Rock Magnetic Cyclostratigraphy of Permian Loess in Eastern Equatorial Pangea (Salagou Formation, South-Central France)

We present the findings from analysis and modeling of a stratigraphic series of magnetic susceptibility (MS) data measured with a portable MS meter from the Permian Salagou Formation loessite (south-central France). The results reveal discernible Milankovitch-scale paleoclimatic variability throughout the Salagou Formation, recording astronomically forced climate change in deep-time loessite of eastern equatorial Pangea. Optimal sedimentation rates are estimated to have ranged between 9.4 cm/kyr (lower Salagou Formation) and 13 cm/kyr (mid-upper Salagou Formation). A persistent 10-m-thick cyclicity is present that likely represents orbital eccentricity-scale (∼100 kyr) variability through the middle to late Cisuralian (ca. 285—275 Ma). Subordinate, higher frequency cycles with thicknesses of ∼3.3–3.5 and ∼1.8 m appear to represent obliquity and precession-scale variability. If the driver of magnetic enhancement is pedogenic, then the ∼10 m thick cyclicity that is consistent over ∼1000 m of section may represent the thickness of loessite–paleosol couplets in the Salagou Formation. Laboratory rock magnetic data show generally low magnetic enhancement compared to analogous Eurasian Quaternary loess deposits. This is related to the predominance of hematite (substantially weaker signal than magnetite or maghemite) in the Salagou Formation which may be explained by different conditions of formation (e.g., syn depositional processes, more arid, and/or oxidizing climate conditions) than in present Eurasia and/or post depositional oxidation of magnetite and maghemite.

Thick loessite deposits have also been associated with the Late Paleozoic Ice Age (e.g., Murphy, 1987;Johnson, 1989;Kessler et al., 2001;Mack and Dinterman, 2002;Soreghan et al., 2002;Tramp et al., 2004;Soreghan and Soreghan, 2007;Soreghan et al., 2008a,b;Soreghan M. J. et al., 2014;Giles et al., 2013;Sweet et al., 2013;Foster et al., 2014;Pfeifer et al., 2020). Voluminous paleo-loess accumulated in low latitudes of western and eastern Pangea and are exposed today in the western United States (exceeding 700 m thick) and western Europe (up to ∼1500 m thick in southern France). Loessitepaleosol intercalations in the deep-time record of western equatorial Pangea have been posited to represent icehouse glacialinterglacial oscillations on a 100 kyr-scale (Soreghan et al., 1997(Soreghan et al., , 2002Soreghan M. J. et al., 2014;Tramp et al., 2004), as well as higher frequency climatic signals occurring on obliquity and possibly precession scales . Recently, a radioisotopically-dated lower Guadalupian cyclic lacustrine record from the Junggar Basin of eastern Pangea (30-32 • N paleolatitude) was discovered to exhibit astronomical frequencies (Huang et al., 2020). Importantly, this new evidence supports the hypothesis explored in this work for presumably analogous Milankovitch cycles recorded in the thick (∼1000 m) record of upper Paleozoic loess in eastern equatorial Pangea.
The objectives of this work are to (1) document Permian climatic variability and assess Milankovitch characteristics using MS data from the Salagou Formation loessite obtained with a portable MS meter, (2) leverage laboratory rock magnetic measurements to better understand the origin of the magnetic signal, and (3) compare this upper Paleozoic record with analogous successions from the Eurasian Quaternary record.

GEOLOGICAL SETTING
The Permian Salagou Formation is an exclusively fine-grained and internally massive deposit of red mudstone with local pedogenic overprinting and water-reworked interbeds that accumulated in the Lodève continental rift basin at equatorial latitudes (0-3 • N; Domeier and Torsvik, 2014;Muttoni and Kent, 2019). Its characteristics are most consistent with eolian transport and deposition either as loess, or as dust deposited in a seasonal, shallow lacustrine environment (Pfeifer et al., 2020). The provenance of the Salagou Formation is the Montage Noire Dome, a local uplift of Variscan basement composed of gneiss, schist, and granite (Figure 1; Pfeifer et al., 2016). The recognition of great volumes of loess-sized silt, together with known coarsegrained granitic protoliths and deposition during global icehouse conditions, are all most consistent with a hypothesis of upland glaciation in this region (Pfeifer et al., 2020).
The late Artinskian to mid-Kungurian (upper Rotliegend I; Kiersnowski, 2013) age of the lower Salagou Formation (Octon Member) is well constrained by U-Pb ID-TIMS dates from ash layers (Michel et al., 2015; Figure 2). Chronostratigraphic constraints for the top of the formation (Merifons Member) are lacking, so the extent of Guadalupian deposition (cf. Schneider et al., 2006;Pochat and Van Den Driessche, 2011) is unclear. The Octon Member (basal to mid-upper Salagou) makes up the majority of the Salagou Formation and consists primarily of massive red mudstone with loessitic characteristics (Pfeifer et al., 2020). The MS data for this study were collected from the Octon Member to lower Merifons Member (Figures 1, 2; red bar), which represents deposition through the mid-late Cisuralian .
Rock magnetism of Quaternary loess (and loess-based paleosols) has been used as a proxy (directly proportional) for mean annual precipitation (MAP) of ∼300-1500 mm/year (e.g., Porter, 2001;Maher and Possolo, 2013;Maher, 2016;Gao et al., 2019). Depending on the parent material, amount of biological mixing, carbonate leaching, and clay illuviation, soils less penetrable by precipitation in areas with MAP > ∼600 mm may result in water-logged conditions. Together with the presence of organic matter, this promotes anoxia and post-depositional dissolution of (fine) magnetic minerals.

Magnetic Susceptibility Sampling
Sampling Magnetic Susceptibility in the Field (KT-10) A detailed sedimentological investigation of the Salagou Formation (Pfeifer et al., 2020) included measurement and description (at a decimeter scale) of ∼1000 total meters of stratigraphic section exposed at the surface (Figure 3). One MS reading (each flanked by free air measurements) was taken every 0.5 m through the entire formation with the Terraplus KT-10 Plus v2 handheld MS meter that includes a pin and circular coil for uneven surfaces. In mathematical terms, MS (k) is the ratio of the intensity of magnetization (I) to the magnetic field (H) responsible for the magnetization. It is calculated with the KT-10 Plus handheld meter based on the difference in frequency between the rock and free air with a sensitivity of 1 × 10 −6 SI units (Kt-10 manual, 2017). Results (k) in dimensionless units of 10 −3 are reported as a curve constructed from the individual MS values at 0.5 m increments through each section.
The Salagou Formation was measured and described over nine stratigraphic sections ( Figure 3A). Most sections are continuous (from the top of one to the base of the next) and can be concatenated for analysis. However, there is ∼500 m of missing section (inaccessible in outcrop) between the lowermost section (Section 1; Figure 3A) and the remainder of the formation (Sections 2-9; Figure 3A), necessitating the need to analyze Section 1 separately. Thus, time series analysis was done in two large segments (as noted in Figure 3A, lower horizontal bar): the lower (165 m of continuous Section 1; results in Figure 5) and the mid-upper (775 m of eight concatenated Sections 2-9; results in Figure 6).

Sampling for Laboratory Magnetic Measurements
Two sets of samples were collected from the middle Salagou Formation ( Figure 3A) for laboratory rock magnetic analysis (section "Laboratory Magnetic Measurements") at two different increments. In the first set ( Figure 3B), a 10 g sample was collected at a 5 m resolution for ∼80 m of section. The second set ( Figure 3C) is higher resolution (a 10 g sample collected every 30 cm) but covers a shorter interval (∼10.2 m total).

Time Series Analysis and Modeling
Time series analysis of the MS stratigraphic series was carried out in MATLAB and Astrochron for R (Meyers, 2014) (see Supplementary Appendix S1 for commands). MATLAB functions were used to display the stratigraphic series, and for spectral analysis and evolutionary spectrograms to assess the distribution of variability as a function of frequency (cycles/meter) and along stratigraphic position to assess potential changes in sedimentation rate. A low-pass Taner filter was applied FIGURE 2 | A summary of the current Permian chronostratigraphic constraints of the Salagou Formation (Lodève Basin). The section where MS data were measured in this study is highlighted in gray and also indicated with the red bar. Note that the upper age of the Salagou Formation is very poorly constrained. Icon shapes indicate different types of age data, numbered 1-8 according to the sources from which they are derived; Specifically: (1) U/Pb ID-TIMS on ash 291.0 ± 0.2 from the upper Viala Formation (Michel et al., 2015), (2) tetrapod tracks, Dromopus-Erpetopus biochron boundary, ∼290-284 Ma (Schneider, pers. commun. 2020); (3-6) U/Pb ID-TIMS on ash (in stratigraphic order: 284.4 ± 0.1, 284.5 ± 0.1, 283.5 ± 0.1, 282.9 ± 0.1) from the Octon Member of the Salagou Formation (Michel et al., 2015), (7) magnetostratigraphic data from the La Lieude Formation that suggests proximity to the end of the "Kiaman" or Permo-Carboniferous Reversed Superchron (PCRS), ca. 262 Ma (Evans, 2012), and (8) tetrapod track ichnofauna from the La Lieude Formation consistent with a Guadalupian age (Schneider et al., 2010(Schneider et al., , 2019.
to pre-whiten the stratigraphic series, for clarity in evaluating stratigraphic cyclicity potentially related to orbital eccentricity.
The absence of an adequately precise time scale (Figure 2) motivated the application of objective modeling with the average spectral misfit (ASM) method provided in Astrochron (Meyers and Sageman, 2007). The ASM method searches for astronomical frequencies in well-sampled stratigraphic series across a range of plausible sedimentation rates. For each test sedimentation rate, the stratigraphic series is converted to a time series and compared against an astronomical target (in kyr). Here the astronomical target is a collection of astronomical frequencies based on a precession frequency of k = 56.6 arcsec/year estimated for 298 Ma (Berger and Loutre, 1994) and assuming Solar System fundamental frequencies s i , g i from Laskar et al. (2004; Table 1). The sedimentation rate yielding time frequencies in the data that most closely match those of the astronomical target (with Monte Carlo testing for assessing significance) is taken as the "optimal sedimentation rate."

Laboratory Magnetic Measurements
More elaborate rock magnetic analysis was carried out for representative sections from the middle Salagou Formation (Figures 3B,C) to determine the origin of the magnetic signal and strengthen the validity of MS as a proxy for the ancient paleoclimate record. The analyses were done at the rock and paleomagnetic laboratory of the Leibniz Institute for Applied Geophysics (LIAG) in Grubenhagen, Germany. The samples were dried, homogenized, weighed, and filled into non-magnetic plastic boxes of 6.4 cm 3 . The MS for all samples was measured using a frequency-and field-variable Magnon VSM susceptibility bridge at 505 Hz (low frequency susceptibility; χlf) and at 5050 Hz (high frequency susceptibility; χhf), both at a field of 400 A/m. These measurements were also used to detect the frequency dependence of the MS, which can be expressed as simple difference ( χ or %χfd). The temperature-dependent MS was determined using an AGICO MFK1-FA Kappabridge and the CS4 furnace, covering a temperature range of 20-700 • C (Figure 4). Additionally, the system was rinsed by Argon provided in a constant flux and pressure. Thermally stable furnace holders were measured in an empty state and subtracted, based on their MS rusting of the full-range heating. The applied field strength was set to 400 A/m. The correlation between MS measured in the field to χlf is weak (r 2 = 0.1758), but the relationship between the two datasets is highly significant (P-value < 0.05). (∼10+ cm), randomly oriented, semi-radial slickensides (Figures 3B,C). Micromorphologic attributes of samples from these localities include higher proportions of clay, wedgeshaped peds, clay coatings on grains, and porphyroskelic fabrics. Geochemical analyses reveal elevated chemical index of alteration (CIA) values from these intervals (compared to neighboring loessite). These features, together with a lack of horizonation, are most analogous to modern Vertisols which are typically indicative of environments with seasonal precipitation (e.g., Mack et al., 1993). Table 2 summarizes the statistics of MS data obtained by portable MS meter, reporting the mean and standard deviation of susceptibility readings overall and by section (#1-9; Figure 3A). Overall, the mean measurement of k (MS) for the Salagou formation is 0.253 ± 0.100 (×10 −3 ). The section with the lowest readings (uppermost Salagou; Section 9) averages 0.179 ± 0.034 (×10 −3 ) and the section with the highest readings (middle Salagou; Section 5) averages 0.366 ± 0.092 (×10 −3 ).  This table reports the mean and standard deviation of MS measurements overall, and also by section (#1-9; corresponding with the stratigraphic columns in Figure 3A). The number of readings (measurements) for each section is also reported, as well as the sampling interval. Note: the aim was to take a measurement every 0.5 m, but sampling intervals are slightly wider than 0.5 m because of unexposed parts of the section.

Lower Salagou Formation
A repeating pattern appears in the MS series ( Figure 5A), defined by a predominant peak at 0.1 cycles/m (10 m wavelength) in the power spectrum ( Figure 5B) and warm colors in the spectrogram ( Figure 5C). Prominent, shorter frequencies of 0.29 cycles/cm (3.45 m) and 0.57 cycles/m (1.75 m) are also consistent through the section. ASM analysis estimates an optimal sedimentation rate of ∼9.4 cm/kyr and indicates that astronomical forcing is statistically significant (P = 0.002; Figure 5D).

Mid-Upper Salagou Formation
Analysis of the mid-upper Salagou Formation (Figure 6) reveals a strong, very low-frequency signal-especially in the middle part of the section-coincident with observed ∼100-130 m (large-amplitude variations) in the MS series ( Figure 6A). The most significant cyclicity is ∼0.1 cycles/m (0.13-0.08 cycles/m), i.e., 10 m thick cycles. Subordinate, higher frequency peaks at 0.28-0.31 and 0.55 cycles/m appear intermittently throughout the series. ASM analysis estimates an optimal sedimentation rate of 13 cm/kyr and indicates that astronomical forcing is statistically significant (P = 0.002; Figure 6D).

Laboratory Magnetic Measurements
Temperature (Figure 4) and frequency-dependent susceptibility measurements from the representative section (Figures 3B,C) indicate that the origin of the magnetic signal is primarily driven by hematite, as evinced by the decrease of the MS at its Curie temperature at 675 • C, with a varying but generally minor contribution of magnetite at a Curie temperature of 580 • C (Figure 4).

Calibration of Magnetic Susceptibility Measurements by KT-10
The data acquired by portable MS meter (see Section "Cyclicity in the Salagou Formation") demonstrate the tool's ability to recognize cyclic trends through extended stratigraphic successions of loessite. This methodology is relatively rapid, convenient, and inexpensive, but it is important to calibrate in situ outcrop MS data with rock magnetic analysis from the laboratory. The correlation between MS acquired by the portable MS meter in the field and detailed rock magnetic data acquired in the laboratory (χlf and χ) is weak (r 2 = 0.1758; Figure 3C), but a simple t-test indicates with 95% confidence that the relationships between χlf and field MS, and χ and field MS, are both highly significant (P-values < 0.05). The standard deviations in most sections ( Table 2; Sections 1-3 and 7-9) for MS readings obtained with the portable meter (0.179 × 10 −3 -0.234 × 10 −3 ) are between 0.034 × 10 −3 -0.064 × 10 −3 , which further supports reproducibility. Prior to taking measurements through a section, several readings were recorded on the same sample to understand variability in the magnetic signal by lithology, and the tool always measured within 0.05 × 10 −3 . For example, in Section 3 ( Figure 3A), six measurements were taken on calcareous interbeds for a mean reading of 0.158 ± 0.037 × 10 −3 and 11 measurements were taken on pedogenic slickensides for a mean reading of 0.232 ± 0.055 × 10 −3 . In comparison to the average loess reading for this same section (0.220 ± 0.056 × 10 −3 ), this demonstrates the reliability of the tool to capture magnetic signal variations in different lithologies as well as reproducibility of readings on the same sample. However, there are limitations with in situ MS measurements. The poor correlation between field and laboratory data may be caused by inhomogeneous volumes measured in the field (e.g., uneven surfaces or fractures/cavities in outcrop). Elevated mean MS readings from the portable meter (0.303 × 10 −3 -0.366 × 10 −3 ) in the middle Salagou formation ( Table 2; Sections 4-6) have especially high standard deviations (0.092 × 10 −3 -0.135 × 10 −3 ) compared to other sections. It is possible that elevated values in the middle Salagou Formation relate to a decline in calcareous cement observable in this interval.

Origin of the Magnetic Signal and Comparison to Quaternary Eurasian Loess
The direct relationship between χlf and χ (Figure 7; frequency dependence) demonstrates that magnetic enhancement (increased χ) is triggered by ultrafine SP particles close to the domain transition boundary. The production of SP magnetic grains can occur either in situ via pedogenic processes (Singer et al., 1996;Maher, 1998), by incorporation of detrital particles (e.g., Reynolds et al., 2001), and/or through conversion of FIGURE 7 | Frequency-dependent susceptibility (χlf vs. χ) of the Salagou Formation loess samples (red spades from Figure 3B; black circles from Figure 3C) compared to data from the Quaternary CLP red clay (blue triangles; Song et al., 2007) and a loess-paleosol section from Romania (gray crosses; Zeeden et al., 2016). magnetic mineral species during fluid or burial diagenesis (Katz et al., 1998;Woods et al., 2002). Given the burial history of the basin (T max up to 500-600 • C; Copard et al., 2000), we cannot dismiss the possibility that some SP particles may have formed as a result of diagenesis, but minimal evidence occurs for fluid flow or veining in the Salagou Formation exposures. We prefer the interpretation of primarily pedogenically formed magnetic particles for the following reasons: (1) Association with evidence for pedogenesis, namely Vertic features (section "Sedimentology"), and elevated clay mineral content and CIA (compared to parent loessite). Lindquist et al. (2011) show that the seasonal wetting and drying typical of Vertisols facilitate the formation of ferrimagnetic SP minerals. (2) There is no evidence for gleying: The Vertic characteristics of Salagou Formation paleosols are inconsistent with ever-wet conditions (Tabor and Myers, 2015). Numerous studies on recent and modern soils show consistently that non-water-logged (gleyed) soils have higher MS and greater abundances of pedogenically produced ultra-fine-grained ferrimagnetic material than in surrounding loess (cf. Heller and Liu, 1984;Maher and Taylor, 1988; and other references in section "Geological Setting"). Furthermore, (3) the alignment of χlf and χ with other Eurasian loess (Figure 7) suggests pedogenic control on the magnetic signal (see also discussion below) as opposed to trends typical of diagenesis.
Rock magnetic data from the Salagou Formation shows magnetic enhancement, but distinctly lower χlf and lower χ compared to reference data (Figure 7) from the Pliocene red clay on the CLP (Song et al., 2007) and a loess-paleosol sequence in Romania (Zeeden et al., 2016). A systematically lower χlf and χ relates to the predominance of hematite, which generally exhibits a weaker magnetic signal than magnetite and maghemite. The dominance of hematite and lower contribution of magnetite and maghemite in the Salagou Formation is similar to that found in Permian loessite from western equatorial Pangea (Soreghan et al., 1997;Cogoini et al., 2001;Jia, 2020), but unusual relative to Quaternary loess deposits. This difference may be attributable to syndepositional processes (e.g., eolian sorting or dilution from rapid deposition of quartz and other non-iron-bearing minerals), more arid and/or oxidizing climate conditions preferring hematite formation over magnetite genesis or preservation (Maher, 1998), weak or stifled pedogenesis (documented sedimentologically; section "Sedimentology"), and/or post-depositional oxidation (of magnetite and maghemite). Eolian sorting is unlikely, because instead of the direct relationship between χlf and χ in the Salagou Formation (Figure 7), sorting tends to generate an inverse pattern (high χlf with low χ; e.g., Matasova and Kazansky, 2004). Similarly, hematite precipitates formed during post-depositional processes tend to generate a magnetic signal with an inconsistent pattern between χlf and χ. The well-defined alignment of χlf and χ between these datasets (Figure 7) strongly suggests that the magnetic signal in the Salagou Formation is driven by pedogenesis. If there is no postdepositional (diagenetic) conversion of magnetite/maghemite to hematite, then the low contribution of magnetite in the Salagou Formation loess compared to Eurasian Quaternary loess might reflect substantially different paleoclimatic settings in the Permian of France.
If changes in the magnetic signal are pedogenically driven, then samples in this dataset represent a range of pedogenic alteration, where samples with more magnetic enhancement are more enriched in ultrafine particles of hematite. However, the average bulk χ from interpreted paleosol (1.3 ± 0.3 × 10 −3 ) and loess (1.5 ± 0.4 × 10 −3 ) units in the Salagou Formation overlap within error. This is atypical of Quaternary Eurasian loess, and even compared to ancient loessite from western equatorial Pangea wherein the magnetic signal of paleosolsalso driven primarily by hematite-is 1.5-2× higher than in the parent loess. But the rock magnetic measurements in the Salagou Formation (1-1.5 × 10 −3 ) are overall much lower compared to the Permian loessitic paleosols of western equatorial Pangea (∼2-6 × 10 −3 ; Soreghan et al., 1997Soreghan et al., , 2002Cogoini et al., 2001;Tramp et al., 2004;, so it seems that the Salagou Formation preserves distinctively low concentrations of magnetite or maghemite in comparison even to time-equivalent loessite on the other (western) side of Pangea.

Cyclicity in the Salagou Formation
In the lower Salagou MS series (Figure 5), the persistent frequency at 0.1 cycles/m indicates a 10 m thick cyclicity. Higher frequencies at 0.29 and 0.57 cycles/m indicate 3.5 and 1.75 m thick cycles, respectively. This nested cyclicity, i.e., five to six shorter cycles within each 10 m "bundle, " also appears in the MS series ( Figure 5A). If the predominant (10 m thick) cyclicity represents the orbital eccentricity, then this section (160 m) represents a duration of ∼1.6 Myr, and the well-defined 3.5 and 1.75 m cycling represents the obliquity and precession index, respectively. There is also a very low frequency corresponding to a 44-m cycle, i.e., a candidate for 405 kyr orbital eccentricity.
In the mid-upper Salagou MS series (Figure 6), there is longterm amplitude variation that is potentially associated with the decreased presence of calcareous cement in the middle part of the section (resulting in elevated MS values) and/or missing or repeated strata in the concatenation of successive sections ( Figure 3A). Regardless, the predominant cyclicity is ∼9.4 m thick, which appears to be broadened by variable sedimentation rates along the series, with multiple closely spaced spectral peaks from 0.13 to 0.08 cycles/m (7-12 m thick cycles). Higher frequencies at 0.28-0.31 and 0.55 cycles/m translate to ∼3.3 and 1.8 m thick cycles, respectively. If the predominant 9.4 m cycles represent short orbital eccentricity, then this section (775 m) represents ∼7.8 Myr of sedimentation. It is possible that ∼3.3 and 1.8 m thick cycles represent obliquity and precession (respectively) but are inconsistent through the section and locally poorly defined.
The ASM analysis confirms that astronomical frequencies are present and statistically significant (P-values < 0.05 for both MS series), rejecting the null hypothesis of no astronomical forcing for specific sedimentation rates of 9.4 cm/kyr (lower Salagou) and 13 cm/kyr (mid-upper Salagou). The up-section increase of optimal (ASM-based) sedimentation rates from 9.4 to 13 cm/kyr suggests that the Salagou Formation represents a duration of 9-10 Myr. This result is similar to estimates for the Salagou Formation from previous work (11-17 cm/kyr; Pochat and Van Den Driessche, 2011) as well as average sedimentation rates for the middle Permian Lucaogou Formation of northwest China (8.9-10.3 cm/kyr; Huang et al., 2020). It is slightly faster than lower Permian loess deposition in western equatorial Pangea (>700 m over ∼9 Myr so ≥7.8 cm/kyr; Johnson, 1989;Tramp et al., 2004;Soreghan M. J. et al., 2014), and on the lower end of sedimentation rates proposed for the Plio-Pleistocene CLP (Stevens and Lu, 2009).
If the driver of magnetic enhancement is pedogenesis (see section "Origin of the Magnetic Signal and Comparison to Quaternary Eurasian Loess"), then these data demonstrate fluctuation between loess accumulation and soil formation with loessite-paleosol couplets around 10 m thick in the Salagou Formation, which is similar to the lower Permian loessite of western equatorial Pangaea . Irregular cyclicity (e.g., 100-130 m; Figure 5A) may reflect longterm variability in atmospheric circulation, hydroclimate, and/or changes in sedimentation rates.

CONCLUSION
This study documents statistically significant astronomically forced climate change recorded in deep-time loessite of lowlatitude Pangea. The results are as follows: • Spectral analysis of handheld MS stratigraphic data shows a persistent 10 m thick cyclicity through the Permian Salagou Formation that denotes orbital eccentricityscale (∼100 kyr) variability. Accordingly, subordinate, higher frequency cycles, at ∼3.3-3.5 and ∼1.8 m thick (most prominent in the lower Salagou Formation) likely represent 35 kyr obliquity and 17-21 kyr precession indexscale variations. • Average spectral misfit analysis supports the interpretation of the presence of astronomical forcing frequencies, which are statistically significant in the MS series for optimal sedimentation rates of 9.4 cm/kyr (lower Salagou Formation) and 13 cm/kyr (mid-upper Salagou Formation): similar to the Permian of northwest China (∼9-10 cm/kyr) and the western United States (∼8 cm/kyr). • Rock magnetic data from the Salagou Formation indicates magnetic enhancement wherein we interpret that magnetic particles (hematite) likely formed by pedogenesis, possibly with a diagenetic overprint. • The uniquely low magnetic signal in the Salagou Formation compared to analogous Eurasian Quaternary loess deposits relates to the predominance of hematite, which generally has a much weaker magnetic signal than magnetite and maghemite, and may be explained by differing conditions of formation (e.g., syn depositional processes, more arid and/or oxidizing climate conditions) than in present Eurasia (not unexpected), and/or post depositional oxidation of magnetite and maghemite. • The portable MS meter allows for rapid and convenient in situ data acquisition. This study demonstrates the ability to recognize cyclic trends through extended stratigraphic successions of loessite using this method when calibrated with detailed rock magnetic analysis from the laboratory.
The evidence demonstrates that the Salagou Formation records climatic variability over a time interval of ∼9-10 Myr during the early Permian (middle to late Cisuralian). If the driver of magnetic enhancement is pedogenic, then the ∼10 m thick cyclicity that is consistent over ∼1000 m of section may represent the thickness of loessite-paleosol couplets in the Salagou Formation.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

AUTHOR CONTRIBUTIONS
LP and GS (advisor) collected MS data and samples for rock magnetic analysis. LH mentored LP in MS interpretation (including MATLAB and R scripts). CZ conducted advanced rock magnetic analysis and interpreted this data (with close advice from CZ and CR). LP synthesized all data and conversations into manuscript format. All authors have made significant contributions to the editing and revision of this work, and approved the submitted version.