference Gibert, Scott and Deino2024) and Barranco León (Oms et al., Reference Oms, Agustí, Gabàs and Anadón2000).
Sediment accumulation rates
Sediment accumulation rates (SAR) may be estimated for the middle and upper members by two different approaches. On the one hand, for the upper part of the Fuentecica section, N3 and N2 may be attributed to the Jaramillo (1.08–1.01 Ma) and Cobb Mountain (1.22–1.18 Ma) subchrons, respectively, based on the most reasonable GPTS correlation option (2) discussed previously. Given the chronostratigraphic tie points available (1.08 Ma, 1.18 Ma, 1.22 Ma) and the thickness of the magnetozones, SAR values may be estimated to 5 cm/ka for R3 and 9 cm/ka for N2, respectively (Fig. 11). While the former and the latter can reasonably be extrapolated towards the top of the section and the base of the upper member given the overall homogeneity of the lithology in the unit, any further extrapolation through the middle member, which shows very distinct siliciclastic deposits, would be quite speculative.
However, in the absence of magnetic reversals, there is no chronostratigraphic tie point for the lower part of the section, and no SAR value can be interpolated from the paleomagnetic data. Instead, such an estimate can be obtained from the cyclostratigraphic study of the middle member, which yielded two possible SAR estimates. The 18.7 cm/ka sedimentation rate converts the three depth-domain cycles (23.5 m, 6.89 m, and 4.54 m) into time-domain cycles of 125 ka, 36.8 ka, and 24.3 ka, respectively. These cycle durations align perfectly with the astronomical cycles of short eccentricity (125 ka), obliquity (41 ka), and precession (24 ka), respectively (Fig. 10). In contrast, the 4.7 cm/ka sedimentation rate yields 500-ka, 145-ka, and 97-ka cycles, which deviate significantly from the expected long eccentricity (405 ka) and short eccentricity (95–125 ka) cycles, making this option less likely.
To sum up, our results show that the SAR of the deposits along the Fuentecica sequence is clearly not constant. Instead, a continuous and significant decrease is observed from the base to the top of the section, with ∼18.7 cm/ka for the middle member, and 9–5 cm/ka for the upper member. Such results on the gross accumulation rate, and in particular the progressive decrease up section, agree with previous studies on the evolution of the sedimentation rate both at the basin level (Viseras, Reference Viseras1991) and in other sectors of the basin (e.g., Calvache and Viseras, Reference Calvache and Viseras1997; Pla-Pueyo et al., Reference Pla-Pueyo, Viseras, Soria, Tent-Manclús and Arribas2011).
In other words, the SAR values from the middle and upper members differ by a factor of 2, which is compatible with lithological evidence showing dominantly siliciclastic and carbonate deposits in the lower and upper halves of the sequence, respectively. To our knowledge, this is the first time that such variability is documented within one the Early Pleistocene sections in the Orce area. This indirectly suggests that age–depth models based on sub-linear SAR values for these deposits should be treated with caution. In that regard, our results differ from those previously reported by Gibert et al. (Reference Gibert, Scott and Deino2024), who assumed an almost constant SAR value of 6.7 cm/ka for the Early Pleistocene deposits of the Orce area between the Olduvai and Jaramillo subchrons. While this value is roughly consistent with our results for the upper member deposits, it is significantly lower than our calculation of ∼18.7 cm/ka for the middle member.
These SAR values enable us to roughly estimate the time covered by the Fuentecica sequence. First, by extrapolating a SAR value of 9 cm/ka to the base of the upper member, the approximate age of the transition between the middle and the upper members of the Baza Formation may be constrained to ca. 1.3 Ma (Fig. 11). Second, based on the SAR value of ∼19 cm/ka for the terrigenous formation, and the identification of six 24-ka precession cycles, the duration of the 27-m succession studied for cyclostratigraphy is estimated to be 145 ka (Fig. 10). Consequently, the base of the section, at an elevation of 932 m asl, would have a chronology of around 1.45 Ma, an age that is still consistent with the lower age range produced by the ESR Ti-signal (Figs. 9 and 11), and a post-Olduvai chronology for the section. Finally, the correlation of the uppermost N3 to Jaramillo subchron provides a minimum age of 1.01 Ma for the sequence.
Interestingly, it may be noted that this significant variation in the SAR value observed along the sequence makes the whole data set more coherent and robust: while a SAR value of 9 cm/ka extrapolated through the middle member would not have intersected the 2σ ESR age range, the SAR value of ∼19 cm/ka is instead compatible with a modeled ESR age of ca. 1.40 Ma (Fig. 11). This value falls within the 2σ error range calculated for a long-term water content of at least 60%. Importantly, any increase of the SAR value would lead to a younger modeled ESR age. For example, the maximum SAR value of 23 cm/ka derived from the 2σ upper error range estimated by the COCO analysis would yield a younger modeled age of ca. 1.38 Ma. In other words, ESR dating and cyclostratigraphy provide results that are compatible when considering associated uncertainties.
Conclusions
The Fuentecica section, which encompasses the red terrigenous (middle) member and the silty-calcareous (upper) member of the Baza Formation, includes 46 meters of alluvial, lacustrine, and palustrine sediments that reveal the final stages and filling of the eastern Baza Basin during the Pleistocene. According to our combined paleomagnetism, ESR, and cyclostratigraphy results, the entire stratigraphic section falls within the Early Pleistocene, which agrees with the nearby section of Orce (Gibert et al., Reference Gibert, Scott and Deino2024). When combining ESR ages obtained from quartz grains with magnetic reversal stratigraphy and cyclostratigraphy, the most parsimonious interpretation is that the section spans from ca. 1.5 Ma (Matuyama post-Olduvai) to ca. 1 Ma (intra-Jaramillo). Even though the Jaramillo Subchron (1.08–1.01 Ma) has recently been reported in a nearby section (Gibert et al., Reference Gibert, Scott and Deino2024), the Cobb Mountain subchron is observed for the first time in the Baza Basin, hence providing an additional timeline for intra-basin correlation irrespective of lateral facies changes.
Another corollary of the chronostratigraphic results is that the biostratigraphic locality of Fuentecica 5 (FTC-5 in Fig. 2A), which corresponds to a blackish mudstone rich in arvicolid rodents (Agustí et al., Reference Agustí, Piñero, Lozano-Fernández and Jiménez-Arenas2022; Agustí and Piñero, Reference Agustí and Piñero2023) and located just 300 meters farther north (Fig. 2A), falls right below magnetozone N2, constraining the age of the locality between 1.3 Ma and the bottom of Cobb Mountain, or 1.2 Ma.
A corollary of our results is that the assumption of horizontal, continuous strata over kilometric distances in the basin is not sustained (Fig. 12). Our study shows the bottom of the Jaramillo Subchron at a height of 978–975 m asl, whereas in the section of Orce, about one kilometer to the WSW, the bottom of the Jaramillo Subchron is found at 990 m asl (Fig. 11), revealing a subtle tilt of the strata that is imperceptible to the naked eye.
From a methodological point of view, the dating results obtained at Fuentecica may have significant implications in ESR dating, because they show that the Al signal may sometimes return underestimated age results. Consequently, the usual assumption that the Al signal provides a maximum possible burial age for a given quartz sample, given its slower bleaching kinetics, should not be systematically taken for granted, especially when the Al signal returns younger results than the Ti signal in a given quartz sample, such as is the case in the present study.
Finally, our results document for the first time in the Orce area a significant variation of sediment accumulation rate along a given sequence, and especially between the middle and upper members of the Baza Formation. Importantly, such variation is not only coherent with the lithology but also with the independent age control given by ESR dating. In conclusion, the various lines of evidence collected from different independent, and yet complementary, methods, enable us to not only establish a robust and coherent chronostratigraphic framework for the Fuentecica section, but also contribute to a better understanding of the sedimentary dynamics in the Orce area.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/qua.2025.10041.
Acknowledgments
Investigations in the Orce Archaeo-paleontological Zone (OAZ) are performed within the framework of the Regional Research Project (PGI) ‘Evolución humana y paleoecología a partir de los yacimientos de la zona arqueológica ‘Cuenca de Orce’ Retos y desafíos, ReDes-ORCE funded by Junta de Andalucía. This work is part of grants: PID2021-123092NB-C22 and PID2021-122355NB-C33 funded by MCIN/AEI/ 10.13039/501100011033/’ERDF A way of making Europe, and Grant PID2024-155126NB-I00 funded by MICIU/AEI/10.13039/501100011033 and by ERDF/EU. Mathieu Duval’s investigation is currently supported by the Spanish Ramón y Cajal Fellowship RYC2018-025221-I, which is funded by MCIN/AEI/10.13039/501100011033 and by ‘ESF Investing in your future’. Michael Toffolo is supported by grant RYC2021-030917-I funded by the MCIN/AEI/10.13039/501100011033 and by the European Union NextGenerationEU/PRTR. Alexia Serrano-Ramos and Juan Manuel Jiménez-Arenas are supported by grant PID 2021.125098NB.I00 funded by MCIN/AEI/10.13039/501100011033/FEDER (‘Una manera de hacer Europa’), and ProyExcel_00274 funded by ‘Dirección General de Planificación de la Investigación (Consejería de Universidad, Investigación e Innovación, Junta de Andalucía)’; PID2022-140850OB-C21/PID2022-140850OB-C22 from MICIU/AEI/10.13039/501100011033 and ERDF/EU. Additional support was provided by the RNM 369 Group of the Andalusian Research and Innovation Plan (PAIDI). We also thank Maria Jesús Alonso Escarza, and Cristina García Rodríguez, CENIEH, for invaluable assistance with the ESR dating analyses. We would like to thank Conall MacNiocaill and an anonymous reviewer for taking the time and effort necessary to review the manuscript. We sincerely appreciate all the valuable comments and suggestions, which helped us to improve the quality of the manuscript.
Author contributions
J.M. Parés, D. Anastasio, C. Berti: conceptualization, formal analysis, investigation, writing, supervision. A. Santamaría: sample preparation, data acquisition, formal analysis, investigation, writing. S. Colino García, C. Saiz: sample preparation, data acquisition. L. Hongliang: time series analysis. M. Toffolo: petrographic analyses. L.M. Yeste: drone surveying. T. Fujioka: investigation, validation. J.M. Jiménez Arenas, M. Duval: conceptualization, formal analysis, investigation, writing, funding acquisition.
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
