436 ISSN 1819-7140, Russian Journal of Pacific Geology, 2024, Vol. 18, No. 4, pp. 436–451. © Pleiades Publishing, Ltd., 2024. Holocene Geomagnetic Excursions in Peat Deposits A. Yu. Peskov a , *, A. N. Didenko a , b , A. S. Karetnikov a , M. A. Klimin c , M. V. Arkhipov a , N. V. Kozhemyako a , and A. I. Tikhomirova a a Kosygin Institute of Tectonics and Geophysics, Far Eastern Branch, Russian Academy of Sciences, Khabarovsk, Russia b Geological Institute, Russian Academy of Sciences, Moscow, Russia c Institute of Water and Environmental Problems, Far Eastern Branch, Russian Academy of Sciences, Khabarovsk, Russia * e-mail: peskovitig@yandex.ru Received March 5, 2024; revised March 26, 2024; accepted March 29, 2024 Abstract —The paper presents the results of comprehensive (microprobe, paleomagnetic and magnetic) stud- ies on peats from the Tyapka peat section (Russian Far East). Radiocarbon dating placed the start of peat for- mation at ∼ 11.7 ky B.P. The principle carriers of magnetization were found to be magnetite, to a lesser extent hematite and, possibly, greigite. The relative paleointensity values obtained through calculations are in good agreement with the literature data. Intervals with negative inclinations of the magnetization vector were iden- tified in peats, which most likely correspond to geomagnetic excursions in the Holocene. The research con- strained the duration of the geomagnetic excursions, as well as the geomagnetic field intensity behavior not typical for such variations. Keywords: paleomagnetism, peat, geomagnetic excursions, relative paleointensity, Holocene DOI: 10.1134/S1819714024700143 1. INTRODUCTION Geomagnetic excursions are an important feature of the fine structure of the geomagnetic field and con- stitute a special type of its variations. The discovery of excursions and the identification of their characteristic features allow us to suggest a more complex structure of the liquid core than previously thought and a special type of interactions between the inner core and the outer core (for example, [28]). Currently, a geomag- netic excursion is defined as a short-time change (up to 2 000 years) in the geomagnetic field direction (by 60–180 ° , >45 ° virtual geomagnetic pole (VGP) co- latitude), the amplitude of which is no less than three times exceeds the normal range of geomagnetic secu- lar variation [26, 35]. In the latest version of [TSCreator 8.0, 2023], men- tion is made of 30 geomagnetic excursions, 29 of which are known to occur during the Quaternary, including the Matuyama (21) and Brunhes (8) polar- ity chrons. According to other data [37], 17 geomag- netic excursions are proposed in the Brunhes Chron. The difference in the number of excursions in different magnetostratigraphic time scales of the Brunhes Chron, we believe, is due to objective difficulties in their detecting and dating. Based on the duration and behavior of the geomagnetic field, two types of geo- magnetic excursions are recognized, the nature of which may also be different. Geomagnetic excursions of the 1st type are long-term events; their duration is close to the period of the main dynamo ( ∼ 9000 years [2]); and they occur against the background of reduced intensity of the dipole component of the geo- magnetic field. Geomagnetic excursions of the 2nd type are short-lived episodes lasting from hundreds of years to a few thousand; they can occur while the intensity of the dipole component of the geomagnetic field is either decreasing or increasing. We believe that the youngest geomagnetic excursion Etrussia (Sterno- Etrussia) with an age of about 2.7–2.8 thousand years discovered in the Holocene is of this type, as set out below. The first mention of the anomalous behavior of the geomagnetic field in Etruscan times dates back to the end of the 19th century [10]. Almost a century later, archaeomagnetic research in Georgia yielded evidence of the anomalous behavior of the geomagnetic field during the 7th and 6th centuries BC, which was ascribed to the geomagnetic excursion named Nam- cheduri [22]. Most notably, the excursion occurred against the background of the highest values of the geomagnetic field paleointensity in the Holocene. To date, at least 15 marine and continental sedimentary sections are known, in which the Sterno-Etrussia excursion is recorded [9] in varying detail. The main questions are still the same, that is, its duration and paleointensity during the excursion. RUSSIAN JOURNAL OF PACIFIC GEOLOGY Vol. 18 No. 4 2024 HOLOCENE GEOMAGNETIC EXCURSIONS IN PEAT DEPOSITS 437 It is not an easy task to qualitatively record the anomalous behavior of the Earth’s paleomagnetic field (the transition from normal behavior to anoma- lous, and vice versa) lasting only the first hundreds or thousands of years; unique geological objects are required either with high rates of sedimentation (sedi- mentary sections), or with almost continuous effusive activity over the first thousand years (volcanic edi- fices). Recently, the use of high-sensitivity magne- tometers has made it possible to conduct such studies on chemogenic sediments (speleothems) from karst caves to construct high-resolution records of chrono- logically constrained geomagnetic field behavior [6, 11]. Our previous research has shown that peats indi- cate a good potential to be used for studying the geo- magnetic field fine structure [24, 25]; a number of fac- tors contribute to this as well. Firstly, high rates of ver- tical growth which according to [33] are as high as 2.6 mm/year for peat deposits in the southern taiga zone of Western Siberia. Secondly, in most cases an oxygen-free reducing environment of all diagenetic processes in peats, which contributes to the preserva- tion of allochthonous magnetic grain-size particles transported by wind and water flows. Thirdly, except for periods when peatlands were on fire, virtual conti- nuity of the record of palaeomagnetic, magnetic, and geochemical characteristics throughout the Holocene. The latter circumstance is very important for veri- fying the obtained time dependencies for a specific object with records for other objects and processes due to both global climate changes and regional phenom- ena, including catastrophic ones. This paper presents the results of comprehensive studies of the section through the Tyapka peat bog, one of the most ancient peat bogs in the Far East of Russia, which formed during the Holocene. The goal of the research was to obtain a complete magneto- stratigraphic record for the section with the lowermost portion at ∼ 11.7 ky B.P. 2. DESCRIPTION OF THE PEAT SECTION AND RESULTS OF GEOCHEMICAL STUDIES Tyapka peat section is located in the Far East of Russia (53.69 ° N, 140.09 ° E, Fig. 1). The peat massif measuring 400 × 200 m with frozen peat in the middle is part of the vast flattened, swampy lowland with ele- vations of 20–70 m above sea level, stretching 18 to 20 km east of the section site, where it grades off to a steep drop leading down to the coast of the Sea of Okhotsk. This peat bog is a raised bog drained by open ditches, which ensured “clean” sampling (samples without inputs of allochthonous materials) for assay- ing, pollen analysis, and radiocarbon dating, paleo- magnetic and rock magnetic measurements. The sec- tion consists of peats varying in composition and degree of decomposition (Fig. 2a). The ash content in the upper 2 m thick stratum of peat ranges from 1 to 2.5%, except for the uppermost 10 cm interval, where the peat ash content is 5–6% due to the decomposi- tion of plant residues after drainage of the peat bog. At a depth of 3.6 m, the ash content ranges from 3 to 4% suggesting very limited inputs of allochthonous min- eral inclusions during peat formation. The behavior of the ash content in the section (Fig. 2a) indicates the absence of floods and major wildfires during peat for- mation. Radiocarbon ( 14 C) ages of eight samples [24] from the peat bog are available, which are fairly evenly dis- tributed across the section (Table 1). The correlation coefficient (r) of the calibrated age-depth relationship of these eight samples is 0.993. These dates were used to calculate the age of all the undated horizons in the section. Layer by layer sampling of the section yielded 200 azimuthally oriented samples to undertake paleomag- netic and magnetic studies. Samples were collected by continuous sampling by pressing 2-cm plastic cubi- tainers with magnetic susceptibility of –2.35 × 10 –6 SI units (diamagnet) into the exposed vertical side of the peat section. A mining compass was used for azi- muthal orientation of the samples. As mentioned above, the peat massif hosts frozen peat overlain by unfrozen peat soils with land-surface Table 1. Radiocarbon and calibrated ages of peat samples from the Tyapka section Depth (h), cm Sample ID Radiocarbon age, 14 C yr BP Calibrated age, cal yr BP 49 ИМКЭС-14С1095 1765 ± 67 1653 +171 / –122 95 ИМКЭС-14С1097 2604 ± 98 2687 +192 / –324 141 ИМКЭС-14С1096 3663 ± 72 3995 +237 / –264 195 ИМКЭС-14С1098 4414 ± 66 5026 +258 / –169 247 ИМКЭС-14С1100 5437 ± 66 6230 +162 / –227 297 ИМКЭС-14С1102 7519 ± 102 8314 +226 / –274 345 СОАН-7076 9110 ± 105 10293 +277 / –378 393 СОАН-7079 10005 ± 140 11543 +494 / –367 438 RUSSIAN JOURNAL OF PACIFIC GEOLOGY Vol. 18 No. 4 2024 PESKOV et al. elevations between 1 and 1.5 m, which is relict perma- frost that would cover a much larger, if not the entire, area of the peat bog in the past. In the mid-1990s, drainage of the peat massif using 3-m deep open ditches triggered the gradual degradation of perma- frost. Two or three years after the drainage work com- pletion, the first sections revealed that the frozen peat is dome-shaped repeating the shape of a mound with the top at a depth of about 0.5 m and the bottom at about 3.1 m and underlain by a 2-m thick layer of unfrozen peat. By now, the surface of the frost peat mound has subsided significantly; peat deposits sur- rounding it thaw out completely during the warm sea- son. The section we studied cut through the horizontal peat stratum as a result cleaning up the side of the drainage ditch. Geochemical studies on 40 samples from the section evenly distributed in depth revealed that the content of Fe and a number of other elements (Co, Sr, Ba, and Cd) increases with depth (Fig. 2b); the correlation coefficient (r) for Fe = 0.94. At the same time, in the interval 140–230 cm ( ∼ 3.9–5.9 ky B.P.) there is a horizon where this relationship is not observed (the “frozen peat” interval in Fig. 2b). The Fig. 1. Tyapka peat section in the Far East of Russia (53.69 ° N, 140.09 ° E). R u s s i a C h i n a Khabarovsk Amur River Amur River R u s s i a Strainght of Tatary Sakhalin I. A s i a Tyapka peat bog U s s u r i R i v e r Fig. 2. Tyapka peat section: (a) degree of decomposition and ash content, (b) Fe content. Symbols: ( 1 ) degree of decomposition; ( 2 ) ash content; ( 3 – 4 ) data points ( 3 ) and linear regression ( 4 ) of the geochemical analysis results (Fe content). 1 2 3 4 0 1 2 3 4 2000 4000 6000 Fe, g/t Frozen peat r = 0.95 r = 0.46 r = 0.92 (b) r = 0.94 0 1 2 3 4 100 % 50 (a) h , m h , m RUSSIAN JOURNAL OF PACIFIC GEOLOGY Vol. 18 No. 4 2024 HOLOCENE GEOMAGNETIC EXCURSIONS IN PEAT DEPOSITS 439 correlation coefficient for this interval is significantly lower and amounts to 0.46, while for the other two inter- vals in the section (above 140 cm and below 220 cm) the correlation coefficient is high (0.95 and 0.92, respec- tively) (Fig. 2b). Based on the above, we can conclude that in the section under consideration, permafrost degradation (hydrodynamic factor) and compaction of the previ- ously frozen peat, which currently occurs in the range of 140 to 220 cm, caused the redistribution of Fe-con- taining (magnetic) minerals, which affected the pres- ervation of the magnetic field record in this interval. In the following sections of the article, the paleomagnetic data obtained for the “frozen peat” interval are not interpreted. 2.1. Magnetic Remanence Carriers: Findings of Microprobe and Thermomagnetic Studies Magnetic separation of peat samples was imple- mented using a Nd 2 Fe 14 B permanent magnet put into a plastic bag. A magnetic fraction of ∼ 1 g was isolated to study temperature dependence of magnetic susceptibility (k) in the 0–200 cm interval of the study peat bog, which allowed one analysis to be performed. Thermomag- netic analysis (TMA) was performed on a Kappa- bridge MFK-1FA magnetic susceptibility meter equipped with a CS3 temperature-susceptibility mod- ule (Agico, Czech Republic). A total of 28 samples were prepared for microprobe analysis, which characterize intervals in the peat sec- tion as thick as 2 to 5 cm and are evenly distributed in depth. Microprobe measurements were made on a scanning (raster) electron microscope VEGA 3 LMH (Tescan, Czech Republic) coupled with an X-Max 80 energy-dispersive spectrometer with an AztecTM microanalysis system (Oxford Instruments, UK). Investigations were carried out at the Khabarovsk Innovative-Analytical Center of the Kosygin Institute of Tectonics and Geophysics, Far East Branch, Rus- sian Academy of Sciences (KhIAC ITiG FEB RAS). It was found that the main magnetic minerals in the peat samples are detrital magnetite particles ranging in size from 1 to 70 μ m, often with an admixture of Cr, more rarely Ni (Figs. 3f–3m) with Tc ∼ 580 ° С (Fig. 3a). Less common are microspherules of magnetite with sizes from 1 to 10 μ m (Figs. 3b–3e), more rarely hematite with Tc ∼ 670 ° С (Fig. 3a), titanomagnetite (Fig. 3l), particles of native iron (Fig. 3p), wolframite (?) (Fig. 3q), and sulfides (Figs. 3n, 3o). During TMA, heating of the magnetic fraction to 300 ° C causes k to gradually increase by approximately 4.5% (Fig. 3a). In the interval of 320 to 370 ° C, a sharp increase in k by ~16% of the initial value is observed, which may be associated with the alteration of sulfides (possibly greigite) and the formation of new magnetite during heating. Then, when heated to approximately 550–560 ° C, a fairly gradual increase in k is again observed, which is associated with the Hopkinson effect [32]. The heating curve for k clearly shows two Curie points (Tc) (Fig. 3a), that is, magnetite ( ∼ 580 ° C) and hematite ( ∼ 670 ° C). The cooling curve clearly shows the Curie point in the interval of 580 to 590 ° C and the k value after heating is ∼ 1.33 of the ini- tial value, which is related to the formation of second- ary magnetite. From the results of microprobe and thermomag- netic analyses we can confidently conclude that the main magnetic carrier of the investigated peats is mag- netite, and to a lesser extent, hematite and, possibly, greigite. 2.2. Relative Paleointensity In [24], the first relative paleointensity values prior to 4.0 ky B.P. were obtained from Tyapka peats. Rela- tive paleointensity of the geomagnetic field was assessed using the Bagina-Petrova technique [26, 32], which invokes the comparison of AF demagnetization curves of natural remanent magnetization (NRM) and ideal (anhysteretic) remanent magnetization (ARM) for the same sample. The NRM/ARM ratio in the range of alternating magnetic fields, where their coer- citivity spectra coincide (black squares aligned on the vertical line in Fig. 4), is equal to the Ha/Hl ratio, where Ha – paleointensity, Hl – magnetic field intensity when ARM was induced; in our case Hl = 40 A/m. The studies were performed on a cryogenic magnetometer (2G Enterprise, USA, sensitivity 1 × 10 –7 A/m) at the KhIAC ITiG FEB RAS. The relative paleointensities obtained in [24] for the Tyapka peat section with the lowermost portion at 4.0 ky B.P. are in good agreement with paleointensity estimates based on the archaeomagnetic data on the Iberian Peninsula [21] and Central Asia [23], as well as with the data on relative paleointensity of sedimentary rocks of the Western Pacific [27] (Fig. 5). The agree- ment between the variations in relative paleointensities we obtained and the above literature data for different locations of the Earth is a strong argument in favor of the magnetic nature of paleomagnetic anomalies iden- tified in the Tyapka peat section during magnetostrati- graphic investigations (see Section Paleomagnetic and magnetic measurements ). The paper presents the relative paleointensities obtained from peat samples of the Tyapka section with the lowermost part at ~11.7 ky B.P. As a result of inves- tigations, the “(NRM/ARM)-age” relationship was constructed, which determines the relative magnetic field intensity at the time of peat layers emplacement (Fig. 6). A comparison between obtained values for the last 11.7 ky B.P. and archaeomagnetic data for the Iberian Peninsula (Fig. 6) shows that the overall correlation coefficient between them is as high as 0.63. At the 440 RUSSIAN JOURNAL OF PACIFIC GEOLOGY Vol. 18 No. 4 2024 PESKOV et al. same time, some underestimation of r occurs due to the error in determining the age of peat deposits, espe- cially in the interval 0–1.7 ky B.P., where the age of peats is calculated by interpolation from the youngest radiocarbon date. A correlation coefficient of 0.89 indicates a good correlation between archaeomagnetic data and results obtained in the range of 1.9 to 3.9 ky B.P. Down-sec- tion, in the “frozen peat” interval, the correlation is weak ( r = 0.31, Fig. 6). Along with geochemical data (Fig. 2b), this suggests the destruction of the primary magnetic record in this layer. Below the “frozen peat” layer, in the range of 5.9–7.3 ky B.P., the correlation dependence increases and is equal to 0.65. The data obtained on relative paleointensity of peats allow us to conclude about the synchronism of changes in the Earth’s magnetic field (EMF) intensity, at least along the Iberian Peninsula coast [21]), in Central Asia [23], and along the coast of the Sea of Okhotsk (this study), as well as to confirm the mag- netic nature of the paleomagnetic record established in the peats. 2.3. Paleomagnetic and Magnetic Measurements Initial magnetic susceptibility (k) was measured by a Kappabridge MFK-1FA magnetic susceptibility meter (AGICO, Czech Republic) with a sensitivity of 5 × 10 –8 SI units. Alternating field (AF) demagnetiza- tion (H-cleaning) was performed on a 2G Enterprise cryogenic magnetometer (2G Enterprise, USA) with a sensitivity of 1 × 10 –7 A/m. Stepwise AF demagnetiza- tion was performed in 10- to 40-mT steps (from 3 to 7 steps in total) up to an alternating magnetic field (H) of 100 mT. In order to eliminate the influence of the viscous component of magnetization to calculate vari- ations in the geomagnetic field (variations of geomag- netic declination (D) and inclination (I) versus age), Fig. 3. Results of thermomagnetic (a) and microprobe (b–q) analyses. The composition of minerals is in weight percent; (h) depth in the section; the black circles are microprobe pits. h = 384–390 cm O–33.7 Fe–64.6 Ni–1.7 h = 378–384 cm O–43.5 Fe–30.8 Al–8.7 Si–9.3 Ca–4.1 Mg–3.6 O–33.6 Fe–66.4 .4 h = 198–209 cm O–42 Fe–58 h = 90–96 cm (e) h = 378–384 cm O–22.2 Fe–77.9 O–44.5 Fe–55.5 h = 190–192 cm O–29.4 Fe–69.6 h = 166–168 cm O–22.2 Cr–2.8 Fe–75 h = 132–134 cm S–51.6 Fe–48.4 h = 390–395 cm h = 122–124 cm O–10 Cr–2 Fe–88 h = 112–114 cm O–48.9 Ti–27.8 Mn–4.5 Fe–18.8 O–46 Cr–8.5 Fe–44.5 Ni–1 h = 48–50 cm h = 166–168 cm O–7.1 S–32.5 Fe–27.5 Cu–32.9 h = 102–104 cm Cr–2.3 Fe–97.7 h = 48–50 cm O–17.8 Mn–1.1 Fe–17.3 W–63.8 O–29.4 Fe–69.4 Ti–1.2 h = 390–395 cm 0 100 200 300 400 500 600 700 T , C Heat Cooling k / k o 1.0 1.2 1.1 1.3 0.9 (a) (b) (c) (d) (f) (g) (h) (j) (k) (l) (m) (q) (n) (o) (p) T c ~580 C T c ~670 C o = 1.08 × 10 –4 SI 2.5 m 1 m 2 m 2 m 2 m 2 m 5 m 5 m 5 m 2 m 1 m 1 m 20 m 10 m 1 m 2 m (i) RUSSIAN JOURNAL OF PACIFIC GEOLOGY Vol. 18 No. 4 2024 HOLOCENE GEOMAGNETIC EXCURSIONS IN PEAT DEPOSITS 441 magnetization vectors at H = 60 mT were used (62% of samples, Fig. 7a); after complete demagnetization of the sample at H = 40 mT (31% of samples, Fig. 7b); and after incomplete demagnetization at H = 100 mT (7% of samples, Fig. 7c). Age dependences of the initial magnetic suscep- tibility of the examined samples are shown in Fig. 8. The measured k is in the range from –1.4 × 10 –5 to 1.3 × 10 –5 SI units (Fig. 8b); the mean value of k is negative; its numerical value in SI units is –1.7 × 10 –6 Figure 8a shows changes in the natural remanent magnetization (NRM) of the samples. The mean NRM value for the peat deposit is 1.1 × 10 –5 A/m. The spread of the measured values is over two orders of magnitude, from 1.9 × 10 –7 to 5.4 × 10 –5 A/m. A weak positive correlation is observed between measured val- ues of k and NRM ( r = 0.34). Figure 8c, d shows depth variations in declination (D) and inclination (I) in the section. As mentioned above, the time interval between 3.9 and 5.9 ky B.P. (“frozen peat”) is not considered when interpreting paleomagnetic data due to the possible destruction of primary magnetization under the influence of hydro- dynamic forces. Excluding this interval, variations in D when aver- aged by the moving average of 5 points on the natural remanent magnetization full vector over the studied interval in the peat section are very large – from –180 ° to +156 ° . Remanent magnetization inclination values vary over a wide range from –88 ° to +88 ° (Figs. 9a, 9b). We split “D-age” and “I-age” plots into several intervals and consider each of them. 1. The 0–3.9 ky B.P. interval is divisible into three zones: upper, middle and lower. 1.1. The upper zone restricted to the 0–55 cm hori- zon (0–1.788 ky B.P.) hosts an apparently “anoma- lous” interval of ~300 years (0.5–0.8 ky B.P.) with negative inclination (I) values along with declination (D) values different from modern ones (Figs. 9a, 9b). Excluding the “anomalous” interval, variations in D and I values in the upper zone are 67 ° and 60 ° with average values of –15 ° and 24 ° , respectively. The mean value of declination in the upper zone is close to the modern one (–13 ° ). However, we emphasize that the mean I value differs from the inclination of the geo- magnetic field at the sampling site (+68 ° ). It is likely that, by analogy with most sedimentary rocks, peats in the section experience inclination flattening of mag- netization, and the number of demagnetization steps does not allow principal component analysis to be employed in full measure to properly isolate the pri- mary paleomagnetic signal. 1.2. The middle zone is localized in the horizon 57–109 cm ( ∼ 1.85–3.1 ky B.P.). Variations in aver- aged D values here are about 250 ° with an average value of –78 ° , which differs significantly from the mod- ern one. Inclination values averaged by the moving aver- age are negative here and vary from –88 to –29 ° with an average I value of –70 ° 1.3. The lower zone is confined to the horizon 111– 139 cm ( ∼ 3.1–3.9 ky B.P.) and also hosts a 6-cm thick “anomalous” interval, similar to that in the upper zone, with negative inclination values along with dec- lination values different from modern ones, which corresponds to 170 years (~3.3–3.5 ky B.P.). The mean D value in the lower zone, excluding the “anomalous” area, is close to the modern one and amounts to 3 ° with variations from –32 to 100 ° . The mean I value is positive (32 ° with variations from –10 to 68 ° ) and close to that in the upper zone. 2. The 5.9–7.5 ky B.P. interval is identified in the 233–279 cm horizon. Variations in I values are from +11 ° to –48 ° with an average of –20 ° . The mean D value differs from the modern one and is –53 ° with variations from –13 to –91 ° 3. The 7.5–11.2 ky B.P. interval in the horizon of 279–379 cm has positive I values with an average of +79 ° , which is close to the modern one (+68 ° ). The mean D value in this interval is –44 ° Fig. 4. Examples of paleointensity assessment using Tyapka peat samples: NRM/ARM ratios during stepwise demagnetization. Explanation: square—a pair of NRM and ARM for each demagnetization step (mT), which was used (black square) and not used (white square) when plotting the linear NRM/ARM relationship (dashed line) to calculate the relative paleointensity. 4E–05 2E–05 0 2E–06 4E–06 T041 ( h = 81 cm) 4E–05 2E–05 0 2E–06 4E–06 ARM, A/m 60 mT 60 mT 80 mT 80 mT NRM, A/m NRM, A/m R 2 = 0.995 у = 1.53 х R 2 = 0.967 у = 1.98 х T005 ( h = 10 cm) 0 mТ 0 mТ 20 mT 20 mT 10 mT 40 mT 40 mT 442 RUSSIAN JOURNAL OF PACIFIC GEOLOGY Vol. 18 No. 4 2024 PESKOV et al. Fig. 5. Comparison of relative paleointensity estimates for the Tyapka peat section [24] with (a) archaeomagnetic data for the Ibe- rian Peninsula [21], (b) Central Asia [23], and (c) data on the relative paleointensity of sedimentary rocks from the Western Pacific (the authors performed data averaging by a 5-point moving average) [27]. Symbols: ( 1 ) [24]; ( 2 ) [21]; ( 3 ) [23]; ( 4 ) [27]. 0 1000 2000 3000 4000 100 50 Year (BP) Year (BP) Year (BP) 20 1 2 12 8 4 NRM/ARM 0 1000 2000 3000 4000 100 50 Ha, T Ha, T 20 12 8 4 NRM/ARM 3 4 0 1000 2000 3000 4000 4 1 NRM/ARM 12 8 4 NRM/ARM (а) (b) (с) Fig. 6. Comparison of relative paleointensity estimates from the Tyapka peat section and paleointensities obtained from archae- omagnetic data on the Iberian Peninsula [21]. Symbols: ( 1 ) paleointensities obtained from archaeomagnetic data on the Iberian Peninsula [21]; ( 2 ) paleointensities obtained from peat samples of the Tyapka section with the lowermost part at ∼ 11.7 ky B.P; and ( 3 ) paleointensity values smoothed by a 5-point moving average which were obtained from peat samples of the Tyapka section with the lowermost part at ∼ 11.7 ky B.P. 1.9–3.9 ky B.P. r = 0.89 5.9–7.3 ky B.P. r = 0.65 Frozen peat r = 0.31 0 2000 4000 6000 8000 10 000 12000 100 50 1 4 8 12 2 3 Ha, T NRM/ARM Year (BP) RUSSIAN JOURNAL OF PACIFIC GEOLOGY Vol. 18 No. 4 2024 HOLOCENE GEOMAGNETIC EXCURSIONS IN PEAT DEPOSITS 443 Fig. 7. Examples of AF demagnetization of Tyapka peat samples (orthogonal diagram, stereogram, demagnetization curve) in geographic coordinates. Numbers on the orthogonal diagram and stereogram are demagnetization steps (mT). N, N E, UP NRM 10 20 40 60 M/Mmax 1 0 20 40 60 N E Up Down NRM 20 40 60 (b) T008 N, UP E, E NRM 20 40 60 (a) T010 80 80 1 60 40 20 80 0 H , mT H , mT H , mT M/Mmax S E NRM 20 40 60 (с) T052 N, N E, UP NRM 20 40 60 100 60 100 S NRM 60 40 100 M/Mmax 1 0 40 80 40 80 Fig. 8. Age dependences of paleomagnetic and magnetic properties of peats: NRM (a), k (b) D (c), I (d). Symbols: ( 1 ) original values; and ( 2 ) values smoothed by a 5-point moving average. 0 2 4 Age, ky B.P. 8 6 10 120 0 0 0 6 NRM, 10 –5 A/m k , 10 –5 SI D , deg I , deg –1 1 90 180 –90 –180 45 90 –45 –90 (a) (b) (с) (d) 1 2 444 RUSSIAN JOURNAL OF PACIFIC GEOLOGY Vol. 18 No. 4 2024 PESKOV et al. 4. The 11.2–11.7 ky B.P. interval is only 20 cm thick. A change from negative to positive inclination values occurs at the top of the interval (Fig. 9). 2.4. Analysis of the Time Dependence of the Initial Magnetic Susceptibility in the Tyapka Peat Section Rock-magnetic methods are widely used in paleo- ecological research (for example, [17]), including analysis of initial magnetic susceptibility in sedimen- tary sections. Given a continuous peat section for the entire Holocene, we attempted to reveal a regular, dif- ferent-order periodicity in the time-dependent behav- ior of the magnitude of the initial magnetic suscepti- bility. Figure 8b clearly shows nonmonotonic, sawtooth oscillations of the time-dependent magnitude of the initial magnetic susceptibility. Calculations of 8 har- monics (cosine waves) in the PAST program to approximate the observed time series showed that periods of these harmonics spread over quite a wide range—10590, 334, 1692, 559, 443, 211, 267, and 4614 years, but they differ only insignificantly in abso- lute amplitude values. After excluding from consider- ation harmonics with a period approximately equal to the length of the entire series and probably having no physical meaning, the ratio between absolute values of the largest (about 1700 years) and the smallest (about 4600 years) amplitudes will be only 1.3. Noteworthy also is the presence of 5 harmonics with close periods from 211 to 559 years and close absolute values of their amplitudes (6.37–7.94) × 10 –7 For the purpose of more accurate estimation of the harmonics and their physical interpretation, the fol- lowing two transformations of the original experimen- tal time series of the initial magnetic susceptibility in Tyapka peats were performed: (1) Nine values outside the 95% confidence ellipse were excluded from the observed time series; and (2) The trend (cubic depen- dence) corresponding to the longest-period harmonic determined earlier (see above) was removed from the time series using the least squares method. Then, har- monics for the new time series were estimated in the frequency range from 0 to 0.1 using the multitaper spectral estimation technique in the PAST program (Fig. 10). Analysis of the resulting spectrogram showed that harmonic amplitudes of three frequency intervals of 385–320, 232–216, and 194 years exceed Fig. 9. Age dependences of D values (a), I values (b), and relative paleointensity values (c) of the peats; (d) computed magneto- stratigraphic scale. The values are smoothed by a 5-point moving average. 0 2 4 8 6 10 12 –180 0 180 Frozen peat (3.9–5.9 ky B.P.) 0 –90 90 0 10 Relative paleointensity 5 ~1.85 ky B.P. ~3.1 ky B.P. STERNO– –ETRUSSIA ~5.9 ky B.P. ~7.5 ky B.P. SOLOVKI ~11.2 ky B.P. GOTHENBURG ~0.5 ky B.P. ~0.8 ky B.P. ~3.5 ky B.P. (а) (b) (с) (d) D , deg I , deg Frozen peat (3.9–5.9 ky B.P.) Frozen peat (3.9–5.9 ky B.P.) Age, ky B.P. RUSSIAN JOURNAL OF PACIFIC GEOLOGY Vol. 18 No. 4 2024 HOLOCENE GEOMAGNETIC EXCURSIONS IN PEAT DEPOSITS 445 the 95% confidence level, and for the first two inter- vals the confidence level is greater than 99%. Note the interval of 528–492 years, the confidence level of which exceeds 90%. A reasonable explanation and physical justification for the simultaneous presence of all the above period- icities in the time series of the initial magnetic suscep- tibility is hardly possible. This is also evidenced by a visual analysis of the original time series; different quasi-periodicity is clearly observed for individual intervals of the peat section. As examples, time depen- dences are given for the intervals of 0–2000 and 4000– 6000 years which exhibit different quasi-periodicity with characteristic times from 120 to 500 years (Figs. 10b and 10c). This was also confirmed by spec- tral analysis with estimation of the four largest in amplitude harmonics for five intervals (Fig. 10d), based on which it could be argued that in certain sec- tions of time dependence of magnetic susceptibility there are significantly different, but close in time, quasi-periodic components from 120 to 550 years. At the next stage, wavelet analysis was applied [19, 30], since it is better suited than parametric methods for the analysis of non-stationary signals, and the time series of the initial magnetic susceptibility is precisely this kind of signal. The wavelet transform does not simply “cut” the time series under study into several intervals, but extracts from it different-scale compo- nents, and each component is analyzed with that degree of time-varying detail which corresponds to its scale. Moreover, this allows the presentation of all the periods of interest on one diagram and the removal of even harmonics. Having applied the time-frequency decomposition of the time series of the initial mag- netic susceptibility, one can see how the dominant modes vary with time (Fig. 11a): (1) Predominance of a single oscillation with a period of about 1500 years in the interval 11500–6000 years, which then “breaks up” into two: one with a period of approximately 2400 years in the interval 600-0 years, the other approxi- mately 1000 years in the interval 3500-0 years; and (2) Mosaic nature of the density spectrum in the fre- quency range between 0.008 and 0.002 (120– 520 years), where in the interval of approximately 11500–4500 years, the highest densities are typical for harmonics with characteristic time of a little more than 500 and 300 years, and in the interval 4500- 0 years, harmonics with characteristic time of 150– 250 years. Interpretation of the presence of quasi-harmonic oscillations with periods of approximately 150 to 550 years does not pose any particular difficulties. Firstly, similar periodicities have been reported in numerous studies to determine wildfire frequency in the Holocene: 320 years for Swedish boreal forests [4], 180 and 430 years for Finnish boreal forests [7], 100 and 176 years for the coastal forests of Vancouver Island, Canada [3], 117 and 330 years for the northern European part of Russia [20], 100 and more than 200 years for the Daihai Lake region of northern Cen- tral China [34]. According to the authors of the last publication, an increase in wildfire frequency starting around 2800 cal. BP is the result of agricultural prac- tices of people. Secondly, determination of the natural (physical) process which is the main, perhaps the only, cause of this quasiperiodicity. After exposure to elevated temperatures during forest and steppe fires, goethite ( α -FeOOH) or hydrogoethite ( α -FeOOH⋅ n H 2 O) occurring in the surface soil and being weak ferromag- nets, pass into a more magnetic phase, such as maghemite ( γ -Fe 2 O 3 ), which is a ferrimagnet (e.g., [8]). It is not at all necessary for peats to burn; maghemite can be transported by wind and water streams. Enhanced initial magnetic susceptibility cor- responding to the quasiperiods indicated above is clearly seen in the plot (Figs. 10b, 10c). Interpretation of harmonics with periods longer than 1 thousand years (Figs. 10a, 10e) is ambiguous and involves difficulties. On the one hand, such irreg- ular periodicity with characteristic times from 800 to 2300 years is known for the Holocene and has a proper name, that is, Bond cycles (events) [1]. On the other hand, a clear definition of the connection (natural physical process) per se between behavior of the mag- netic susceptibility in peats during the Holocene and Bond events is not at all simple and apparent. A model time series of the magnetic susceptibility was calculated to correlate the experimental time series of the magnetic susceptibility in the peats with the Bond events by parametric interpolation using the fast Fourier transform algorithm. Three main compo- nents with periods of 5317 (46.7%), 1700 (26.4%), and 356 (26.89%) years were selected by the least squares method to construct the model time series, whose phases relative to the present time are 262 ° , 315 ° , and 13 ° , respectively (Fig. 11b). The comparison shows good agreement between the model time series (purple line) and discrete values of the initial magnetic suscep- tibility after trend removal (points), with almost all points within the 95% confidence interval of the inter- polated dependence and only few outside (Fig. 10a). The characteristics and a possible cause for quasi- periodicities with characteristic times from 120 to 500 years were considered above, therefore we will not discuss the harmonic with a period of 356 years (green color, Fig. 11b), which was calculated in the course of parametric time series analysis. We will not discuss a harmonic with a period of more than 5317 years (red color, Fig. 11b), since there are only two such cycles in the time interval under consideration, which is cer- tainly not enough to isolate with confidence such a harmonic. We now focus on the harmonic of 1700 years (Fig. 11b), more than 6.5 cycles of which are within the considered interval, that is, the Holocene with 446 RUSSIAN JOURNAL OF PACIFIC GEOLOGY Vol. 18 No. 4 2024 PESKOV et al. nine Bond events. Analysis of (Fig. 11b) shows that there is a time-based correlation between most of the Bond events with extreme values of the harmonic. A clear correlation is observed for four Bond events asso- ciated with cooling: 1.4 ka (Late Antique Little Ice Age), 2.8 ka (Cold Age of the Iron Age), 8.2 ka (Global cooling caused by the Laurentide Ice Sheet collapse), and 9.4 ka (Erdalen event associated with a surge of glacial activity in Europe). There is also a clear tempo- ral correlation between two Bond events related to aridization and drought and an increase in the average global temperature, that is, the origin (4.2 ka) and evo- Fig. 10. Analysis of the time dependence of the initial magnetic susceptibility in the Tyapka peat section. (a) Estimation of time- series spectral density using the multitaper power spectral density estimation method [13] after trend removal. (b, c) Examples of time-dependent behavior of initial magnetic susceptibility in peats for intervals of 10-2000 and 4000–6000 years, respectively. (d) Period estimation for the first four harmonics using the Lomb-Scargle method [18, 29] for individual time series intervals after trend removal. Time, years 0 1000 2000 3 (b) 0 4000 5000 6000 4 Time, years (с) 0 k ,10 –6 (normalized) k ,10 –6 (normalized) Period, years Std er., years % 1177 156 16 554 34 19 366 10 41 243 5 24 1736 232 39 539 36 19 155 3 18 134 2 24 826 54 39 436 20 21 218 5 25 117 2 14 1766 93 66 515 21 14 354 11 11 230 5 9 2695 385 27 497 16 32 331 8 26 217 5 15 0–2000 2000–4000 4000–6000 6000–8000 8000–10000 (d) Frequency (period, years) 90% 95% 99% Power, 10 –9 (a) 4 3 2 1 0 0.001 (1000) 0.002 (500) 0.003 (333) 0.004 (250) 0.005 (200) 0.006 (167) 0.007 (143) 0.008 (125) 0.009 (111) RUSSIAN JOURNAL OF PACIFIC GEOLOGY Vol. 18 No. 4 2024 HOLOCENE GEOMAGNETIC EXCURSIONS IN PEAT DEPOSITS 447 lution (5.9 ka) of the Sahara Desert. Three other Bond events, with two (0.5 ka and 11.1 ka) associated with cooling and one (10.3 ka) with warming, have no dis- tinct temporal correlation with the magnetic suscepti- bility behavior in the Tyapka peat section. 2.5. Trajectory of Virtual Geomagnetic Poles during the Etrussia Excursion The smooth virtual geomagnetic pole (VGP) path during the Etrussia excursion calculated from the paleomagnetic data is in Fig. 12 which shows that at the onset of the excursion, the path is traceable from the polar latitudes of the northern hemisphere approx- imately along the 180-degree meridian to the equato- rial latitudes. Then, it has two clockwise loops over the Pacific Ocean encompassing the adjacent areas of Eurasia and North America to end in the northern part of the North American continent. Similar clockwise looping is observed for the Lachamp excursion [15], but in this case, the area cov- ered by the VGP path is significantly larger, which is probably due to the duration of the excursions. From U-Th dating data on speleothems [16] it follows that the Lachamp excursion occurred in the range of 42.25–39.70 thousand years, that is, its duration was more than 2.5 thousand years, while the Etrussia excursion, according to our data, was about 1.2 thou- sand years, and according to other sources (for exam- ple, [9]) even less. Fig. 11. Analysis of time dependence of the initial magnetic susceptibility in the Tyapka peat section. (a) Morlet wavelet diagrams [13] of the time series after trend removal. (b) Model time series (purple line) of the initial magnetic susceptibility in the peat sec- tion with the 95% confidence interval (black lines) and initial values of the magnetic susceptibility of the smoothed series (points) calculated from three main harmonics with periods of 5317 (red line), 1700 (blue line) and 356 (green line) years. The arrows show a possible correlation between the time-dependent behavior of the harmonic of 1700-year periodicity and the Bond events [1]. 2435 1448 861 512 304 181 107 64 0 1500 3000 4500 6000 7500 9000 10500 Time, years Period, years 0.25 0.50 0.75 1.00 0 Magnitude–squared coherence (а) 1.881e–4/5317 ± 516 5.