science.sciencemag.

org/content/371/6531/811/suppl/DC1

Supplementary Materials for

A global environmental crisis 42,000 years ago

Alan Cooper*†, Chris S. M. Turney*†, Jonathan Palmer, Alan Hogg, Matt McGlone,
Janet Wilmshurst, Andrew M. Lorrey, Timothy J. Heaton, James M. Russell,
Ken McCracken, Julien G. Anet, Eugene Rozanov, Marina Friedel, Ivo Suter,
Thomas Peter, Raimund Muscheler, Florian Adolphi, Anthony Dosseto, J. Tyler Faith,
Pavla Fenwick, Christopher J. Fogwill, Konrad Hughen, Matthew Lipson, Jiabo Liu,
Norbert Nowaczyk, Eleanor Rainsley, Christopher Bronk Ramsey, Paolo Sebastianelli,
Yassine Souilmi, Janelle Stevenson, Zoe Thomas, Raymond Tobler, Roland Zech
*These authors contributed equally to this work.
†Corresponding author. Email: alan.cooper@blueskygenetics.com (A.C.);
c.turney@unsw.edu.au (C.S.M.T.)

Published 19 February 2021, Science 371, 811 (2021)

DOI: 10.1126/science.abb8677

This PDF file includes:

Materials and Methods

Figs. S1 to S34
Caption for Table S1
Tables S2 to S8
References

Other Supplementary Material for this manuscript includes the following:

(available at science.sciencemag.org/content/371/6531/811/suppl/DC1)

Table S1 (.xlsx)
Materials and Methods

Ancient kauri (Agathis australis) from Northland, New Zealand

Four bog-preserved kauri trees associated with the Laschamps Excursion (hereafter
Laschamps) were obtained from Ngāwhā Springs (Nga001; site location: -35.403°,
173.859°), Kai Iwi Lakes (FIN09 and FIN14058; site location: -35.810°, 173.631°) and
Mangawhai (NZKL1855; site location: -36.143°, 174.609°) in northern New Zealand
(Fig. S1). Nga001 came from a power plant development site in the central part of
Northland and appears to have been encased by sediments associated with geothermal
activity (Fig. S2); the first time that we are aware of a kauri log being found encased in
geothermal deposits of any kind. FIN09 and FIN14058 are cross-sections from logs that
were extracted on two (nearly) adjacent farms, along the west coast of Northland that are
owned by the Finlayson family. These two sites are positioned within a late Pleistocene
parabolic dune belt that contains pockets of wood-rich peat deposits. NZKL1855 is a
section of a kauri table top that was obtained after the primary log had been milled; this
sample is believed to originate from Black Swamp Road, Mangawhai, which is
positioned within a Northland east coast relic foredune ridge sequence (of Last
Interglacial age) facing the Pacific Ocean that contains abundant late Pleistocene (Marine
Isotope Stage 3) peat and wood (24).

Multiple cross-sections (or ‘biscuits’) were cut from Nga001, FIN09, FIN14058, and
radial strips were removed for measurement while the residual offcuts were archived
(Fig. S3). Only one set of radial sections was obtained from NZKL1855 because it was a
tabletop panel pre-cut from the center of a tree trunk. When NZKL1855 was cut in the
mill yard, pith was visibly offset from the center of the table, creating two radii of
unequal lengths consistent with asymmetric growth of that tree (common for kauri). All
of the radial strips were dried and then sanded progressively from coarse to fine grit
paper until a highly polished surface was obtained. The finished radii were studied under
a binocular microscope and the annual rings identified and counted. Following this step,
rings were measured using standard dendrochronology techniques employing a travelling
stage with an incorporated digital linear encoder.

The four subfossil kauri tree-ring series associated with the Laschamps did not indicate
any immediate matching patterns of growth (i.e. cross matching). The absence of
dendrochronological cross matching could be due to multiple factors, including
idiosyncratic issues related to individual trees (e.g. NZKL1855 is visibly damaged
and suppressed towards the bark edge), the type of sample (FIN14058 is likely a branch,
despite being nearly one meter in diameter), or attempting comparisons between a
juvenile growth period (i.e. Nga001) with senescent tree phases (i.e. FIN14058,
FIN09, NZKL1855). All are known issues that can hamper linking kauri tree-ring
sequences, especially when sample numbers are limited. As such, the internal ring
sequence count is considered robust for each of these trees, but there is currently
no dendrochronological “lock” between the four samples (and their associated sequential
14
C dates). Future work may help to refine this issue if sufficient coeval samples are
found.
1
Comparisons of the average radial growth rates (mm/year) were made between the
ancient Ngāwhā tree and modern kauri tree-ring collections as well as other subfossil
samples from the late Pleistocene (Fig. S4). The results clearly show that the Ngāwhā tree
(Nga001) was noticeably slower growing than modern trees and other samples from the
late Pleistocene. Furthermore, the Nga001 ring-width sequence (shown in Fig. S4A)
shows a sustained decline in growth that coincides with the Adams Event and into
the Laschamps. There is a natural tendency for radial ring widths to decline slowly over
the life of a tree (commonly described as the age-related growth trend); however, the
Ngāwhā tree shows an abnormally slow growth pattern indicating consistently
unfavorable growing conditions.

The internal ring sequence mark-up for each radial sample was extended onto adjacent
offcuts to enable destructive sampling for radiocarbon dating. Four-decade block samples
were marked on each of the radii where feasible, with slight variation in some cases
where ring width suppression hampered obtaining sufficient weight for high precision
liquid scintillation spectroscopy (starting dry wood weights between 70-80g are required
to produce 7.5g benzene aliquots). Each multi-decadal sample was removed from either
the radii or an offcut via careful chiseling along annual boundaries. As such, the internal
age incorporation for most samples are 40 years. Chronological error due to missing rings
for each tree is estimated to be <1% and would be isolated within millennial scale ring-
sequences for each tree (43, 44). Across the multi-millennial sequential radiocarbon data
set that has been developed in this study, the issue of missing rings is not considered a
significant element that would contribute to compression or drift of the internal tree ring
time scales and the spacing of adjacent 14C blocks through time (43-45).

Sedimentary Sequences
Pillar Rock, subantarctic Auckland Islands: Located in the Pacific Sector of the
Southern Ocean, the New Zealand subantarctic Auckland Islands (50°S, 166°E; Figs S1
and S5) are the eroded remnants of extensive Oligocene to Miocene basaltic volcanism
(46), with rugged topography, reaching up to 664 m.a.s.l. The islands lie in the southwest
Pacific in the core of Southern Hemisphere westerly winds and experience the same
climate regime with relatively mild mean annual temperatures (47). Situated today
between the Subtropical Front (approximately 45° S) and the cool Antarctic waters of the
Antarctic Convergence (around 55°S), these ocean fronts may have shifted by up to 5°
further northwards during the LGM (48) leading to a mean annual air temperature
depression over the islands of ~6°C (49). The Auckland Islands are therefore highly
sensitive to regional climate change during the late Pleistocene (12, 49, 50).

The exposed sedimentary sequence on a north-facing cliff near Pillar Rock, Auckland
Island (-50.518˚, 166.217˚; 30 m.a.s.l.; Fig. S6), consists of glacial till overlain by a
sequence of organic and non-organic units, providing important constraints on late
Pleistocene environmental change in this sector of the Southern Ocean. The exposed
section consists of a laterally continuous (~10 m), horizontally bedded section exposed
by extensive cliff erosion along the north coast of Auckland Island at ~30 m.a.s.l. In
places the section is over 4 m thick, and rests on several meters of deeply indurated

2
diamicton, that we correlate with the upper till exposed on nearby Enderby Island (12).
The ‘Enderby Till’ at Pillar Rock is overlain non-conformably by stratified sands and
gravels, which are iron banded. This unit is overlain by a ~2 m thick conformable
sedimentary succession, that grades upwards from massive silts and sands into laminated
silts and clays and into an organic-rich silt band (Fig. S7). Above this organic silt the
succession continues, with a conformable, non-erosive boundary into sandy silts, which
become increasingly clast rich and less dense. This conformable sequence is topped by a
10 cm thick clast-supported gravel, topped by a 2 cm layer of flat lying gravel/pebbles.
This succession is overlain unconformably by a 27 cm thick sequence of early Holocene
humified dark brown peat (Fig. S8). Above the peat, there is a series of horizontally-
bedded silty sandy loams and low organic soil horizons, with pebble horizons, overlain
by ~20 cm of sandy loam at the surface. The organic units at Pillar Rock were sampled
with monolith tins. Sediments were described in the field, with samples wrapped and
transported back to the laboratory where they were cold-stored (4˚C) prior to detailed
pollen analysis and radiocarbon dating.

A 26-cm long rectangular sedimentary section was dug out of Pillar Rock, capturing an 8
cm layer of highly compressed lignitic peat with sharp boundaries (Fig. S7), sitting on
organic rich silts with weathered boulders, and overlain by grey-brown sandy silt with
abundant stones. The lignite layer thus represents a stable interval in a prolonged period
of periglacial activity. The entire lignite layer was sampled for pollen at 5 mm thick
contiguous samples (1.25ml volume of peat per sample). Samples for radiocarbon dating
were taken from the top and bottom of the lignite, as 5 mm thick samples at the sharp
boundaries, and contiguously down the profile at the same depths as the pollen samples.
Pollen preparation followed standard techniques (51). Pollen was counted under x400
magnification until a sum of 250 grains were reached (excluding fern spores) and pollen
expressed as a percentage of this sum.

Lake Towuti, Sulawesi: Lake Towuti is located at 2.5°S, 121.5°E on the island of
Sulawesi in central Indonesia within the Indo-Pacific Warm Pool (IPWP) (11) (Fig. S9).
The site lies at the center of the humid, unstable air mass overlying the IPWP, and its
climate responds strongly to large-scale changes in regional atmospheric circulation and
sea surface temperature associated with the Australian–Indonesian summer monsoon
(AISM). The Lake Towuti basin receives ∼2,700 mm of precipitation annually and is
surrounded by dense closed-canopy rainforest (52). Today, this region experiences a wet
season from December to May when the Intertropical Convergence Zone (ITCZ)
migrates southward over the region (53).

Lake Towuti is the largest tectonic lake in Indonesia, and at 205 m depth, its sediments
preserve perhaps the longest and most continuous terrestrial record of climate available
from the region. In 2007–2010, 13 sediment piston cores were recovered from Lake
Towuti (11). Here we focus on the most continuous radiocarbon-dated stratigraphy from
core TOW10- 9B (Fig. S10). As part of a comprehensive multi-proxy study, carbon-
isotopic composition of long-chain, even-numbered n-alkanoic acids (δ13Cwax) was
measured, a main component of plant epicuticular waxes. The δ13Cwax is primarily used to
distinguish between plants using C3 and C4 photosynthetic pathways (54) because C4

3
plants use a CO2-concentrating mechanism that improves their photorespiration and
water-use efficiency relative to C3 plants (55). Precipitation is the dominant control on
the distribution of C3 and C4 plants through much of the tropics, such that δ13Cwax has
been widely used to reconstruct past changes in tropical hydroclimate (11, 56, 57). Global
surveys of the δ13C of vegetation indicate these processes can cause ∼6‰ depletion in
the δ13C of leaf matter in tropical rainforests relative to more xeric C3 ecosystems, and
∼4‰ depletion in tropical rainforests relative to drier, tropical deciduous forests (58, 59).
Thus, we interpret more depleted δ13Cwax to represent C3 forests growing in wet
conditions; whereas, enriched δ13Cwax reflects a drier climate and increasing C4 grasses
(11).

Previous work has reported the δ13Cwax values. To summarize, lipid extraction,
purification, and isotopic analysis followed the procedures outlined in ref. (11). Between
2 and 10 g of freeze-dried, powdered sediment samples were extracted using a Dionex
350 accelerated solvent extractor with a 9:1 solution of CH2Cl2:MeOH. Acid fractions of
lipid extracts were separated from neutral and polar fractions using aminopropyl silica gel
column chromatography with 2:1 CH2Cl2:isopropanol and 4% acetic acid in anhydrous
ethyl ether as eluents. The acid fraction was methylated using acetyl chloride in methanol
of a known isotopic composition to allow isotopic correction of fatty acid methyl esters
(FAMEs). FAMEs were further purified using silica gel columns with hexane and
CH2Cl2 as eluents. The CH2Cl2 fraction containing saturated FAMEs was dried under N2
gas and then dissolved in toluene before isotopic analysis. Carbon isotopes of the leaf
waxes (δ13Cwax) were run in duplicate or more on a Thermo Finnegan Delta XL gas
chromatograph isotope ratio mass spectrometer at Brown University and are reported
relative to Vienna Pee Dee Belemnite. All δ13Cwax data are corrected for the isotopic
composition of the methyl group added during methylation. 145 samples were analyzed
for the δ13C of C28 n-acid, the most abundant homolog in our samples (Fig. S11). 119 and
58 samples had sufficient abundances to measure the δ13C of C26 and C30 n-acid,
respectively. The average difference between duplicated samples in the δ13C of C28 n-acid
is 0.17‰, and the pooled SE of samples (n = 20) measured in triplicate or greater is
0.47‰.

The core TOW10-9B was also subjected to measurements of natural remnant

magnetization (NRM) as well as NRM20mT, which is the NRM after alternating field (AF)
demagnetization to a peak field of 20 mT (60). Here, the AF demagnetization was
demonstrably effective in isolating the secondary NRM carried by large multidomain
(MD) grains, which have a coercivity ≤20 mT (61). The core was also subjected to
measurements of ARM20mT and IRM20mT, which are the intensities of ARM and SIRM
after an AF demagnetization to a peak field of 20 mT, respectively. The ARM
susceptibility, denoted by κARM, was calculated by dividing the measured ARM with
the value of a superimposed steady magnetic field of 0.05 mT (62). The measurements of
NRM20mT, ARM20mT and IRM20mT were carried out in the Paleomagnetic Laboratory of
the University of Rhode Island using a 2G-Enterprises cryogenic magnetometer with a
measurement interval of 1 cm (Fig. S11). In the Lake Towuti sediments which contain
high concentrations of MD grains, the paleointensity values were calculated by

4
normalizing the NRM20mT values with respect to other magnetic parameters in order to
isolate the effects of the geomagnetic field.

Radiocarbon Dating and Calibration

Here we report radiocarbon (14C) ages as thousands of years Before Present (CE
1950) i.e. ka 14C BP, with calibrated 14C and calendar ages given as thousands of calendar
years (ka) Before Present.

Ancient kauri (Agathis australis): For radiocarbon dating the ancient kauri, chemical
pretreatment of the four decadal-long wood samples was carried out to extract alpha-
cellulose – the wood fraction deemed the most reliable for minimizing potential
contamination – and measured using High-Precision Liquid Scintillation Counting (HP-
LSC) at the University of Waikato, in order to provide the most robust 14C ages which are
required for such high-precision study (Table S1) (63). Wood pretreatment protocols
followed those outlined in Hogg et al. (64). Seventy-gram samples were ground to pass a
20-mesh sieve. Soxhlet apparatus was used to extract resins utilizing acetone for 6 hours
followed by distilled water for 6 hours. Holocellulose was produced by repeated
bleaching in acidified NaClO2 (15 g/L). The alpha-cellulose fraction (~40% of initial
wood weight) was obtained by a final extraction in 5% NaOH (under N2 gas), followed
by acidification in 5% HCl and washing in distilled water. Radiocarbon ages were
determined by decay counting in Perkin Elmer Wallac QuantulusTM spectrometers.
Cellulose samples were converted to benzene in vacuum systems preconditioned for
activities of approximately seven half-lives, utilizing 7.5 g benzene samples contained
within 'Waikato' synthetic silica counting vials (65) and measured for a minimum of
10,000 minutes per sample. The Modern and Background standards (OXII and Oxygen
Isotope Stage-5 Renton Road sub-fossil kauri wood, respectively) were prepared and
measured in a similar manner. Background blank activity derived from repeat (13
measurements) over a two-year period of the Renton Road kauri alpha-cellulose was
equivalent to 0.00041 ± 0.00009 times Modern (~62.7 kyr BP), with the Background
blank uncertainty obtained from the Gaussian standard deviation of all measurements.
Isotopic fractionation correction was achieved by measuring 13C for all samples using a
Europa Scientific Penta 20/20 isotope ratio spectrometer.

Calibration Curve: A 14C kauri-Hulu Southern Hemisphere (SH) calibration curve was
constructed to cover 30–50 ka (Figs. S12 and S13). The sub-fossil kauri 14C
determinations were combined with the 249 14C determinations from the two 230Th-dated
Hulu Cave speleothems (MSD and MSL) (22) during this time period. Curve
construction was performed using the same Bayesian spline with errors-in-variables
statistical approach used for the IntCal20 and SHCal20 curves (66-68). Non-informative
priors were placed on the absolute calendar ages of the oldest 14C determination of each
floating kauri tree-ring sequence. The robust, internal, ring sequence counts then
provided relative and known age increments with respect to these oldest calendar ages.
Within the Markov Chain Monte Carlo (MCMC) fitting procedure, in addition to
calibration curve construction, posterior absolute calendar ages for each floating tree-ring
sequence were simultaneously estimated based upon fit with both the Hulu data and the
other kauri.

5
The Hulu cave speleothem data were modelled to have an offset ri , measured in
radiocarbon years, to the atmospheric SH kauri due to the dead carbon fraction (DCF)
present in the Hulu speleothems and the North-South (NS) interhemispheric radiocarbon
offset. The offset for any Hulu determination was considered as independent and
identically distributed ri ∼ N(ν, ζ2 ) i.e. additional independent variation around a
constant, unknown mean. Here ν represents the unknown mean kauri-Hulu offset and ζ
the level of independent variation around the mean. A prior was placed on the mean
kauri-Hulu offset of π(ν) ∼ N(480 − 36 = 444, 1002 ) based upon the 480 14C yr mean
DCF offset seen between the H82 Hulu cave speleothem (69) and the Northern
Hemisphere tree ring determinations of IntCal20 (66); with the 36 14C yr mean NS
radiocarbon offset of SHCal20 subtracted (68). Since these initial offset estimates were
based upon the 0–14 ka time period, as opposed to 30-50 ka, a difference in mean of ±
100 14C yr (1σ) for our period of interest was considered prudent. The nature of
speleothems means that the radiocarbon signal recorded in the Hulu Cave is also believed
to be somewhat smoothed, so that atmospheric radiocarbon fluctuations on short
timescales are effectively lost (23). The independent uncertainty ζ around the mean offset
aims to account for this along with both potential annual variation in the Hulu DCF and
the NS radiocarbon offset. The Hulu H82 speleothem is seen to require an additional
uncertainty of ~50 14C yr for consistency with atmospheric tree ring determinations (67),
while the NS offset in SHCal20 has an uncertainty of ~27 14C yr (68). Adding in
quadrature provided an estimate for ζ, the additional independent uncertainty in the kauri-
Hulu offset for any determination, of 57 14C yr.

The Bayesian spline used to model the calibration curve was constructed using 150 knots
to permit potential fine scale detail in the resultant curve to be captured while still
enabling practical estimation. In the 40–45 ka period that covered the potential fitting
locations of the floating kauri trees, 103 evenly spaced knots were used in order to
penalize all possible calendar age placements of the kauri equitably (Table S1). Outside
this range, knots were placed at quantiles of the observed Hulu data. A non-informative
prior was placed on the level of smoothing. The MCMC was run for 250,000 iterations.
The first 125,000 of these iterations were discarded as burn-in. The remaining 125,000
were used to create the final calibration curve. This curve is shown in Figs S12 (in Δ14C)
and S13 (in radiocarbon age), and the posterior calendar age estimate of each kauri in
Fig. S14. Further, 500 individual calibration curve realizations from the posterior (each
representing a plausible historical record of SH atmospheric radiocarbon) were extracted
for use in synchronizing our radiocarbon and ice core chronologies.

For ~800 years prior to the prominent Δ14C increase captured in our composite
radiocarbon record, we also observe a quasi-centennial oscillatory data structure that
tapers from high to low amplitude that is followed by a punctuated radiocarbon
enrichment starting at ~43 ka (68). We consider this pattern in the radiocarbon data to be
sound because of the demonstrable reproducibility in the 14C measurements, and
similarity to previously published records (22, 70). This pattern is also independently
replicated in different ways via similar characteristics seen in 10Be variations captured in
ice cores (Fig. 1) (21). The 14C centennial-scale pattern prior to the Adams Event is

6
comparable to the Holocene 207-year de Vries solar cycle pattern that has also been
observed in the Greenland ice core 10Be data around Laschamps (71), so may represent
low-frequency solar variability. The levels of atmospheric radiocarbon remain elevated
after the Laschamps for >1000 years as has been observed previously, most likely a
consequence of global carbon cycle changes, including Heinrich Event 4 (22).

Peat and lake sediments: The contiguous peat and lake sediments analyzed here were
given an acid-base-acid (ABA) pretreatment. Because of the relatively small amount of
material available (compared to the kauri), samples were measured by accelerator mass
spectrometry (AMS). Here the pretreated peat and lake sediments were combusted and
graphitized in the University of Waikato AMS laboratory, with 14C/12C measurement by
the University of California at Irvine (UCI) on a NEC compact (1.5SDH) AMS system.
The pretreated samples were converted to CO2 by combustion in sealed pre-baked quartz
tubes, containing Cu and Ag wire. The CO2 was then converted to graphite using H2 and
a Fe catalyst, and loaded into aluminum target holders for measurement at UCI. To
generate an age model for the sequences, we used the new combined kauri-Hulu Cave
calibration dataset generated here (see above) using a P_sequence deposition model with
the General Outlier analysis option (with probability value of 0.05) in OxCal 4.2 (ref. (72,
73)) (Tables S2-S5). For Pillar Rock and Lake Towuti, previously ages reported by ref.
(12) and ref. (11) respectively were integrated in the age models (Figs S6 and S10
respectively); to accommodate a changing sediment reservoir age (ΔR) through the Lake
Towuti sequence (11), a Delta R with the prior U(0,1500) was used (Table S4).

Synchronizing radiocarbon and ice core chronologies

To compare our Δ14C record to independent estimates of cosmogenic radionuclide
production we used the ice core 10Be-stack as compiled by ref. (21). For such a
comparison, we need to explore i) timescale differences between our 14C and ice core
10
Be-records, ii) differences that may arise from a different state of the carbon cycle,
specifically the deep ocean ventilation, and iii) differences that may arise from either a
so-called “polar bias” of the 10Be-record, or a different sensitivity of 14C and 10Be
production rates to variations in the cosmic ray flux as predicted by some production rate
models (74).

Using 500 realizations of our Δ14C record as the atmospheric input, we modelled 14C-
production rates using a box-diffusion carbon cycle model (75, 76). For each realization
we determined for which combination of i) timescale shift, ii) state of deep ocean
ventilation (constant values of the model’s ocean diffusivity parameter from 10% to
100% preindustrial in steps of 10%) and iii) scaling of the 10Be record we obtain the best
fit (lowest root mean square error) between p14C and 10Be. The resulting histograms are
shown in Fig. S15 (panels B-D). The most likely combination of parameters (applied in
panel a) is i) a timescale difference of 265 years (GICC05 being younger), ii) a less
efficient mixing of the deep ocean (70% its preindustrial value), and iii) a scaling of 10Be
by a factor of 1.3 (indicating a slight polar bias or lower sensitivity of 10Be to variations
in the cosmic ray flux). These findings are in excellent agreement with earlier findings
(21) but significantly more precise owing to our high-quality and high-resolution truly
atmospheric 14C record.

7
The Earth’s geomagnetic field during the Laschamps Excursion
The Laschamps paleomagnetic record from the Black Sea is the best dated and most
detailed record of this excursion from the Northern Hemisphere (5, 77). Its age model is
constrained by 16 AMS 14C ages, improved by fine-tuning high-resolution records of
IRD (ice-rafted detritus) counts as well as records of element ratios of Ca/Ti and K/Ti
obtained from XRF (X-ray fluorescence) scanning of Black Sea sediments to the NGRIP
oxygen isotope record (41, 42) on the Greenland Ice Core 2005 (GICC05) chronology
(43, 44). A detailed analysis of an improved record of the Black Sea paleomagnetic data,
spanning from 14.5–68.9 ka, has been performed by Liu et al. (5). By comparison of
relative paleointensity data to absolute paleointensity data (42, 46, 47), a full-vector
paleomagnetic record can be established. Then, the three spherical vector components
(inclination, declination, absolute intensity) were transformed into Cartesian components
with one of them being oriented along the direction expected from a pure geocentric axial
dipole. Most of the field decay is seen in this component. Perpendicular to it, in the
second component in East-West direction and in the third component in the vertical
North-South plane, only multipolar components are seen. These exhibit much smaller
oscillations, but of more or less persistent amplitudes across the excursion. This means
that the field decay is mostly related to the dipole component which is a global signal.
Similar to the first study (5, 77) a correction factor of 14.52×1022 Am2 was used to
convert relative paleointensities into Virtual Axial Dipole Moments (VADM).

The lowest field values during the Laschamps recorded in Black Sea sediments were
obtained for the N-R transition at 41.63 ka with VADM values as low as 0.48×1022 A m2,
just ∼6.3% of the modern axial dipole moment in 2010 with 7.628×1022 A m2 (according
to the 11th generation International Geomagnetic Reference Field, IGRF-11) (48). For the
fully reversed phase of the Laschamps from 41.35 to 41.11 ka the VADM recovers to a
mean value of 2.12×1022 A m2, about 28% of the modern axial dipole moment, and then
drops down to 1.14×1022 A m2 at 40.95 ka during the R-N transition. An intermediate
VADM value of around 4.00×1022 A m2 is then reached at about 40.23 ka with directions
already according to a normal polarity. Other studies have suggested even lower values of
paleointensity, potentially down to 0% modern axial dipole moment (2). The field
minima bracketing the Laschamps fully reversed phase were asymmetric not only in
terms of field strength, but also having different durations. If a large directional precursor
is taken into account then the full N-R transition took around 1000 years (5), while the R-
N transition-related minimum lasted only about 2-300 years (Fig. 1). In general, the
Black Sea paleointensity record conforms with other (less detailed) paleointensity stacks,
such as the GLOPIS75 (49) and is therefore considered appropriate for comparison to the
sequences reported in this study.

Atmospheric impacts of geomagnetic changes associated with the Laschamps Excursion

The Earth’s Radiation Environment: To gauge the likely impacts of Laschamps on
multiple geophysical parameters, we have used previous solar, cosmic ray, and space
research studies and examples as a guide (Fig. 5). In the present epoch, the geomagnetic
field prevents a large proportion of the cosmic radiation <14GeV/nucleon from accessing
the atmosphere. During the Laschamps however, the greatly reduced geomagnetic field

8
strength meant that nearly all of the cosmic radiation reaching the orbit of Earth gained
access to the atmosphere at all latitudes (that is, the “cosmic ray cut-off rigidity” was
<1.0GV). In addition, solar activity (as indicated by the sunspot cycle) results in further
changes in the radiation reaching Earth of up to orders of magnitude with observed
periodicities 11<T<10,000yrs. All these variations will influence the tropospheric and
stratospheric ionization.

There are two major sources of cosmic radiation that produce the ionization in the
atmosphere that influence the ozone column and UV incident on Earth; galactic and solar
cosmic radiation. The galactic cosmic radiation (GCR) originates in supernova
throughout the galaxy; has an average age of 2x106 years; an energy spectrum
approximating E-2.65 and has been essentially invariant over the past 100 ka. The
spectrum is referred to as the “local interstellar spectrum”, and has now been measured in
situ by the Voyager 1 spacecraft that exited the solar system in 2012 (78). The second
source is the explosive acceleration of cosmic rays following the annihilation of portions
of the extremely strong sunspot magnetic fields in large solar flares. Solar cosmic
radiation has a vastly different spectrum from that of the GCR, sometimes approximating
E-6, and intensities at energies at and below 10Gev that can be many orders of magnitude
greater than those of the GCR (79). These short-lived events, ranging from 1 to 10 days,
are referred to as “solar energetic particle” (SEP) events (previous acronym SPE for solar
proton event). Around 71 SEP have been observed by ground level detectors since 1942,
and hundreds by satellite detectors since 1963. Based on that knowledge we use the SEP
event of 23 February 1956, the largest event observed since 1942, to estimate the effects
of SEP during the Laschamps (Fig. S16). Based on neutron monitor and ionization
chamber (ground level) records for the interval 1942–2012 (79) we estimate that there
were approximately 3 SEP per decade with intensities within a factor of ten of those
displayed in Figures S16. Studies of several SEP events since 1942 indicate increases in
atmospheric ionization on the order of 90 times (80), which would have had far more
impact at equatorial and low latitudes during the Laschamps due to the reduced
geomagnetic cut off rigidities. The visual impacts on the atmosphere at night during these
SEP periods may have been considerable, potentially influencing paleolithic cave artists
(see below). Cosmogenic data indicate that there have also been “super-SEP” events in
the past with fluences up to 40-50 times those in Fig. S16 with a frequency of about two
per millennium (81, 82).

The variability and intensities of both the GCR and SEP are strongly dependent on solar
magnetic activity (“the sunspot cycle”). The very tenuous low energy “solar wind”
transports the solar magnetic field to the limits of the solar system, at the termination
shock 120 astronomical units (1AU= distance from Sun to Earth) from the Sun. En route,
the solar magnetic field partially reflects and also decelerates the GCR ( “modulation”),
resulting in substantial variations in the intensity reaching Earth. Historical optical
observations going back to Galileo (1609 CE) and before that to the Chinese (to 3 ka BP),
show that there is a persistent ~11 year periodicity in solar activity. In addition, there
have been repeated periods of “Grand Solar Minima” (26 in the past 10 ka, each of
duration 40–140 years (83) when there were very few sunspots – the Spörer (1400-1540
CE), Maunder (1645-1715 CE) and Dalton (1790-1820 CE) Minima being the most

9
recent and intensively studied using cosmogenic isotopes and sunspot records. During
these intervals, the solar-driven interplanetary magnetic field decreases greatly, allowing
the interstellar cosmic radiation to reach Earth with little attenuation. This effect is
usually quantified by the parameter ɸ, the modulation potential, measured in MV. In the
present epoch, ɸ varies in the range 400 < ɸ <1400MV between sunspot minimum and
maximum. Studies of the radiogenic radionuclides (10Be, 14C and 36Cl) have indicated
that ɸ approaches zero during Grand Solar Minima, and the cosmic ray intensity at Earth
then approximates the interstellar value (83).

During periods of high solar activity (i.e. high ɸ) the GCR intensity reaching Earth is
strongly decreased, but the sunspot size and sunspot magnetic field strengths are much
greater, resulting in powerful solar flares, and very intense SEP events. The overall
radiation environment at Earth surface during the Laschamps is summarized in Fig. S16.
When the Sun is relatively inactive, particularly during Grand Minima, the GCR intensity
is persistently high for periods of 30–140 y (the ɸ=0 curve), during which there is also
the potential for SEP events (84-86). When the Sun is active, the average ɸ approximates
800MV and the GCR intensity is low, but very intense short-lived SEP events occur with
a frequency of about three per decade. The differential spectra for the 23 February 1956
SEP event is given in Fig. S16 and is discussed below.

A substantial fraction of both the GCR and SEP cosmic radiation is deflected from
reaching Earth by the geomagnetic field. The lowest “cut-off” rigidity that can reach
geomagnetic latitude λ (from the west) in a dipole field is given by the Stoermer equation
Eco = 10 (M/Mo). cos λ4 MV where M and Mo are the geomagnetic dipole magnetic
moments for the period of interest and the modern value, respectively. The cut-off
energies for M=0, 0.1M0, and M0 for λ=45⁰ are given in the lower left-hand axis of Fig.
S16. For each value, all the radiation to the right of the symbol is accessible to the
atmosphere. Comparison of the ɸ = 0 MV and 800 MV curves shows that the
combination of the Grand Minimum (ɸ going from 800 to ~0 MV) and the reduction in
the geomagnetic moment to 0.1 M0 during Laschamps results in an increase in the
radiation entering the atmosphere of 1–2 orders of magnitude. Using the well-known
range-energy relationship for protons in air, the scale in the top left-hand upper corner
shows that the greatest increase during Grand Solar Minima occurs in the “ozone layer”
in the lower stratosphere below 32 km.

Consider now the difference between the GCR curve during active solar conditions (ɸ =
800 MV) and the SEP spectrum. For M=1.0M0, the 10 to 1000-fold SEP enhancements
are primarily in the upper troposphere. For M = 0.1M0, three to five orders of magnitude
enhancements occur in the lower stratosphere. While large SEP events only occur on
average at about three per decade, calculations show that these intermittent pulses of
radiation could theoretically result in greater increases in the stratosphere than those
associated with the Grand Solar Minima.

We stress that the substantial increases in cosmic ray intensity, and the resulting large
increases in ionization in the troposphere during the Laschamps occurred primarily at
equatorial and low latitudes; there were minimal increases in the polar and mid-latitude
troposphere. Conversely, there were substantial increases in ionization in the stratosphere
10
at all latitudes. During the period of increased atmospheric ionization, the elevated
production of nitrogen and hydrogen oxides would have resulted in a decrease in the
concentration of ozone in the stratosphere, leading to a substantial increase in the UV
levels on Earth (4). The marked increases in ion density in the troposphere may have
influenced cloud formation (via atmospheric electric charges), cooling, and wind
direction as recently suggested by studies of modern meteorological data (87). Large
scale ionization would potentially impact rainfall, cloudiness, temperatures, and ocean
circulation in equatorial and mid-latitudes (9, 88), whilst also increasing lightning
activity, particularly in the Southern Hemisphere (89).

The cosmic radiation inputs to SOCOL and associated models used to estimate UV
incident on Earth during the Laschamps
Ice-core records of 10Be fluctuations at four polar locations during the Laschamps provide
a well resolved record of the cosmic ray intensity from 45-39 ka (20). Fig. 1 presents
those data for 44.5–39.5 ka showing a number of distinct peaks consistent with Grand
Solar Minima, both in duration and spacing. During the ~1.6 ky period of Laschamps
(42.35–40.75 ka) with M< 0.1 M0, there appear to have been around nine Grand Solar
Minima of total duration ~1000 y, with five particularly intense (Fig. 1). We note that the
two most intense Grand Solar Minima occurred during the latter part of the Laschamps,
and this will be studied at a later time. Assuming the Sun was in an active phase between
all the Grand Solar Minima, contemporary observations suggests that within that
cumulative total 600 yr period there may have been a total of ~180 SEP events similar to
that of 23 February 1956. In addition, there is the potential for ‘super SEP’ events as have
been observed in the cosmogenic record millennium (81, 82). Currently, the lack of
knowledge about the frequency, timing, and extent of SEP events during Laschamps
complicates detailed modelling of the climatic impacts (90), and we leave this issue for
future work.

Auroral Zones: The balance of pressure between the solar wind and Earth’s geomagnetic
field results in the formation of the magnetopause (at around 10–15 Earth radii) and the
geomagnetic tail of the Earth. Processes in both accelerate electrons that precipitate to
earth creating the auroral zones which presently form a circle of 22 degrees in latitude
radius around the geomagnetic pole (which wanders by 10 degrees or more over
millennia). The reduction in dipole geomagnetic field strength to 0-6% during the
Laschamps would have major effects on these parameters. Estimates suggest that the
magnetopause would have shrunken during the Laschamps to a radius of about 5 Earth
radii during quiet periods of solar wind, and potentially to only 3 Earth radii during
geomagnetic storms. As a result, the auroral zones would have shifted 10 to 20 degrees
towards the equator such that they may have reached to mid or even low latitudes. Intense
aurora excite oxygen emissions that are deep red, in the form of rapidly moving curtains,
vertical rays, etc. It is likely that these visual phenomena would have been noticed by
Paleolithic human populations, and potentially even represented in artistic depictions (see
below). These parameters and approximate calculations are summarized in Fig. 5.

11
Defining the timing of the Adams Event using geomagnetic, ice, and kauri records
To define the onset of the Adams Event we used the first appearance of the marked
peaks in 10Be flux recorded in ice records consistent with Grand Solar Minima (see
above), as this is an indication of the reduced geomagnetic field strength allowing
increased amounts of ionizing radiation to penetrate to atmospheric levels.
Approximately nine contiguous large fluctuations in 10Be can be observed between 42.35
and 40.8 ka (Fig. 1), with some peaks also apparent in 14C flux. As a result, we define the
start of the Adams Event at 42.35 ka, while the end is demarcated by the rapid increase in
relative paleomagnetic intensity at 41.56 ka as the polarity reverses. This reversed
polarity phase of Laschamps continued until around 41.05 ka, and was followed by the
final transition phase back to normal polarity which was completed around 40.85 ka. The
estimates of both polarity reversal and recovery have been chosen to be consistent with
the reconstructed geographic position of the Virtual Geomagnetic Pole (VGP) crossing 0
degrees latitude (5, 77).

Global chemistry climate modelling

To examine the potential impact of GCRs on regional and global climate we carried
out a series of different simulations. Here we report the results from three 72-year-long
model runs that we consider provides insights into the impact of a Laschamps-like event
on global atmospheric chemistry and climate. The differences between the three runs are
only the geomagnetic field strength (B) and solar modulation potential configuration. The
geomagnetic field parameters affect the energetic particles penetration depth into the
Earths’ atmosphere depending on the incoming energy spectrum and latitude, while the
solar modulation potential steers the galactic cosmic ray flux into the planetary atmosphere.
Here we use the modelling tool for Solar-Climate-Ozone Links studies (SOCOL v3.0) (91),
consisting of the global circulation model MA-ECHAM5 (MPI-M) (92, 93), a modified
version of the chemistry model MEZON (94-96) and coupled to the ocean model MPI-OM
(97). Versions 1 and 2 of SOCOL, based on MA-ECHAM4, have been validated by
Egorova et al. (98) and Schraner et al. (99). The chemical processes triggered by solar
energetic particles (SEP) and galactic cosmic rays have recently been implemented in
SOCOL v2 (100). The CRAC:CRII (Cosmic Ray Atmospheric Cascade: Application for
Cosmic Ray Induced Ionization) (101, 102) model was used to describe the effect and
parameterize the ionization caused by GCRs in the entire horizontal and vertical model
domain. CRAC:CRII is able to account for the effects of the solar activity and the
geomagnetic field on ionization rates. The solar modulation potential φ is used to describe
the deceleration of precipitating particles due to their interaction with the solar wind. The
ionization rates retrieved from this model are then converted to a NOx and HOx production
rate, since SOCOL does not treat ion-chemistry explicitly. The conversion factor for NOx
is 1.25 nitrogen atoms per ion pair (103) of which 45% yield N(4S) and 55% yield N(2D).
The N(2D) instantaneously converts into NO via N(2D) + O2 → NO + O. For HOx
production we used the parametrization by Solomon et al. (104). The latter study examined
the thermodynamics of ion and neutral chemistry during charged particle precipitation to
describe odd hydrogen production depending on altitude and ionization rate and found
values between 1.9 and 2 odd hydrogen particles produced per ion pair below 60 km
altitude. The model does not include any potential effect of GCRs on cloud properties via
aerosol formation (105). We use SOCOL-MPIOM (106) in T31 horizontal resolution i.e.

12
with an approximate grid spacing of 3.75˚ × 3.75˚ and 39 vertical hybrid sigma-p levels,
terrain-following sigma levels at the bottom and isobaric coordinates aloft. The vertical co-
ordinates span the atmosphere from surface to 1 Pa (∼80 km). The chemical module treats
41 chemical species, which interact via 140 gas-phase, 46 photolysis and 16 heterogeneous
reactions. MPI-OM is based on primitive equations with hydrostatic and Boussinesq
approximations. Horizontal discretization is on an orthogonal curvilinear C-grid, in the
vertical isopycnal coordinates are used. MPI-OM includes a dynamic and thermodynamic
sea ice model.

To examine the impacts of different geomagnetic and solar conditions we performed three
72-year-long model runs. All runs were set up with glacial conditions. Before the three
runs were started, a 398-year-long spin-up run was made under reference conditions, to
ensure that the ocean-atmosphere system was in a stable state, with no large drifts in climate
parameters. The Earths’ orbital parameters were fixed to historic values (Table S6). The
initial land surface dataset (e.g. ice cover) and ocean temperatures are from an earlier
glacial simulation with ECHAM5 and were kindly provided by Uwe Mikolajewicz from
MPI Hamburg (Fig. S17); this includes the presence of the Laurentide Ice Sheet. The
atmosphere was run in a ‘pristine’ state and concentrations of greenhouse gases (GHGs)
and ozone depleting substances (ODSs) are low (Table S7), resulting in a model globally
averaged annual mean 2m (surface) temperature of ∼284.68K. Trace gas concentrations
and solar spectral irradiance were held fixed during the entire runs. The reference run
(REF) has a modern geomagnetic dipole moment, a position of the geomagnetic North Pole
at 78.5°N and 291.5˚E, a solar modulation potential φ of 800 MV, and ionization rates from
GCR and auroral electrons. To emulate a weakened geomagnetic field during the
Laschamps, we undertook a second run (P800M0) where we applied zero geomagnetic
field strength keeping the other parameters as in REF setup. For the third run we applied
solar modulation potential φ of 0 MV and exclude auroral electron precipitation effects to
explore the impacts arising from a weakened geomagnetic field (zero) and Grand Solar
Minima (P0M0). To eliminate the transition from the reference spin-up to the experimental
conditions, the first 12 years of the simulations were not used. The 60-year simulations
were compared applying a two-tailed student’s t-test. Summary plots of seasonal
differences in chemistry and climate parameters compared to REF are shown in Figs. S18-
S30; these include the relative erythemal dose change which was calculated by applying a
RAF factor of -1.1 to relative total ozone changes. It is important to note that low solar
activity UV decreases relative to other wavelengths of the solar spectrum (during Grand
Solar Minima), therefore generating an even stronger response in ozone depletion
compared to the decrease from additional GCR ionization (107). As a consequence, total
ozone may have decreased by a further 10-20%, resulting in an additional 5-10% surface
UV increase, potentially amplifying the climate impacts described below.

We focus on the impact of a weakened geomagnetic field and Grand Solar Minima (P0M0)
during the boreal winter (December-February) which exhibits the greatest impact of all the
seasons in our simulations (Figs. S18-S30), and is similar to some modern empirical
observations (87). We find that NOx increases over most of the atmosphere due to increased
ionization levels from GCRs, which is due to a higher flux of energetic particles into the
atmosphere (no geomagnetic field disturbance) and a higher particle energies allowing

13
penetration through the atmosphere down to low altitudes (no solar modulation potential
breaking down the energetic particles). The negative anomaly in the Northern Hemispheric
mesosphere is likely due to the decrease in solar low energetic electrons. As for NOx, the
generation of HOx increases over the entire atmosphere, down to low altitudes. However,
due to the increased amount of NOx, in the austral summer, HOx is destroyed by reaction
with NOx to HNO3, which is removed through for instance, precipitation, resulting in a net
negative anomaly for HOx in the austral Southern Hemisphere. Importantly, both NOx and
HOx have been demonstrated to cause the catalytic destruction of stratospheric ozone
(while also producing tropospheric ozone) (Fig. 2C) (8).

The temperature patterns (Fig. S21) are, for a large part, linked to the ozone changes. The
negative anomalies in stratospheric ozone lead to a cooling effect, except where no light
penetrates the atmosphere (polar night); here, the positive anomaly in ozone over the
Northern Hemisphere pole is associated with a negative temperature anomaly. In
contrast, the increase in ozone in the troposphere is accompanied by positive temperature
anomalies, reducing the loss of infrared thermal radiation leave the lower troposphere to
higher altitudes. Intriguingly, the remarkable bi-pole pattern of temperature anomalies
over the Northern Hemisphere polar night (positive anomalies centred on 10-40km versus
negative anomalies at 50-80 km) cannot be explained by ozone anomalies only. We
assume that the acceleration of the Brewer-Dobson circulation leads to an increased
transport of ozone-rich air from the equatorial region to the northern pole (positive ozone
anomaly). As well, the accelerated Brewer-Dobson circulation increases downwelling of
air over the northern pole, from altitudes of 70 km down to altitudes of 20 km (108). The
air, heating up adiabatically, therefore generates a significant positive temperature
anomaly at lower stratospheric altitudes. The zonal wind (Fig. S27) reacts accordingly.
At 20 km altitude, a warm air anomaly over the polar region disrupts the polar vortex
through a decrease in the equator-to-pole temperature gradient. This process is self-
sustaining, if not self-enhancing: a disrupted polar vortex will mix with warmer
extratropical air masses, warming the region further, which in turn further disrupts the
vortex. This disturbance then propagates upward to high altitudes (65 km), as seen in the
zonal mean wind speed. One possible cause for the acceleration of the Brewer-Dobson
circulation is the increased surface (10 m altitude) wind speed in front of the most
important orographic barriers (for instance, the Norwegian Alps, the central Rockies,
Kamchatka, and Japan) (Figs. 3A and S28) (109). Here, increases in the wind stream
above the orography will generate additional gravity wave drag, potentially vertically
propagating into the stratosphere and driving the ‘extratropical pump’. Our modelled sea-
level pressure increase over the Arctic and North America and decreases over western
Europe corresponds to a negative phase of the Arctic Oscillation (AO) and/or North
Atlantic Oscillation-like pattern (NAO) (110, 111), consistent with contemporary studies
that have linked persistent solar minimum-induced long-term changes in these modes
(112). Importantly, the mechanisms described above for the Laschamps are several orders
of magnitude greater than current solar-induced phenomena, and as result cannot explain
historic global temperature trends (113-117).

It is important to note that in many ways our findings appear to confirm aspects of much
earlier work that suggested geomagnetic excursions may be connected to major climatic

14
shifts and biological impacts (4, 118-120), potentially including extinctions and human
evolution (4, 118-120).

Potential impacts of VGP route

The inferred geographic route taken by the projected Virtual Geomagnetic North
Pole (VGP) during the Laschamps has previously been reconstructed (5, 77). In
summary, the projected VGP initially moved eastwards across northern North America
(over the Laurentide Ice Sheet; LIS) as far as the northern Atlantic and Greenland
(around 42.25 ka) before reversing track across central North America (and the LIS once
more) back to the northern Pacific Ocean (around 41.80 ka). The VGP then moved
rapidly southwards over the Pacific as far as Antarctica (around 41.51 ka) before crossing
the southern continent and exiting south of Western Australia (around 40.35 ka). The
return route of the VGP northwards through the Indian Ocean was only slightly slower
than the Pacific leg, and reached mid-Eurasia by 40.83 ka. The northwards movement
then continued to the current Arctic position of the VGP. It is notable that this route took
the VGP across or close to many sites in our Pacific transect recording climatic or biotic
impacts during the Laschamps. These sites include Auckland Island, New Zealand, the
Australian mainland, New Caledonia and to a lesser extent Sulawesi. One interesting
possibility is that the reduced geomagnetic field strength expressed directly above the
VGP may have allowed increased amounts of ionizing radiation to penetrate to the local
atmosphere, with ensuing enhanced climatic and biological impacts. In this regard it is
potentially important that the change in ice sheet growth rate of the LIS (from retreat to
rapid expansion) occurred very close in time to when the VGP passed by geographically,
around 42.3-42.2 ka in eastward and then westward transects (18, 19, 121). This is also
very close in time to the first Grand Solar Minima apparent during Laschamps (~42.3 ka)
from the 10Be records, raising the potential that one or both of the two events may have
amplified the ice sheet response. The timing of both events closely corresponds to the
shift in ITCZ (Lake Towuti), Southern Hemisphere westerlies (Auckland Island), and
increased aridity suggested across Australia (Lake Mungo and Lynch’s Crater). To date,
the potential local impacts of the VGP position has not been included in the climate
modelling runs.

Summary of environmental and archaeological changes during the Adams Event

Note: All radiocarbon ages reported in this study have been calibrated using the new
combined kauri-Hulu Cave calibration dataset.

Potential signals in Greenland ice records: In the Northern Hemisphere around 42 ka, the
Adams Event and broader Laschamps appears to be associated with a rapid re-expansion
phase of the Laurentide Ice Sheet from a glacial minima that included an ice-free Hudson
Bay (17-19, 121). Analyses of Greenland ice records have revealed atypical signals in the
stadial (GS-10) and interstadial (GI-9) events that occur during and following the
Laschamps. The relatively insubstantial GS-10 signal is absent or only weakly recorded
in many other climate records, including marine signals such as Cariaco Basin, Arabian
Sea, and Iberian Margin, and has an unusually low signal of atmospheric dust in
Greenland ice records (29, 30). The ensuing GI-9 has an unusually gradual onset while
the WAIS Antarctic ice record shows minimal evidence of a corresponding Antarctic

15
isotope maximum event at this time, which would be anticipated from a bipolar seesaw
response (41). This has led to suggestions that GI-10 and GI-9 in fact represent a single
longer interstadial event that has been interrupted by a pronounced cold phase that was
not a true Greenland (Dansgaard-Oeschger) stadial, but which altered the temperature
gradient between the mid-latitudes and Greenland (29, 30). This would be consistent with
a negative phase of a AO/NAO-like pattern during the Adams Event causing expansion
of sea ice around Greenland, and producing a local signal recorded as GS-10.
Importantly, the enigmatic GS-10 event occurred at exactly the same time as the weak
geomagnetic field strength of the final transition phase of Laschamps (Fig. 1). A further
observation is that the Adams Event itself is closely coincident with the prolonged cold
period of GS-11, while the Laschamps reversed polarity phase overlaps markedly with
interstadial GI-10. Together, the coincident timing of these events raises the prospect that
marked cooling signals corresponding to periods of low geomagnetic field strength
during the Laschamps may be preserved in Greenland ice records, but that they have not
been previously identified due to the close resemblance to the pattern of surrounding
stadial-interstadial events.

Subantarctic environmental changes: The pollen and spore results from the lower organic
lignite in Pillar Rock indicate a shrub-grassland, dominated by Dracophyllum longifolium
and tussock grasses (Fig. S6). A previously unreported and unidentified pollen type –
which we are here calling ‘Pseudopanax moarii’ – almost certainly represents an extinct
and close relative of the Pseudopanax group of large leaved small trees on the New
Zealand mainland. Megaherbs were present and for most of the sequence the local ground
surface was a Plantago sward. Interpretation is complicated by the fact that the site is
currently on an exposed cliff subjected to strong winds and thus in tussock grassland,
while shrub and forest cover is some 200 m distant downslope. The extant vegetation
cover is therefore strongly influenced by wind. Growth of a Dracophyllum shrubland
close to the cliff edge, and the absence of stones in the sediments implies reduced gale
force winds. The annual temperatures at the time cannot have been much lower than
those of the present.

The lower half of the lignite sequence, centered on 52 ka has consistent representation of
pollen from the tall New Zealand endemic trees Nothofagus and Phyllocladus with
occasional representation of other coniferous trees (Podocarpus, Prumnopitys taxifolia
and Halocarpus). These trees are abundant in mainland lowland forests on New Zealand
and their long-distance transport to Auckland Island indicates substantial areas of
lowland forest in southern regions of the South Island of New Zealand at this time. This
period coincides with the highest levels of woody plants and lowest of grasses at the site
and thus indicates peak warmth. Mean annual temperatures at this time were likely to
have been within 1.5 °C of the present. From 50 ka onwards, the long-distance pollen
assemblage consists of Nothofagaceae, Phyllocladus, Halocarpus and Australian sourced
types such as Casuarina and Eucalyptus. This period coincides with the highest levels
of woody plants and lowest of grasses at the Pillar Rock site and thus indicates peak
warmth. From 50 ka onwards, the long-distance pollen assemblage also includes
Australian sourced types such as Casuarina and Eucalyptus. Locally, grasses increase
and woody plants decline indicating increasingly cooler climates. Mean annual

16
temperatures at the time of deposition of the lower organic lignite were probably within
1.0–1.5 °C of the present.

The subsequent influx of silt, stones and boulders indicates severe periglacial activity,
followed from 11.5 ka by stabilization of the ground surface. The upper peat sequence,
covering the transition from the full glacial to the beginning of the Holocene, is initially
dominated by megaherbs and relatively low levels of grass with almost no woody plants.
This is typical of much higher elevation sites above 250 m and with estimated annual
temperatures at least 1.5 °C lower than at current sea level. Lack of stones and silt in the
peat suggest relatively weaker westerly winds. Shortly after the increase of Myrsine
shrubland at 8 ka, the influx of stones and silt in the section recommenced, indicating a
resumption of intense westerly gale-force winds (12, 122). The summary pollen record
from the early Holocene upper organic peat unit is shown in Fig. S8.

Equatorial West Pacific changes: The sediment deposits in Lake Towuti, Sulawesi,
demonstrate a marked shift in 13C (leaf wax) to more negative values (depleted in 13C)
in close association with a minimum in geomagnetic intensity (60). Radiocarbon dates
through this period align well with the magnetic changes, and indicate a transition at 42
ka (Fig. S11). These signals indicate increasing environmental aridity, which has been
suggested to be a consequence of migration of the ITCZ south (11) and the general
conditions remained in place until the Holocene (comparable to the onset and duration of
inferred periglacial conditions on Auckland Island). Similarly, on New Caledonia (which
was uninhabited in the Late Pleistocene) a change in vegetation from Araucaria to more
arid conditions around this time (123) parallels the patterns seen in Australia and
Sulawesi.

Megafaunal extinctions: High-resolution Pleistocene records of biodiversity change

appear quite episodic. A key example is megafaunal extinctions, which appear to be
staggered throughout the late Pleistocene (38) in many areas. Interestingly, within this
generally distributed pattern, a cluster of megafaunal genetic transitions (woolly rhino,
mammoth, bison) were previously observed around the timing of the Mono Lake
geomagnetic excursion (38). Importantly, a global trend is that megafaunal extinctions
seem to occur long after the arrival of modern humans in given areas such as Eurasia, the
Americas (124), and even large islands such as Madagascar (125).

The global pattern of extended human overlap with megafauna, and lack of intensive
signs of megafaunal hunting, makes it increasingly hard to explain humans as the primary
cause of extinctions. Australia is an extreme example with an overlap of Anatomically
Modern Humans (AMH) and megafauna of at least ~10 ka, if not more (32, 126). Despite
the ITCZ migrating south, exposure of the Sunda and Sahul shelves with falling global
sea level may explain reduced precipitation over northern Australia from 42 ka (11). In
the south, recent work suggests the equatorward-migration of westerly winds over the
Pacific subantarctic islands was accompanied by expanded sea ice (12), decreasing
delivery of precipitation over southern Australia. The discovery that the Laschamps is
contemporaneous with a sustained climatic shift to arid conditions and megafaunal
extinctions in Australia raises interesting questions about potential causes. Many studies

17
have pointed out that global megafaunal populations have survived numerous previous
glacial cycles, suggesting that something must have been different in the last cycle, with
the appearance of AMH being the most likely key factor. In the Australian situation, a
model where the Adams Event led to major environmental shifts including increased
aridity may have allowed a critical resource e.g. water supplies, to be occupied or
dominated by humans, resulting in outsized impacts on megafaunal populations without a
marked increase in human population size or activity. Recent modelling work has favored
such a model to explain the Australian megafaunal extinctions at 42 ka (14, 127).
Similarly, fires from increased lightning strikes, potentially increased by atmospheric
ionization during the Adams Event (89), and other consequences of the climatic changes
may have played significant roles and could explain increased charcoal levels in sites
such as Lynch’s Crater from 42 ka (15) and the lack of relationship with archaeological
signs of human activity (128). It is worth noting that many large Australian endemic
terrestrial taxa that survived this megafaunal extinction phase show relatively limited
genetic diversity, such as the thylacine, whose mitochondrial diversity indicates a
bottleneck around 42 ka (129). This suggests that even surviving large animal
populations may preserve genomic signals of population bottlenecks due to the impact of
the Adams Event. It is also possible that species which rely on biomagnetic direction
sensing may have been impacted by the Adams Event. For short generation time species
such as insects and birds, the rate of geomagnetic field change may have been sufficiently
slow to allow adaptation, however for long-lived species such as cetaceans this may not
have been possible (130), and confusion over their geo-positioning may have occurred.

Neandertal extinction: The extinction of European Neandertal populations appears to

have occurred in a temporally and geographically staggered pattern moving across
Europe from east to west (31, 37, 131, 132). This pattern has been modelled as resulting
from the westward spread of steppe landscapes and permafrost zones during the intense
cold spells of Greenland Stadials 12 and 10; 44.3–40.8 ka) in combination with the
arrival and rapid spread of AMH from eastern Europe around 46-47 ka (31, 131).
Stalactite records suggest a steep regional climate gradient, with a prolonged period of
what has been interpreted as extreme low summer temperatures in the East Carpathians
that persisted from GS-10 through into GS-9 and subsequent Heinrich Event (HE) 4,
despite the interstadial GI-9 recorded in Greenland ice records (31). This period of low
summer temperatures commences with the latter phase of Laschamps, and was
accompanied by shifts in conditions across Europe to increased aridity, altered biomes
and changed seasonality in areas such as the Black Sea. Importantly, these environmental
changes were accompanied by widespread depopulation and archaeologically sterile
layers in western Europe lasting several centuries, which are commonly observed to
separate the earlier Mousterian and Chatelperronian (Neandertal) from the later
Aurignacian (AMH) layers. This disjunct stratigraphy has been used to suggest the
immediate cause of Neandertal extinction may not have been direct competition, but
better or more rapid environmental adaptation by invading AMH into environments
vacated by Neandertal populations during these periods of environmental change (31).
Importantly, this prolonged period of environmental change overlaps with Laschamps.
Together with the preceding extreme cold stadial GS-12, this suggests the impacts of the
Adams Event and broader Laschamps on climate and environments may have been

18
sufficient to negatively affect the remaining Neandertal populations, perhaps due to their
dietary specialization of terrestrial animals, allowing invading AMH populations to
increasingly occupy their former range. While it has been suggested that a higher UVB
sensitivity of Neandertals might have led to their extinction during the Laschamps (4), the
associated climatic and environmental changes seems a far more plausible explanation
given the wide range of melanian pigmentation alleles observed in Neandertal individuals
(133).

To test whether the extinction of Neandertals was contemporaneous with either the
Adams Event, or Laschamps in general, we re-calibrated a recent compilation of high-
quality radiocarbon dated specimens that represent last occurrences at different
Neandertal sites, using the kauri-Hulu calibration curve reported here (Table S8). Using
the OxCal Phase function (72, 73), we calibrated published radiocarbon age estimates
from bone collagen (ultrafiltered) (37, 132) to refine the estimated age of Neandertal
extinction to 40.9–40.5 ka at 1 (41.0–40.3 ka at 2). This is synchronous with the final
transition phase to normal magnetic polarity 41–40.8 ka and the enigmatic GS-10 event
(Fig. S31). Comparisons of early AMH cultures in Europe that overlap with Neandertals
have shown that the Uluzzian in Italy also appears to end contemporaneously at this point
(31), indicating the Laschamps may have also negatively impacted AMH populations in
Europe at the time. In this regard it is potentially significant that the Laschamps appears
to mark the period when regional variants of the Initial Upper Paleolithic cultures (such
as the Uluzzian) coalesced into the Aurignacian technocomplex, which seems to appear
roughly simultaneously across western Eurasia (134). For example, the Aurignacian
culture appears to arrive as far west as Portugal, as shown by the Lapa do Picareiro site,
around or shortly after 42-41 ka. At this site the Aurignacian population appears shortly
after the local extinction of Neandertals, and around the same time or slightly before the
severe cold and dry conditions of GS-9/Heinrich Event 4. A similar replacement pattern
is seen at Bajondillo Cave in southern Spain, albeit potentially slightly earlier (135). In
both cases, the Laschamps appears to mark a distinct boundary in the sequence of AMH
cultural layers.

The warmer period leading up to the Laschamps in the Northern Hemisphere locations
might also explain why the archaeological record appears to indicate that multiple Arctic
sites were occupied by AMH populations relatively rapidly after the 50–55 ka
colonization of Eurasia. The cold conditions following Laschamps, including Heinrich
Event 4, leading into the LGM, may have limited the ability of AMH to survive at such
northern sites e.g. the Russian early paleolithic site of Mamontovaya Kurya (136).

Human movement and behavior in Sahul: The human burials at Mungo Lake, Australia,
which have been dated by Optically Stimulated Luminescence (OSL) analysis of the sand
layers in which they were buried, are within or close in age to the Adams Event (13). The
Mungo I skeleton is bracketed by OSL ages of 42.7 and 42.5 ka, while Mungo III is
similarly constrained at 42.2 and 41.9 ka, with errors on all dates around 2.5 k (13). Stone
tools are recorded in older layers, with a maximum age at the site of 50 ka, similar to
many other early sites around Australia (126). The human burials are close to the time
when climatic patterns appear to have changed in the Mungo stratigraphy, marked by

19
gravel layer III, dated around 42 ka. Given the much earlier occupation by humans of this
site, it is intriguing as to why the burials themselves seem to be limited to the time of the
Adams Event. Around the same time, Timor Leste was first occupied, with evidence of
cave habitation and red ochre use around 42 ka (35, 137). The role of decreasing global
sea level (in response to the rapid growth of the Laurentide Ice Sheet) (18, 19) in
allowing AMH exploration of Timor Leste and the wider region remains unclear.

Southern African records: A key vertebrate fossil archive from the southern Cape of
South Africa shows that the magnitude of large mammal faunal change across the
Laschamps exceeded that which occurred across the Pleistocene-Holocene transition, and
corresponded with expansion of C4 grassland and shifting westerly airflow, suggesting
hemispheric-wide impacts (42). The southern Cape of South Africa, which includes the
region encompassing the southern coastal plains and the east-west trending ranges of the
Cape Fold Belt, is uniquely positioned to inform on the history of temperate and tropical
circulation systems in the region because it receives winter rainfall from frontal systems
associated with the westerly storm track as well as summer rainfall related to tropical
easterly flow (138). Multiple lines of evidence from the region’s fossil and climatic
archives highlight significant changes that correspond in age with the Adams Event.

The archaeological deposits at Boomplaas Cave (139) provide the only southern Cape
mammalian fossil record that encompasses the Adams Event (Fig. S32) (42). At this site,
the faunas from stratum BP, the earliest 14C ages for which are ~40.3 ka (140), are
distinct from those found in overlying and underlying stratigraphic intervals (42). This is
shown here in Fig. S33, which plots chord distance across adjacent pairs of stratigraphic
units (see 141), and indicates that the magnitude of faunal change across the transition to
and from stratum BP exceeds that which occurred across the Pleistocene-Holocene
transition. Stable carbon isotope (δ13C) evidence indicates elevated consumption of C4
grasses by BP grazers (142), matching a positive excursion in δ13C in the nearby (~4 km
east) Cango stalagmite (143). The onset of this excursion in the Cango record is ~41 ka
(Fig. S33) though it may be closer to ~42.5 ka if the radiocarbon reservoir correction
applied by ref. (143) is too large (144). The implication of these archives is an expansion
of C4 grassland immediately following the Adams Event, likely reflecting greater
influence of summer rainfall and relatively less influence of winter precipitation systems
delivered by Southern Hemisphere westerly airflow (142). Similar transitions are
apparent in Border Cave in KwaZulu-Natal, where several discrete stratigraphic horizons
immediately before and during the Laschamps record the abrupt appearance of the
earliest archaeological evidence of modern hunter-gatherer adaptations (the San material
culture) including digging sticks, eggshell beads, and bone points (145).

Cave Art: Early cave art in ISEA overlaps with the Adams Event (from 42 ka; Fig. S23),
and AMH cave art appears in Europe almost contemporaneously (36, 146). This
coincidence has been noted, and regarded as currently unexplained (36, 146). Similarly,
sites in Timor Leste, and Southeast Asia more generally, show an onset of cave habitation
and red ochre use around 42 ka (35), consistent with increased use of caves for shelter. It
is possible that this is related to the elevated UV levels and other impacts of the increased
atmospheric ionization that would have been associated with Grand Minima, and

20
particularly SEP events. The relatively sophisticated nature of the early cave art when it
appears suggests that figurative painting was already well-established prior to it being
recorded on cave walls, presumably on other substrates such as rock shelter or cliff walls
or other materials (e.g. wood, bark shields), but that these have not survived subsequent
(long term) environmental exposure. The globally common appearance of negative and
positive ochre hand stencils in Laschamps-aged cave art is also interesting given the
utility of red ochre-based skin coverings as protection against ultraviolet radiation, which
is still practiced in places such as Africa (147), given the likely UV impacts of Grand
Minima and SEP events. In terms of the potential use of ochre skin coverings as a
sunscreen it is also possible that in temperate areas such as Europe where animal skins
are likely to have been routinely used, hands and arms would have been an important area
for ochre sunscreen application. If applied by blowing (e.g. through a straw, or from the
mouth), then a negative stencil would have been a routine consequence and may have
been adopted as an artistic feature. Either way, the contemporaneous appearance of hand
stencils and complex art in different parts of the world would appear to represent the first
detection of the activity, rather than the origin of the behavior itself as has sometimes
been suggested, particularly with reference to the origins of cognitive thought and
language. In a similar fashion, the apparent origins of other behavior suddenly recorded
in cave deposits at 42 ka, such as deep ocean fishing in Timor Leste (35) or modern
hunter gatherers in South Africa (145), might represent the same preservation bias due to
the more intense utilization of cave environments during the Adams Event. Again, this
would suggest that the behaviors themselves could considerably precede 42 ka.

The earliest AMH artistic depictions involve red ochre hand stencils, symbols, and
figurative animal art in both Europe and ISEA. In El Castillo cave, Spain, a large stippled
red disk near a series of such images (Panel de las Manas) has been dated by overlying
stalactites at >41.4 ka (34), and represents the earliest figurative art currently known in
Europe. Given the temporal proximity of this date to Laschamps, this raises the question
of whether the curving series of large red disks observed in El Castillo could potentially
represent the movement of the Sun (or Moon), whose appearance and damaging UVB
outputs are likely to have changed significantly and rapidly during this period of time.
Intriguingly, a second large red disk image from another gallery of such images in El
Castillo (Galeria de los Discos) was also dated in the same study, and was tightly
bracketed between 34.2 and 35.7 ka (34) (Fig. 3), which overlaps the Mono Lake
geomagnetic excursion 34.5 ka). This would appear to add further support to the idea that
these images are related to phenomena during geomagnetic excursions, and that this
behavior shows some continuity over a period of at least 7 ka. In addition to potential
images of the Sun or Moon, such phenomena may relate to the reduced magnetosphere
diameter during such excursion events, which would have allowed aurora to spread to
low latitudes, including southern Europe. During Grand Solar Minima or Solar Energetic
Particle events, these atmospheric conditions are likely to have been far more extreme
(Fig. 5). Intense auroral conditions excite oxygen which produces deep red emissions,
and it is highly likely that intense and unusual auroral visual phenomena would have
made a considerable impression on human populations who observed them, potentially
then being reflected in artistic depictions. It will be important to perform largescale
dating studies on cave art to investigate whether visual styles or images might be

21
potentially related to geomagnetic excursions. This could potentially include analysis of
the magnetic signals preserved in the ochre paintings themselves.

22
Fig. S1
Location of sites investigated in this study and key records discussed in the text. Map
generated from OpenStreetMap contributors.

23
Fig. S2
The Ngāwhā Springs ancient kauri log during extraction. Credit: Nelson Parker/ Nelson’s
Kaihu Kauri.

24
Fig. S3
The cross-section of the Ngāwhā kauri log. A) The full ‘biscuit’ of the tree. B) Close-up
of the radii used for measurements and sampling for radiocarbon. Note, the innermost
centre of the tree was missing (rotten).

25
Fig. S4
Ngāwhā tree-ring measurements and comparative kauri growth rates. A) Nga001 ring-
width measurements (mm/year). B) Box-and-whisker plot of modern kauri growth
measurements obtained from 240 trees (17 sites) living over the last millennium (43),
a late Pleistocene collection of 169 trees (15 sites; Palmer unpubl. data) and the Ngāwhā
tree (Nga001). Note, for the modern kauri and late Pleistocene series, the plot shows the
average growth rates between all the trees, whereas Nga001 shows the distribution of
the individual tree measurements.

26
Fig. S5
Location of Pillar Rock, subantarctic Auckland Islands (New Zealand). Map generated
from OpenStreetMap contributors.

27
Fig. S6
Location, stratigraphy, vegetation and chronology for the Pillar Rock sedimentary
sequence (subantarctic Auckland Islands) during the last glacial period. (a) Location of
the Pillar Rock site and key stratigraphic units (12), including the lower organic lignite
and upper peat (b) Pollen values from the lignite expressed as percentage total pollen.
The Holocene pollen record from the same sequence is shown in Fig. S8. The previously
unidentified taxa ‘Pseudopanax moarii’ and Liliacitides spp. (that went extinct prior to
the Holocene) are shown as a dashed tree line and red filled (unknown) line respectively.
The chronology for the sequence calibrated against the new kauri-Hulu Cave curve
(plotted against depth) is given alongside the pollen. The timing of the weakest phase of
the magnetic field during the transition into the Laschamps is given as a dark grey
column. Wk = Waikato Radiocarbon Dating Laboratory identifier.

28
Fig. S7
Cleaned section of the lower organic unit, Pillar Rock, subantarctic Auckland Islands
(New Zealand).

29
Fig. S8
Summary Holocene pollen record from Pillar Rock, subantarctic Auckland Islands (New
Zealand). Percentages expressed as total pollen. Ages calibrated using SHCal20 (68).

30
Fig. S9
Location of Lake Towuti, Sulawesi, Indonesia. Map generated from OpenStreetMap
contributors.

31
Fig. S10
Bayesian age model for Lake Towuti, Sulawesi, Indonesia. The model incorporates the
radiocarbon likelihoods (calibrated probability distributions; brackets under the
distributions and blue envelopes represent the 68.2% and 95.4% ranges, respectively).
Dashed line denotes the magnetic field minimum (60).

32
Fig. S11
Leaf wax 13C compared to sediment relative magnetic paleointensity (RPI) against
depth. Sampling depths of radiocarbon ages shown in upper panel; open circles are
radiocarbon ages that have been previously reported (11); solid circles denote new
contiguous radiocarbon ages generated during this study (upper side panel shows
magnified detail of radiocarbon sampling). Note, the coincidence of the minimum in the
magnetic field strength (60) with the long-term inflection of the leaf wax 13C (11) (grey
column).

33
 Posterior mean
95% CI
800

600

400
D C (‰)
14

200

Speleothem: MSD
Speleothem: MSL
Kauri: FIN14058
−200 Kauri: FIN09
Kauri: NZKL1855b
Kauri: Ngawha

30 35 40 45 50

Calendar Age (cal kBP)

Fig. S12
Posterior mean of the kauri-Hulu Southern Hemisphere radiocarbon calibration curve
from 30-50 ka in Δ14C space together with its 95% credible interval. All Δ14C
observations are shown with their 95% probability intervals. The four floating kauri trees
are located at their posterior mean calendar ages as estimated during curve creation. The
data from the two Hulu speleothems (MSD and MSL) are plotted at their prior calendar
age estimates based upon U/Th dating (shown with 95% probability intervals) and
adjusted in Δ14C by the posterior mean offset to the Southern Hemisphere kauri estimated
in curve construction. Repeat 14C measurements within MSD and MSL have been
combined (using a weighted F14C mean) before plotting.

34
 Kauri−Hulu Laschamps Calibration Curve

Speleothem: MSD
Speleothem: MSL
Kauri: FIN14058
Kauri: FIN09
45 Kauri: NZKL1855b
Kauri: Ngawha
Radiocarbon Age (14C yrs )

40

35

30

Posterior mean
25 95% CI

30 35 40 45 50

Calendar Age (cal kBP)

Fig. S13
Posterior mean of the kauri-Hulu Southern Hemisphere radiocarbon calibration curve
from 30-50 ka in radiocarbon age space together with its 95% credible interval. The data
are shown with their 1σ uncertainties in both radiocarbon and calendar age. Shown as a
rug along the bottom are the knot locations for our spline, with the higher density of
evenly-spaced knots covering the potential calendar ages of the four kauri floating trees.
As in Figure S12, the kauri trees are located at their posterior mean calendar ages (the
uncertainty in these posterior calendar ages is not indicated on the plot); the Hulu
speleothems their priors based upon U/Th and adjusted in radiocarbon age according to
the posterior mean of the offset to the SH data estimated during curve
creation. Repeat 14C measurements within MSD and MSL have been combined (using a
weighted F14C mean) before plotting.

35
Fig. S14
Histograms of the posterior calendar ages of the oldest 14C determination for each of the
four kauri trees estimated during the construction of the radiocarbon calibration curve.
(A) and (B) Kai Iwi Lakes trees (FIN09 and FIN14058), (C) Mangawhai (NZKL1855)
and (D) Ngāwhā Springs (Nga001).

36
Fig. S15
Comparison between kauri 14C and ice core 10Be records. A) Modelled 14C production
rates assuming a reduced ocean ventilation of 70% of its preindustrial value (orange) and
the 10Be-ice core stack (blue, 21) scaled by a factor of 1.3 and shifted by +265 years from
its original GICC05-age) (148). The shown uncertainty envelopes are derived from
modelling 500 realizations of the Δ14C record, and from the 10Be-measurement
uncertainty and stacking of the different ice cores (see ref. (21) for more details). B)
Histogram (bars) and kernel density (line) of the obtained best shifts for each iteration.
The dot with error bars shows the most likely shift and its 68.2 and 95.4% probability
intervals. C) Histogram of the obtained 10Be-scaling factors for the best fit of each
iteration. D) Histogram of the obtained best scaling of the ocean diffusivity parameter in
the model, relative to its preindustrial value.

37
 10
55 47 32 18
Height- km
TRO PO PAUSE

7
23 Feb 1956
Log Fluence (particles/cm2.MeV)

ɸ = 0 MeV

ɸ = 400 MeV

ɸ = 800 MeV
1

Stoermer cut-off at 45⁰ geomagnetic

M=0% M=10% M=present
0
1 1.5 2 2.5 3 3.5 4 4.5

Log Energy (MeV)

Fig. S16
Summary of the radiation environment during the Laschamps. To allow easy comparison
of the relative importance of the SEP events and the Grand Solar Minima, this figure
plots the fluences (logarithmic scale) of the largest SEP event (total duration), and the 20-
day fluence of the GCR for modulation potentials ɸ= 0, 400, and 800 MeV. The x axis is
the logarithm of the cosmic ray energy in MeV. The symbols M=0; M= 10% M0 and M0
(modern day) indicate the lowest cosmic ray energy accessible to 45⁰ latitude for those
values of the geomagnetic dipole moment. The altitudes given along the top indicate the
depth at which a cosmic ray particle of that energy comes to the end of its range.

38
Fig. S17
Global ice coverage (blue) as used for initiation of the global chemistry-climate model
SOCOL-MPIOM runs.

39
 Geomagnetic Minimum Geomagnetic Minimum +
Grand Solar Minima

September,
October
and November

December,
January
and February

March,
April
and May

June,
July
and August

Fig. S18
Simulated seasonal changes in nitrogen oxide (NOx; %) anomalies for different
geomagnetic and solar scenarios. Anomalies are shown for a weakened geomagnetic field
during the Laschamps (M0P800) and a weakened geomagnetic field and Grand Solar
Minima (M0P0). Anomalies calculated as a 60-year average of 60 seasons. Colored areas
denote 10% significance.

40
 Geomagnetic Minimum Geomagnetic Minimum +
Grand Solar Minima

September,
October
and November

December,
January
and February

March,
April
and May

June,
July
and August

Fig. S19
Simulated seasonal changes in hydrogen oxide (HOx; %) anomalies for different
geomagnetic and solar scenarios. Anomalies are shown for a weakened geomagnetic field
during the Laschamps (M0P800) and a weakened geomagnetic field and Grand Solar
Minima (M0P0). Anomalies calculated as a 60-year average of 60 seasons. Colored areas
denote 10% significance.

41
 Geomagnetic Minimum Geomagnetic Minimum +
Grand Solar Minima

September,
October
and November

December,
January
and February

March,
April
and May

June,
July
and August

Fig. S20
Simulated seasonal changes in ozone (O3; %) anomalies for different geomagnetic and
solar scenarios. Anomalies are shown for a weakened geomagnetic field during the
Laschamps (M0P800) and a weakened geomagnetic field and Grand Solar Minima
(M0P0). Anomalies calculated as a 60-year average of 60 seasons. Colored areas denote
10% significance.

42
 Geomagnetic Minimum Geomagnetic Minimum +
Grand Solar Minima

September,
October
and November

December,
January
and February

March,
April
and May

June,
July
and August

Fig. S21
Simulated seasonal changes in temperature (te; ˚K) anomalies for different geomagnetic
and solar scenarios. Anomalies are shown for a weakened geomagnetic field during the
Laschamps (M0P800) and a weakened geomagnetic field and Grand Solar Minima
(M0P0). Anomalies calculated as a 60-year average of 60 seasons. Colored areas denote
10% significance.

43
 Geomagnetic Minimum Geomagnetic Minimum +
Grand Solar Minima

September,
October
and November

December,
January
and February

March,
April
and May

June,
July
and August

Fig. S22
Simulated seasonal changes in zonal wind (u; m/s) anomalies for different geomagnetic
and solar scenarios. Anomalies are shown for a weakened geomagnetic field during the
Laschamps (M0P800) and a weakened geomagnetic field and Grand Solar Minima
(M0P0). Anomalies calculated as a 60-year average of 60 seasons. Colored areas denote
10% significance.

44
 Geomagnetic Minimum Geomagnetic Minimum +
Grand Solar Minima

September,
October
and November

December,
January
and February

March,
April
and May

June,
July
and August

Fig. S23
Simulated seasonal changes in total ozone (O3; %) anomalies for different geomagnetic
and solar scenarios. Anomalies are shown for a weakened geomagnetic field during the
Laschamps (M0P800) and a weakened geomagnetic field and Grand Solar Minima
(M0P0). Anomalies calculated as a 60-year average of 60 seasons. Hatched areas denote
10% significance level.

45
 Geomagnetic Minimum Geomagnetic Minimum +
Grand Solar Minima

September,
October
and November

December,
January
and February

March,
April
and May

Missing

June,
July Missing
and August

Fig. S24
Simulated seasonal changes in relative erythemal dose (H_er; percentage change)
anomalies for different geomagnetic and solar scenarios. Anomalies are shown for a
weakened geomagnetic field during the Laschamps (M0P800) and a weakened
geomagnetic field and Grand Solar Minima (M0P0). Anomalies calculated as a 60-year
average of 60 seasons. Hatched areas denote 10% significance level.

46
 Geomagnetic Minimum Geomagnetic Minimum +
Grand Solar Minima

September,
October
and November

December,
January
and February

March,
April
and May

June,
July
and August

Fig. S25
Simulated seasonal changes in sea-level pressure (SLP; hPa) anomalies for different
geomagnetic and solar scenarios. Anomalies are shown for a weakened geomagnetic field
during the Laschamps (M0P800) and a weakened geomagnetic field and Grand Solar
Minima (M0P0). Anomalies calculated as a 60-year average of 60 seasons. Hatched areas
denote 10% significance level.

47
 Geomagnetic Minimum Geomagnetic Minimum +
Grand Solar Minima

September,
October
and November

December,
January
and February

March,
April
and May

June,
July
and August

Fig. S26
Simulated seasonal changes in surface temperature anomalies (˚K) for different
geomagnetic and solar scenarios. Anomalies are shown for a weakened geomagnetic field
during the Laschamps (M0P800) and a weakened geomagnetic field and Grand Solar
Minima (M0P0). Anomalies calculated as a 60-year average of 60 seasons. Hatched areas
denote 10% significance level.

48
 Geomagnetic Minimum Geomagnetic Minimum +
Grand Solar Minima

September,
October
and November

December,
January
and February

March,
April
and May

June,
July
and August

Fig. S27
Simulated seasonal changes in surface zonal wind (u; 10 metres; m/s) anomalies for
different geomagnetic and solar scenarios. Anomalies are shown for a weakened
geomagnetic field during the Laschamps (M0P800) and a weakened geomagnetic field
and Grand Solar Minima (M0P0). Anomalies calculated as a 60-year average of 60
seasons. Red colors denote west-to-east airflow; blue, east-to-west airflow; hatched areas
denote 10% significance level.

49
 Geomagnetic Minimum Geomagnetic Minimum +
Grand Solar Minima

September,
October
and November

December,
January
and February

March,
April
and May

June,
July
and August

Fig. S28
Simulated seasonal changes in surface wind speed (10 metres; m/s) anomalies for
different geomagnetic and solar scenarios. Anomalies are shown for a weakened
geomagnetic field during the Laschamps (M0P800) and a weakened geomagnetic field
and Grand Solar Minima (M0P0). Anomalies calculated as a 60-year average of 60
seasons. Hatched areas denote 10% significance level.

50
 Geomagnetic Minimum
Geomagnetic Minimum Geomagnetic Minimum
Geomagnetic Minimum + +
Grand Solar Minima
Grand Solar Minima

September,
October
and November

December,
January
and February

March,
April
and May

June,
July
and August

Fig. S29
Simulated seasonal changes in surface meridional wind (v; 10 metres; m/s) anomalies for
different geomagnetic and solar scenarios. Anomalies are shown for a weakened
geomagnetic field during the Laschamps (M0P800) and a weakened geomagnetic field
and Grand Solar Minima (M0P0). Anomalies calculated as a 60-year average of 60
seasons. Red colors denote south-to-north airflow; blue, north-to-south airflow; hatched
areas denote 10% significance level.

51
 Geomagnetic Minimum Geomagnetic Minimum +
Grand Solar Minima

September,
October
and November

December,
January
and February

March,
April
and May

June,
July
and August

Fig. S30
Simulated seasonal changes in precipitation (mm/day) anomalies for different
geomagnetic and solar scenarios. Anomalies are shown for a weakened geomagnetic field
during the Laschamps (M0P800) and a weakened geomagnetic field and Grand Solar
Minima (M0P0). Anomalies calculated as a 60-year average of 60 seasons. Hatched areas
denote 10% significance level.

52
 OxA−13914
OxA−13881
OxA−13915
OxA−V−2333−35
OxA−21752
OxA−21807
OxA−21754
OxA−21791
OxA−11796
OxA−V−2344−12
NZA−2315
OxA−21986
Gro−2574/75
OxA−V−2344−13
OxA−19459
OxA−21825
OxA−21702
OxA−13592
OxA−V−2344−11
OxA−21808
OxA−21751
OxA−V−2344−14
GrN−2579
NZA−2312
OxA−V−2344−18
NZA−2314
CAMS−
OxA−22120
OxA−21790
GrN−13984
OxA−21753
RTT−3844
RTT−3845
X−2286−13
OxA−X−2286−13
OxA−X−2286−14
X−2286−14
OxA−X−2275−45
OxA−21638
EVA−33*
OxA−11976
OxA−21690
OxA−16998
OxA−11797
EVA−34*
OxA−17566
OxA−21757
NZA−2313
OxA−21758
OxA−21805
OxA−17980
OxA−X−2226−12
OxA−X−2279−45
OxA−X−2300−21
Lab Number

OxA−X−2279−18
OxA−21789
OxA−22121
OxA−21576
OxA−21824
OxA−21809
EVA−26*
OxA−21712
OxA−21765
MAMS−10832
OxA−4782
OxA−12025
RTT−3525
OxA−21839
MAMS−10831
EVA−41*
OxA−21657
OxA−21706
OxA−V−2384−11
OxA−21637
OxA−21704
OxA−V−2384−13
OxA−19994
OxA−21685
GrN−13982
EVA−42*
EVA−30*
OxA−X−2279−46
OxA−X−2226−13
OxA−23199
OxA−21574
OxA−X−2286−10
X−2286−10
OxA−21658
OxA−22562
OxA−21707
OxA−22561
OxA−X−2226−7
OxA−22564
OxA−21565
OxA−22563
OxA−21595
OxA−21777
OxA−X−2300−19
OxA−22153
OxA−22560
OxA−21662
MAMS−25149
OxA−X−2326−22
GrA−32626
OxA−21636
P27311 comb
OxA−21700
OxA−21594
OxA−21792
OxA−10560
OxA−21836
OxA−18099
GrA−32623
OxA−21699
OxA−21592
EVA−29*
OxA−21720
OxA−21593
OxA−V−2384−12
OxA−21591
OxA−21577

40000 45000 50000 55000

Calendar Age, BP

Fig. S31
Recalibration of Neandertal extinction. High-quality radiocarbon dates of last occurrence
dates of Neandertal specimens and Mousterian levels in European sites (37, 132)
recalibrated with the kauri-Hulu curve and the OxCal Phase function (72) to estimate the
end date of Neandertal presence (40.9-40.5k cal BP; blue column). Radiocarbon dating
lab numbers are given, along with confidence intervals (1), and the relationship to
Laschamps (red column) is shown. The extinction of Neandertals overlaps with the final
transition phase of Laschamps, contemporaneous with GS-10.

53
Fig. S32
Location of Boomplaas Cave, South Africa. Map generated from OpenStreetMap
contributors.

54
 1.5
Squared Chord Distance
A Pleistocene-Holocene boundary
Taxonomic Change

1
0.5
0

BRL

CL

YOL

BOL
LP

BP

OLP
BLD

LPC

OCH

LOH
BLA

GWA
Stratum

-4 B
δ13C (‰ PDB)

-8

-12
0 10 20 30 40 50
Age (ka)

Fig. S33
Abundance data from Boomplaas Cave, South Africa. (A) Squared chord distance—a
measure of taxonomic dissimilarity for abundance data—calculated across successive
pairs of stratigraphic units across the Boomplaas Cave large mammal sequence (42).
Grey shading highlights the marked faunal change that occurred in the transition into
stratum OLP to BP at ~40 ka (stratigraphy and ages follow refs. (139, 140)). (B) δ13C of
the Cango Cave stalagmite (143); shaded bands indicate 95% confidence interval for age-
depth model (142).

55
Fig. S34
Summary age distributions of rock art (cave paintings and petroglyphs). Probability age
distributions for rock art compiled from 209 age estimates obtained from Europe, Island
Southeast Asia and Australia (34, 36, 146, 149-175). For radiocarbon ages older than 25
ka BP, calibration was undertaken using the new kauri-Hulu curve reported here; for ages
younger than 25 ka BP, we used IntCal20 (see Table S8). Note: the oldest ages reported
are of possible Neandertal origin (174).

56
Table S1.
Ancient kauri radiocarbon ages calibrated against the Hulu Cave 14C dataset (22).
[Provided as an Excel file].

57
Table S2.
The CQL code for the Bayesian (OxCal) age model used to generate the Pillar Rock
(subantarctic Auckland Islands) chronology.

Plot()
{
Curve("KauriHulu.14c", "KauriHulu.14c");
Outlier_Model("General",T(5),U(0,4),"t");
P_Sequence("Pillar Rock",1,10,U(-2,2))
{
Boundary();
R_Date(" 48790",50170,1876)
{
Outlier(0.05);
z=13.0;
};
R_Date(" 48789",50679,2016)
{
Outlier(0.05);
z=12.0;
};
R_Date(" 48788",50681,2026)
{
Outlier(0.05);
z=11.25;
};
R_Date(" 48787",46376,1126)
{
Outlier(0.05);
z=10.75;
};
R_Date("48786",46162,1091)

58
{
Outlier(0.05);
z=10.25;
};
R_Date(" 48785",47914,1378)
{
Outlier(0.05);
z=9.75;
};
R_Date(" 48784",43047,727)
{
Outlier(0.05);
z=9.25;
};
R_Date(" 48783",44727,916)
{
Outlier(0.05);
z=8.75;
};
R_Date(" 48782",42519,683)
{
Outlier(0.05);
z=8.25;
};
R_Date(" 48781",40780,569)
{
Outlier(0.05);
z=7.75;
};
R_Date("48780",38627,414)
{

59
 Outlier(0.05);
z=7.25;
};
R_Date("43916",37956,636)
{
Outlier(0.05);
z=6.75;
};
Boundary();
};
};

60
Table S3.
Pillar Rock (subantarctic Auckland Islands) radiocarbon ages. Depths are given relative
to an arbitrary datum near the top of each stratigraphic unit. * denotes anomalous ages
not used in age model. AMS 14C dates obtained by University of Waikato (Wk).

Depth, Lab. Mean cal

cm no 14C age 1 age BP Mean cal 1
Upper
Peat
5 43918 9368 20 10,540 50
18 43919 11,695 28 13,500 40

Lignite
6.5-7.0 48779* 35,108 272
6.5-7.0 43916 37,956 636 42,230 220
7.0-7.5 48780 38,627 414 42,580 180
7.5-8.0 48781 40,780 569 43,780 420
8.0-8.5 48782 42,519 683 45,100 460
8.5-9.0 48783 44,727 916 46,150 510
9.0-9.5 48784 43,047 727 46,660 550
9.5-10.0 48785 47,914 1378 47,820 520
10.0-10.5 48786 46,162 1091 48,250 440
10.5-11.0 48787 46,376 1126 48,630 380
11.0-11.5 48788 50,681 2026 48,990 320
11.5-12.5 48789 50,679 2016 49,360 270
12.5-13.5 48790 50,170 1876 49,800 190
13.5-14.5 48791* 44,602 890
13.5-14.5 43917* 41,948 1077

61
Table S4.
The CQL code for the Bayesian (OxCal) age model used to generate the Lake Towuti
(Sulawesi) chronology.

Plot()
{
Curve("KauriHulu.14c",KauriHulu.14c);
Outlier_Model("General",T(5),U(0,4),"t");
Delta_R("uniform",U(0,1500));
P_Sequence("Lake Towuti",1,0.1,U(-2,2))
{
Boundary();
R_Date("OS-86972",42100,620)
{
Outlier(0.05);
z=895;
};
R_Date("OS-86951",39400,340)
{
Outlier(0.05);
z=865;
};
R_Date("OS-86950",37600,420)
{
Outlier(0.05);
z=785;
};
R_Date("Wk-50026",40174,505)
{
Outlier(0.05);
z=750;
};

62
R_Date("Wk-50025",40174,506)
{
Outlier(0.05);
z=749;
};
R_Date("Wk-50024",40545,530)
{
Outlier(0.05);
z=748;
};
R_Date("Wk-50023",40511,551)
{
Outlier(0.05);
z=747;
};
R_Date("Wk-50022",40131,506)
{
Outlier(0.05);
z=746;
};
R_Date("OS-86949",36600,360)
{
Outlier(0.05);
z=745;
};
R_Date("Wk-50020",39891,488)
{
Outlier(0.05);
z=744;
};
R_Date("Wk-50019",39549,472)

63
{
Outlier(0.05);
z=743;
};
R_Date("Wk-50018",38951,433)
{
Outlier(0.05);
z=742;
};
R_Date("Wk-50017",38121,393)
{
Outlier(0.05);
z=741;
};
R_Date("Wk-50016",39320,456)
{
Outlier(0.05);
z=740;
};
R_Date("Wk-50015",38781,427)
{
Outlier(0.05);
z=739;
};
R_Date("Wk-50014",38025,393)
{
Outlier(0.05);
z=738;
};
R_Date("OS-92843",30400,140)
{

64
 Outlier(0.05);
z=713;
};
R_Date("OS-89973",28400,180)
{
Outlier(0.05);
z=675;
};
R_Date("OS-86836",29300,230)
{
Outlier(0.05);
z=639;
};
R_Date("OS-86855",25900,240)
{
Outlier(0.05);
z=569;
};
Boundary();
};
};

65
Table S5.
Lake Towuti (Sulawesi) radiocarbon ages calibrated against the kauri-Hulu Cave dataset
used during this study. * denotes magnetic minimum identified in the Lake Towuti
sediments. To accommodate a changing sediment reservoir age (ΔR) through the Lake
Towuti sequence (11), a Delta_R with the prior U(0,1500) was used.

Mean cal age

Depth, cm Lab. no 14C age 1 BP Mean cal 1
569 OS-86855 25900 240 30,510 140
639 OS-86836 29300 230 32,900 720
675 OS-89973 28400 180 33,490 700
713 OS-92843 30400 140 34,810 250
738* Wk-50014 38025 393 42,350 170
739 Wk-50015 38781 427 42,430 120
740 Wk-50016 39320 456 42,460 120
741 Wk-50017 38121 393 42,480 120
742 Wk-50018 38951 433 42,570 170
743 Wk-50019 39549 472 42,800 210
744 Wk-50020 39891 488 42,930 180
745 OS-86949 36600 360 42,990 180
746 Wk-50022 40131 506 43,060 150
747 Wk-50023 40511 551 43,090 140
748 Wk-50024 40545 530 43,110 150
749 Wk-50025 40174 506 43,130 150
750 Wk-50026 40174 505 43,140 160
785 OS-86950 37600 420 43,720 450
865 OS-86951 39400 340 45,240 940
895 OS-86972 42100 620 45,910 1110

66
Table S6.
Orbital parameters for global chemistry climate model runs.

Eccentricity OMEGA Obliquity Precession

0.014215 16 23.769 0.00342

67
Table S7.
Greenhouse gas and ozone depleting substances (ODS) for global chemistry climate
model runs.

CH4 CO2 N2O CH3Br CH3Cl Other ODS

450 ppbv 210 ppmv 230 ppbv 3 pptv 450 pptv 0

Table S8.
The CQL code for the Bayesian (OxCal) age model used to generate the summed
probability plot for rock art (Fig. S34). Names given to individual ages relate to lead
author of study (34, 36, 146, 149-175).

Options()

};

Plot()

Curve("KauriHulu.14c", "KauriHulu.14c");

Sum("Art")

R_Date("O'Connor", 42800, 925);

R_Date("O'Connor", 33600, 250);

R_Date("Valladas", 32410, 720);

R_Date("Valladas", 32360, 490);

R_Date("Valladas", 31350, 620);

R_Date("Valladas", 30940, 610);

R_Date("Valladas", 30340, 570);

R_Date("Ambert", 30260, 110);

68
R_Date("Valladas", 30230, 530);

R_Date("Valladas", 29000, 400);

R_Date("Valladas", 28370, 440);

R_Date("Valladas", 27740, 410);

R_Date("Valladas", 27110, 400);

R_Date("Valladas", 26980, 410);

R_Date("Valladas", 26250, 350);

R_Date("Valladas", 26230, 280);

R_Date("Valladas", 26120, 400);

C_Date("Aubert", calBP(35700), 570);

C_Date("Aubert", calBP(44410), 490);

C_Date("Aubert", calBP(40830), 340);

C_Date("Aubert", calBP(41320), 380);

C_Date("Aubert_2019", calBP(35700), 285);

C_Date("Aubert_2019", calBP(44410), 245);

C_Date("Aubert_2019", calBP(41260), 120);

C_Date("Aubert_2019", calBP(41320), 190);

C_Date("Hoffmann", calBP(79660), 7450);

C_Date("Hoffmann", calBP(70080), 1910);

C_Date("Hoffmann", calBP(68130), 1480);

C_Date("Hoffmann", calBP(62970), 345);

C_Date("Aubert", calBP(50281), 775);

C_Date("Hoffmann", calBP(48230), 2215);

69
C_Date("Hoffmann", calBP(47640), 535);

C_Date("Hoffmann", calBP(46060), 405);

C_Date("Aubert", calBP(44000), 4550);

C_Date("Hoffmann", calBP(42470), 1535);

C_Date("Pike", calBP(41400), 285);

C_Date("Aubert", calBP(40881), 422);

C_Date("Aubert", calBP(40700), 435);

C_Date("Hoffmann", calBP(40420), 895);

C_Date("Aubert", calBP(40300), 1550);

C_Date("Aubert", calBP(39927), 286);

C_Date("Aubert", calBP(39670), 160);

C_Date("Aubert", calBP(38800), 800);

C_Date("Pike", calBP(37630), 170);

C_Date("Aubert", calBP(36900), 800);

C_Date("Pike", calBP(36160), 305);

C_Date("Pike", calBP(35720), 130);

C_Date("Pike", calBP(35540), 195);

C_Date("Aubert", calBP(34980), 205);

C_Date("Pike", calBP(34250), 85);

C_Date("Aubert", calBP(32600), 380);

C_Date("Hoffmann", calBP(32350), 135);

C_Date("Aubert", calBP(30900), 850);

C_Date("Pike", calBP(29650), 275);

70
C_Date("Aubert", calBP(29300), 600);

C_Date("Aubert", calBP(29300), 600);

C_Date("Aubert", calBP(29100), 1600);

C_Date("Aubert", calBP(28100), 330);

C_Date("Aubert", calBP(26086), 1248);

C_Date("Aubert", calBP(24900), 1550);

C_Date("Pike", calBP(24340), 60);

C_Date("Aubert", calBP(24000), 550);

C_Date("Aubert", calBP(24000), 750);

C_Date("Pike", calBP(22880), 135);

C_Date("Pike", calBP(22110), 65);

C_Date("Aubert", calBP(20739), 64);

C_Date("Aubert", calBP(19795), 59);

C_Date("Aubert", calBP(19700), 500);

C_Date("Aubert", calBP(17770), 210);

C_Date("Aubert", calBP(17572), 52);

C_Date("Aubert", calBP(16291), 47);

C_Date("Aubert", calBP(15645), 50);

C_Date("Aubert", calBP(14930), 180);

C_Date("Pike", calBP(14400), 850);

C_Date("Aubert", calBP(13800), 135);

C_Date("Pike", calBP(13020), 210);

C_Date("Plagnes", calBP(9870), 30);

71
C_Date("Aubert", calBP(9490), 90);

C_Date("Tacon", calBP(9400), 3000);

C_Date("Pike", calBP(8200), 430);

C_Date("Aubert", calBP(6300), 28);

C_Date("Tacon", calBP(2300), 125);

C_Date("Aubert", calBP(659), 14);

C_Date("Roberts", calBP(17453), 898);

C_Date("Ross", calBP(15953), 499);

C_Date("Ross", calBP(3233), 94);

C_Date("Ross", calBP(653), 19);

C_Date("Ross", calBP(1373), 116);

C_Date("Ross", calBP(1583), 53);

C_Date("Ross", calBP(483), 36);

C_Date("Ross", calBP(883), 24);

C_Date("Ross", calBP(603), 56);

C_Date("Ross", calBP(5053), 12);

Curve("IntCal20","IntCal20.14c");

R_Date("Valladas", 24840, 340);

R_Date("Valladas", 24730, 300);

R_Date("Cole", 24600, 110);

R_Date("David", 22965, 109);

R_Date("Clottes", 21650, 400);

R_Date("Valladas", 20790, 340);

72
R_Date("Clottes", 20210, 130);

R_Date("Clottes", 19970, 125);

R_Date("Valladas", 19720, 210);

R_Date("Valladas", 19340, 200);

R_Date("Valladas", 19290, 340);

R_Date("Valladas", 19200, 240);

R_Date("Valladas", 18840, 250);

R_Date("Valladas", 18010, 200);

R_Date("Valladas", 17800, 160);

R_Date("Valladas", 14330, 190);

R_Date("Valladas", 14260, 130);

R_Date("Valladas", 14060, 70);

R_Date("Clottes", 13940, 125);

R_Date("Valladas", 13940, 170);

R_Date("Clottes", 13900, 60);

R_Date("Clottes", 13810, 370);

R_Date("Clottes", 13640, 55);

R_Date("Valladas", 13570, 190);

R_Date("Valladas", 13520, 120);

R_Date("Clottes", 13400, 60);

R_Date("Clottes", 13400, 500);

R_Date("Clottes", 13200, 55);

R_Date("Valladas", 12910, 180);

73
R_Date("Clottes", 12900, 70);

R_Date("Clottes", 12850, 30);

R_Date("Clottes", 12770, 21);

R_Date("Clottes", 12760, 85);

R_Date("Valladas", 12620, 110);

R_Date("Clottes", 12540, 52.5);

R_Date("Morwood", 12400, 40);

R_Date("Clottes", 11650, 100);

R_Date("Valladas", 10510, 100);

R_Date("Cole", 9160, 35);

R_Date("Cole", 8500, 30);

R_Date("Jones", 8290, 20);

R_Date("Steelman", 7920, 30);

R_Date("Steelman", 7370, 25);

R_Date("Cole", 7230, 375);

R_Date("Cole", 7230, 55);

R_Date("Fontugne", 7165, 15);

R_Date("Jones", 6570, 97.5);

R_Date("Cole", 6530, 690);

R_Date("Jones", 6260, 17.5);

R_Date("Steelman", 6030, 55);

R_Date("Jones", 5404, 22.5);

R_Date("Jones", 5340, 25);

74
R_Date("Jones", 5290, 17.5);

R_Date("Fontugne", 4520, 15);

R_Date("Fontugne", 4390, 15);

R_Date("Watchman", 3880, 55);

R_Date("Morwood", 3780, 30);

R_Date("Jones", 3710, 17.5);

R_Date("Morwood", 3540, 25);

R_Date("Watchman", 3140, 175);

R_Date("Cole", 2850, 57.5);

R_Date("Cole", 2810, 75);

R_Date("Cole", 2785, 42.5);

R_Date("Bischoff", 2760, 195);

R_Date("Fontugne", 2455, 15);

R_Date("Morwood", 2220, 50);

R_Date("Morwood", 2200, 30);

R_Date("Watchman", 2200, 50);

R_Date("Cole", 2195, 27.5);

R_Date("Morwood", 1850, 20);

R_Date("Fontugne", 1770, 15);

R_Date("Morwood", 1680, 25);

R_Date("Watchman", 1680, 25);

R_Date("Fontugne", 1590, 15);

R_Date("Morwood", 1580, 30);

75
R_Date("Morwood", 1510, 40);

R_Date("Watchman", 1510, 40);

R_Date("Watchman", 1490, 145);

R_Date("Watchman", 1490, 25);

R_Date("Morwood", 1480, 30);

R_Date("Morwood", 1440, 60);

R_Date("Watchman", 1440, 60);

R_Date("Watchman", 1430, 90);

R_Date("Morwood", 1370, 35);

R_Date("Morwood", 1370, 15);

R_Date("Ross", 1290, 15);

R_Date("Ross", 1285, 15);

R_Date("Cole", 1275, 47.5);

R_Date("Morwood", 1270, 45);

R_Date("Watchman", 1270, 45);

R_Date("Ross", 1230, 17.5);

R_Date("Morwood", 1220, 40);

R_Date("Watchman", 1220, 40);

R_Date("Cole", 1210, 122.5);

R_Date("Morwood", 1210, 70);

R_Date("Morwood", 1200, 45);

R_Date("Morwood", 1170, 30);

R_Date("Watchman", 1170, 30);

76
R_Date("Morwood", 1160, 20);

R_Date("Morwood", 1120, 45);

R_Date("Morwood", 1090, 25);

R_Date("Morwood", 1060, 25);

R_Date("Watchman", 1060, 25);

R_Date("Morwood", 910, 20);

R_Date("Morwood", 860, 15);

R_Date("Morwood", 805, 27.5);

R_Date("Morwood", 730, 80);

R_Date("Watchman", 730, 80);

R_Date("Morwood", 695, 20);

R_Date("Morwood", 670, 45);

R_Date("Watchman", 670, 45);

R_Date("Morwood", 630, 62.5);

R_Date("Morwood", 610, 20);

R_Date("Morwood", 550, 50);

R_Date("Watchman", 550, 50);

R_Date("Morwood", 430, 65);

R_Date("Watchman", 430, 65);

R_Date("Morwood", 375, 17.5);

R_Date("Morwood", 300, 42.5);

R_Date("Ross", 155, 15);

};

77
};

78
References and Notes
1. C. Laj, H. Guillou, C. Kissel, Dynamics of the Earth magnetic field in the 10–75 kyr period
comprising the Laschamp and Mono Lake excursions: New results from the French
Chaîne des Puys in a global perspective. Earth Planet. Sci. Lett. 387, 184–197 (2014).
doi:10.1016/j.epsl.2013.11.031
2. M. Brown, M. Korte, R. Holme, I. Wardinski, S. Gunnarson, Earth’s magnetic field is
probably not reversing. Proc. Natl. Acad. Sci. U.S.A. 115, 5111–5116 (2018).
doi:10.1073/pnas.1722110115 Medline
3. E. J. Oughton, A. Skelton, R. B. Horne, A. W. P. Thomson, C. T. Gaunt, Quantifying the daily
economic impact of extreme space weather due to failure in electricity transmission
infrastructure. Space Weather 15, 65–83 (2017). doi:10.1002/2016SW001491
4. J. E. T. Channell, L. Vigliotti, The role of geomagnetic field intensity in late Quaternary
evolution of humans and large mammals. Rev. Geophys. 57, 709–738 (2019).
doi:10.1029/2018RG000629
5. J. Liu, N. R. Nowaczyk, S. Panovska, M. Korte, H. W. Arz, The Norwegian-Greenland Sea,
the Laschamps and the Mono Lake recorded in Black Sea paleosecular variation records
between 68.9 and 14.5 ka. J. Geophys. Res. 125, e2019JB019225 (2020).
doi:10.1029/2019JB019225
6. G. Wagner, D. M. Livingstone, J. Masarik, R. Muscheler, J. Beer, Some results relevant to the
discussion of a possible link between cosmic rays and the Earth’s climate. J. Geophys.
Res. 106, 3381–3387 (2001). doi:10.1029/2000JD900589
7. R. Zech, J. Zech, C. Kull, P. W. Kubik, H. Veit, Early Last Glacial Maximum in the southern
Central Andes reveals northward shift of the westerlies at ~39 ka. Clim. Past 7, 41–46
(2011). doi:10.5194/cp-7-41-2011
8. I. Suter, R. Zech, J. Anet, T. Peter, Impact of geomagnetic excursions on atmospheric
chemistry and dynamics. Clim. Past 10, 1183–1194 (2014). doi:10.5194/cp-10-1183-
2014
9. J. Kirkby, J. Duplissy, K. Sengupta, C. Frege, H. Gordon, C. Williamson, M. Heinritzi, M.
Simon, C. Yan, J. Almeida, J. Tröstl, T. Nieminen, I. K. Ortega, R. Wagner, A. Adamov,
A. Amorim, A.-K. Bernhammer, F. Bianchi, M. Breitenlechner, S. Brilke, X. Chen, J.
Craven, A. Dias, S. Ehrhart, R. C. Flagan, A. Franchin, C. Fuchs, R. Guida, J. Hakala, C.
R. Hoyle, T. Jokinen, H. Junninen, J. Kangasluoma, J. Kim, M. Krapf, A. Kürten, A.
Laaksonen, K. Lehtipalo, V. Makhmutov, S. Mathot, U. Molteni, A. Onnela, O. Peräkylä,
F. Piel, T. Petäjä, A. P. Praplan, K. Pringle, A. Rap, N. A. D. Richards, I. Riipinen, M. P.
Rissanen, L. Rondo, N. Sarnela, S. Schobesberger, C. E. Scott, J. H. Seinfeld, M. Sipilä,
G. Steiner, Y. Stozhkov, F. Stratmann, A. Tomé, A. Virtanen, A. L. Vogel, A. C.
Wagner, P. E. Wagner, E. Weingartner, D. Wimmer, P. M. Winkler, P. Ye, X. Zhang, A.
Hansel, J. Dommen, N. M. Donahue, D. R. Worsnop, U. Baltensperger, M. Kulmala, K.
S. Carslaw, J. Curtius, Ion-induced nucleation of pure biogenic particles. Nature 533,
521–526 (2016). doi:10.1038/nature17953 Medline
10. P. D. Strand, J. M. Schaefer, A. E. Putnam, G. H. Denton, D. J. A. Barrell, T. N. B. Koffman,
R. Schwartz, Millennial-scale pulsebeat of glaciation in the Southern Alps of New
Zealand. Quat. Sci. Rev. 220, 165–177 (2019). doi:10.1016/j.quascirev.2019.07.022
11. J. M. Russell, H. Vogel, B. L. Konecky, S. Bijaksana, Y. Huang, M. Melles, N. Wattrus, K.
Costa, J. W. King, Glacial forcing of central Indonesian hydroclimate since 60,000 y B.P.
Proc. Natl. Acad. Sci. U.S.A. 111, 5100–5105 (2014). doi:10.1073/pnas.1402373111
Medline
12. E. Rainsley, C. S. M. Turney, N. R. Golledge, J. M. Wilmshurst, M. S. McGlone, A. G.
Hogg, B. Li, Z. A. Thomas, R. Roberts, R. T. Jones, J. G. Palmer, V. Flett, G. de Wet, D.
K. Hutchinson, M. J. Lipson, P. Fenwick, B. R. Hines, U. Binetti, C. J. Fogwill,
Pleistocene glacial history of the New Zealand subantarctic islands. Clim. Past 15, 423–
448 (2019). doi:10.5194/cp-15-423-2019
13. J. M. Bowler, H. Johnston, J. M. Olley, J. R. Prescott, R. G. Roberts, W. Shawcross, N. A.
Spooner, New ages for human occupation and climatic change at Lake Mungo, Australia.
Nature 421, 837–840 (2003). doi:10.1038/nature01383 Medline
14. F. Saltré, M. Rodríguez-Rey, B. W. Brook, C. N. Johnson, C. S. M. Turney, J. Alroy, A.
Cooper, N. Beeton, M. I. Bird, D. A. Fordham, R. Gillespie, S. Herrando-Pérez, Z.
Jacobs, G. H. Miller, D. Nogués-Bravo, G. J. Prideaux, R. G. Roberts, C. J. A. Bradshaw,
Climate change not to blame for late Quaternary megafauna extinctions in Australia. Nat.
Commun. 7, 10511 (2016). doi:10.1038/ncomms10511 Medline
15. S. Rule, B. W. Brook, S. G. Haberle, C. S. M. Turney, A. P. Kershaw, C. N. Johnson, The
aftermath of megafaunal extinction: Ecosystem transformation in Pleistocene Australia.
Science 335, 1483–1486 (2012). doi:10.1126/science.1214261 Medline
16. C. S. M. Turney, T. F. Flannery, R. G. Roberts, C. Reid, L. K. Fifield, T. F. G. Higham, Z.
Jacobs, N. Kemp, E. A. Colhoun, R. M. Kalin, N. Ogle, Late-surviving megafauna in
Tasmania, Australia, implicate human involvement in their extinction. Proc. Natl. Acad.
Sci. U.S.A. 105, 12150–12153 (2008). doi:10.1073/pnas.0801360105 Medline
17. H. W. Hill, B. P. Flower, T. M. Quinn, D. J. Hollander, T. P. Guilderson, Laurentide Ice
Sheet meltwater and abrupt climate change during the last glaciation. Paleoceanography
21, PA1006 (2006). doi:10.1029/2005PA001186
18. A. S. Dalton, S. A. Finkelstein, S. L. Forman, P. J. Barnett, T. Pico, J. X. Mitrovica, Was the
Laurentide Ice Sheet significantly reduced during marine isotope stage 3? Geology 47,
111–114 (2019). doi:10.1130/G45335.1
19. A. S. Dalton, S. A. Finkelstein, P. J. Barnett, S. L. Forman, Constraining the Late Pleistocene
history of the Laurentide Ice Sheet by dating the Missinaibi Formation, Hudson Bay
Lowlands, Canada. Quat. Sci. Rev. 146, 288–299 (2016).
doi:10.1016/j.quascirev.2016.06.015
20. G. M. Raisbeck, A. Cauquoin, J. Jouzel, A. Landais, J.-R. Petit, V. Y. Lipenkov, J. Beer, H.-
A. Synal, H. Oerter, S. J. Johnsen, J. P. Steffensen, A. Svensson, F. Yiou, An improved
north–south synchronization of ice core records around the 41 kyr 10Be peak. Clim. Past
13, 217–229 (2017). doi:10.5194/cp-13-217-2017
21. F. Adolphi, C. Bronk Ramsey, T. Erhardt, R. L. Edwards, H. Cheng, C. S. M. Turney, A.
Cooper, A. Svensson, S. O. Rasmussen, H. Fischer, R. Muscheler, Connecting the
Greenland ice-core and U/Th timescales via cosmogenic radionuclides: Testing the
synchroneity of Dansgaard–Oeschger events. Clim. Past 14, 1755–1781 (2018).
doi:10.5194/cp-14-1755-2018
22. H. Cheng, R. L. Edwards, J. Southon, K. Matsumoto, J. M. Feinberg, A. Sinha, W. Zhou, H.
Li, X. Li, Y. Xu, S. Chen, M. Tan, Q. Wang, Y. Wang, Y. Ning, Atmospheric 14C/12C
changes during the last glacial period from Hulu Cave. Science 362, 1293–1297 (2018).
doi:10.1126/science.aau0747 Medline
23. R. A. Staff, M. Hardiman, C. Bronk Ramsey, F. Adolphi, V. J. Hare, A. Koutsodendris, J.
Pross, Reconciling the Greenland ice-core and radiocarbon timescales through the
Laschamp geomagnetic excursion. Earth Planet. Sci. Lett. 520, 1–9 (2019).
doi:10.1016/j.epsl.2019.05.021
24. A. M. Lorrey, G. Boswijk, A. Hogg, J. G. Palmer, C. S. M. Turney, A. M. Fowler, J. Ogden,
J.-M. Woolley, The scientific value and potential of New Zealand swamp kauri. Quat.
Sci. Rev. 183, 124–139 (2018). doi:10.1016/j.quascirev.2017.12.019
25. C. S. M. Turney, J. Palmer, M. A. Maslin, A. Hogg, C. J. Fogwill, J. Southon, P. Fenwick, G.
Helle, J. M. Wilmshurst, M. McGlone, C. Bronk Ramsey, Z. Thomas, M. Lipson, B.
Beaven, R. T. Jones, O. Andrews, Q. Hua, Global Peak in Atmospheric Radiocarbon
Provides a Potential Definition for the Onset of the Anthropocene Epoch in 1965. Sci.
Rep. 8, 3293 (2018). doi:10.1038/s41598-018-20970-5 Medline
26. K. Kodera, Y. Kuroda, Dynamical response to the solar cycle. J. Geophys. Res. 107, ACL 5-
1–ACL 5-12 (2002).
27. J. Kidston, A. A. Scaife, S. C. Hardiman, D. M. Mitchell, N. Butchart, M. P. Baldwin, L. J.
Gray, Stratospheric influence on tropospheric jet streams, storm tracks and surface
weather. Nat. Geosci. 8, 433–440 (2015). doi:10.1038/ngeo2424
28. V. Masson-Delmotte, J. Jouzel, A. Landais, M. Stievenard, S. J. Johnsen, J. W. White, M.
Werner, A. Sveinbjornsdottir, K. Fuhrer, GRIP deuterium excess reveals rapid and
orbital-scale changes in Greenland moisture origin. Science 309, 118–121 (2005).
doi:10.1126/science.1108575 Medline
29. S. O. Rasmussen, M. Bigler, S. P. Blockley, T. Blunier, S. L. Buchardt, H. B. Clausen, I.
Cvijanovic, D. Dahl-Jensen, S. J. Johnsen, H. Fischer, V. Gkinis, M. Guillevic, W. Z.
Hoek, J. J. Lowe, J. B. Pedro, T. Popp, I. K. Seierstad, J. P. Steffensen, A. M. Svensson,
P. Vallelonga, B. M. Vinther, M. J. C. Walker, J. J. Wheatley, M. Winstrup, A
stratigraphic framework for abrupt climatic changes during the Last Glacial period based
on three synchronized Greenland ice-core records: Refining and extending the
INTIMATE event stratigraphy. Quat. Sci. Rev. 106, 14–28 (2014).
doi:10.1016/j.quascirev.2014.09.007
30. G. Deplazes, A. Lückge, L. C. Peterson, A. Timmermann, Y. Hamann, K. A. Hughen, U.
Röhl, C. Laj, M. A. Cane, D. M. Sigman, G. H. Haug, Links between tropical rainfall and
North Atlantic climate during the last glacial period. Nat. Geosci. 6, 213–217 (2013).
doi:10.1038/ngeo1712
31. M. Staubwasser, V. Drăgușin, B. P. Onac, S. Assonov, V. Ersek, D. L. Hoffmann, D. Veres,
Impact of climate change on the transition of Neanderthals to modern humans in Europe.
Proc. Natl. Acad. Sci. U.S.A. 115, 9116–9121 (2018). doi:10.1073/pnas.1808647115
Medline
32. J. F. O’Connell, J. Allen, M. A. J. Williams, A. N. Williams, C. S. M. Turney, N. A.
Spooner, J. Kamminga, G. Brown, A. Cooper, When did Homo sapiens first reach
Southeast Asia and Sahul? Proc. Natl. Acad. Sci. U.S.A. 115, 8482–8490 (2018).
doi:10.1073/pnas.1808385115 Medline
33. D. Adams, M. Carwardine, Last Chance to See (Arrow Books, 2009).
34. A. W. Pike, D. L. Hoffmann, M. García-Diez, P. B. Pettitt, J. Alcolea, R. De Balbín, C.
González-Sainz, C. de las Heras, J. A. Lasheras, R. Montes, J. Zilhão, U-series dating of
Paleolithic art in 11 caves in Spain. Science 336, 1409–1413 (2012).
doi:10.1126/science.1219957 Medline
35. S. O’Connor, R. Ono, C. Clarkson, Pelagic fishing at 42,000 years before the present and the
maritime skills of modern humans. Science 334, 1117–1121 (2011).
doi:10.1126/science.1207703 Medline
36. M. Aubert, P. Setiawan, A. A. Oktaviana, A. Brumm, P. H. Sulistyarto, E. W. Saptomo, B.
Istiawan, T. A. Ma’rifat, V. N. Wahyuono, F. T. Atmoko, J.-X. Zhao, J. Huntley, P. S. C.
Taçon, D. L. Howard, H. E. A. Brand, Palaeolithic cave art in Borneo. Nature 564, 254–
257 (2018). doi:10.1038/s41586-018-0679-9 Medline
37. J.-J. Hublin, N. Sirakov, V. Aldeias, S. Bailey, E. Bard, V. Delvigne, E. Endarova, Y.
Fagault, H. Fewlass, M. Hajdinjak, B. Kromer, I. Krumov, J. Marreiros, N. L. Martisius,
L. Paskulin, V. Sinet-Mathiot, M. Meyer, S. Pääbo, V. Popov, Z. Rezek, S. Sirakova, M.
M. Skinner, G. M. Smith, R. Spasov, S. Talamo, T. Tuna, L. Wacker, F. Welker, A.
Wilcke, N. Zahariev, S. P. McPherron, T. Tsanova, Initial Upper Palaeolithic Homo
sapiens from Bacho Kiro Cave, Bulgaria. Nature 581, 299–302 (2020).
doi:10.1038/s41586-020-2259-z Medline
38. A. Cooper, C. Turney, K. A. Hughen, B. W. Brook, H. G. McDonald, C. J. A. Bradshaw,
Abrupt warming events drove Late Pleistocene Holarctic megafaunal turnover. Science
349, 602–606 (2015). doi:10.1126/science.aac4315 Medline
39. B. S. Singer, B. R. Jicha, N. Mochizuki, R. S. Coe, Synchronizing volcanic, sedimentary, and
ice core records of Earth’s last magnetic polarity reversal. Sci. Adv. 5, eaaw4621 (2019).
doi:10.1126/sciadv.aaw4621 Medline
40. K. Hughen, S. Lehman, J. Southon, J. Overpeck, O. Marchal, C. Herring, J. Turnbull, 14C
activity and global carbon cycle changes over the past 50,000 years. Science 303, 202–
207 (2004). doi:10.1126/science.1090300 Medline
41. WAIS Divide Project Members, Precise interpolar phasing of abrupt climate change during
the last ice age. Nature 520, 661–665 (2015). doi:10.1038/nature14401 Medline
42. J. T. Faith, Taphonomic and paleoecological change in the large mammal sequence from
Boomplaas Cave, western Cape, South Africa. J. Hum. Evol. 65, 715–730 (2013).
doi:10.1016/j.jhevol.2013.09.001 Medline
43. J. Palmer, A. Lorrey, C. S. M. Turney, A. Hogg, M. Baillie, K. Fifield, J. Ogden, Extension
of New Zealand kauri (Agathis australis) tree-ring chronologies into Oxygen Isotope
Stage (OIS)-3. J. Quaternary Sci. 21, 779–787 (2006). doi:10.1002/jqs.1075
44. J. Palmer, C. S. M. Turney, A. G. Hogg, A. M. Lorrey, R. J. Jones, Progress in refining the
global radiocarbon calibration curve using New Zealand kauri (Agathis australis) tree-
ring series from Oxygen Isotope Stage 3. Quat. Geochronol. 27, 158–163 (2015).
doi:10.1016/j.quageo.2015.02.028
45. C. S. M. Turney, L. K. Fifield, A. G. Hogg, J. G. Palmer, K. Hughen, M. G. L. Baillie, R.
Galbraith, J. Ogden, A. Lorrey, S. G. Tims, R. T. Jones, The potential of New Zealand
kauri (Agathis australis) for testing the synchronicity of abrupt climate change during the
Last Glacial Interval (60,000–11,700 years ago). Quat. Sci. Rev. 29, 3677–3682 (2010).
doi:10.1016/j.quascirev.2010.08.017
46. P. G. Quilty, Origin and evolution of the sub-antarctic islands: The foundation. Pap. Proc. R.
Soc. Tasman. 141, 35–58 (2007). doi:10.26749/rstpp.141.1.35
47. C. S. M. Turney, C. J. Fogwill, J. G. Palmer, E. van Sebille, Z. Thomas, M. McGlone, S.
Richardson, J. M. Wilmshurst, P. Fenwick, V. Zunz, H. Goosse, K.-J. Wilson, L. Carter,
M. Lipson, R. T. Jones, M. Harsch, G. Clark, E. Marzinelli, T. Rogers, E. Rainsley, L.
Ciasto, S. Waterman, E. R. Thomas, M. Visbeck, Tropical forcing of increased Southern
Ocean climate variability revealed by a 140-year subantarctic temperature reconstruction.
Clim. Past 13, 231–248 (2017). doi:10.5194/cp-13-231-2017
48. C. S. Nelson, P. J. Cooke, C. H. Hendy, A. M. Cuthbertson, Oceanographic and climatic
changes over the past 160,000 years at Deep Sea Drilling Project Site 594 off
southeastern New Zealand, southwest Pacific Ocean. Paleoceanography 8, 435–458
(1993). doi:10.1029/93PA01162
49. M. S. McGlone, The Late Quaternary peat, vegetation and climate history of the Southern
Oceanic Islands of New Zealand. Quat. Sci. Rev. 21, 683–707 (2002).
doi:10.1016/S0277-3791(01)00044-0
50. C. S. M. Turney, M. McGlone, J. Palmer, C. Fogwill, A. Hogg, Z. A. Thomas, M. Lipson, J.
M. Wilmshurst, P. Fenwick, R. T. Jones, B. Hines, G. F. Clark, Intensification of
Southern Hemisphere westerly winds 2000-1000 years ago: Evidence from the
subantarctic Campbell and Auckland Islands (52-50°S). J. Quaternary Sci. 31, 12–19
(2016). doi:10.1002/jqs.2828
51. P. D. Moore, J. A. Webb, M. E. Collinson, Pollen Analysis (Blackwell Science, 1991).
52. G. Hope, Environmental change in the late Pleistocene and later Holocene at Wanda site,
Soroako, South Sulawesi, Indonesia. Palaeogeogr. Palaeoclimatol. Palaeoecol. 171,
129–145 (2001). doi:10.1016/S0031-0182(01)00243-7
53. E. Aldrian, R. Dwi Susanto, Identification of three dominant rainfall regions within Indonesia
and their relationship to sea surface temperature. Int. J. Climatol. 23, 1435–1452 (2003).
doi:10.1002/joc.950
54. M. H. O’Leary, Carbon isotope fractionation in plants. Phytochemistry 20, 553–567 (1981).
doi:10.1016/0031-9422(81)85134-5
55. M. D. Hatch, C4 photosynthesis: A unique blend of modified biochemistry, anatomy and
ultrastructure. Biochim. Biophys. Acta 895, 81–106 (1987). doi:10.1016/S0304-
4173(87)80009-5
56. I. S. Castañeda, S. Mulitza, E. Schefuss, R. A. Lopes dos Santos, J. S. Sinninghe Damsté, S.
Schouten, Wet phases in the Sahara/Sahel region and human migration patterns in North
Africa. Proc. Natl. Acad. Sci. U.S.A. 106, 20159–20163 (2009).
doi:10.1073/pnas.0905771106 Medline
57. K. A. Hughen, T. I. Eglinton, L. Xu, M. Makou, Abrupt tropical vegetation response to rapid
climate changes. Science 304, 1955–1959 (2004). doi:10.1126/science.1092995 Medline
58. A. F. Diefendorf, K. E. Mueller, S. L. Wing, P. L. Koch, K. H. Freeman, Global patterns in
leaf 13C discrimination and implications for studies of past and future climate. Proc. Natl.
Acad. Sci. U.S.A. 107, 5738–5743 (2010). doi:10.1073/pnas.0910513107 Medline
59. M. J. Kohn, Carbon isotope compositions of terrestrial C3 plants as indicators of
(paleo)ecology and (paleo)climate. Proc. Natl. Acad. Sci. U.S.A. 107, 19691–19695
(2010). doi:10.1073/pnas.1004933107 Medline
60. K. H. Kirana, S. Bijaksana, J. King, G. H. Tamuntuan, J. Russell, L. O. Ngkoimani, D.
Dahrin, S. J. Fajar, A high-resolution, 60 kyr record of the relative geomagnetic field
intensity from Lake Towuti, Indonesia. Phys. Earth Planet. Inter. 275, 9–18 (2018).
doi:10.1016/j.pepi.2017.12.007
61. C. S. G. Gogorza, J. M. Lirio, H. Nuñez, M. Chaparro, H. R. Bertorello, A. M. Sinito,
Paleointensity studies on Holocene–Pleistocene sediments from lake Escondido,
Argentina. Phys. Earth Planet. Inter. 145, 219–238 (2004).
doi:10.1016/j.pepi.2004.03.010
62. G. Tamuntuan, S. Bijaksana, J. King, J. Russell, U. Fauzi, K. Maryunani, N. Aufa, L. O.
Safiuddin, Variation of magnetic properties in sediments from Lake Towuti, Indonesia,
and its paleoclimatic significance. Palaeogeogr. Palaeoclimatol. Palaeoecol. 420, 163–
172 (2015). doi:10.1016/j.palaeo.2014.12.008
63. A. G. Hogg, L. Fifield, C. Turney, J. Palmer, R. Galbraith, M. Baillie, Dating ancient wood
by high sensitivity liquid scintillation counting and accelerator mass spectrometry—
Pushing the boundaries. Quat. Geochronol. 1, 241–248 (2006).
doi:10.1016/j.quageo.2006.11.001
64. A. Hogg, C. Turney, J. Palmer, J. Southon, B. Kromer, C. B. Ramsey, G. Boswijk, P.
Fenwick, A. Noronha, R. Staff, M. Friedrich, L. Reynard, D. Guetter, L. Wacker, R.
Jones, The New Zealand kauri (Agathis australis) Research Project: A radiocarbon dating
intercomparison of Younger Dryas wood and implications for IntCal13. Radiocarbon 55,
2035–2048 (2013). doi:10.2458/azu_js_rc.v55i2.16217
65. A. Hogg, in Liquid Scintillation Spectrometry 1992, J. E. Noakes, H. A. Polach, F.
Schonhofer, Eds. (Radiocarbon, 1993), pp. 135–142.
66. T. J. Heaton, M. Blaauw, P. G. Blackwell, C. Bronk Ramsey, P. J. Reimer, E. M. Scott, The
IntCal20 approach to radiocarbon calibration curve construction: A new methodology
using Bayesian splines and errors-in-variables. Radiocarbon 62, 821–863 (2020).
doi:10.1017/RDC.2020.46
67. P. J. Reimer, W. E. N. Austin, E. Bard, A. Bayliss, P. G. Blackwell, C. Bronk Ramsey, M.
Butzin, H. Cheng, R. L. Edwards, M. Friedrich, P. M. Grootes, T. P. Guilderson, I.
Hajdas, T. J. Heaton, A. G. Hogg, K. A. Hughen, B. Kromer, S. W. Manning, R.
Muscheler, J. G. Palmer, C. Pearson, J. van der Plicht, R. W. Reimer, D. A. Richards, E.
M. Scott, J. R. Southon, C. S. M. Turney, L. Wacker, F. Adolphi, U. Büntgen, M.
Capano, S. M. Fahrni, A. Fogtmann-Schulz, R. Friedrich, P. Köhler, S. Kudsk, F.
Miyake, J. Olsen, F. Reinig, M. Sakamoto, A. Sookdeo, S. Talamo, The IntCal20
Northern Hemisphere radiocarbon age calibration curve (0–55 cal kBP). Radiocarbon 62,
725–757 (2020). doi:10.1017/RDC.2020.41
68. A. G. Hogg, T. J. Heaton, Q. Hua, J. G. Palmer, C. S. M. Turney, J. Southon, A. Bayliss, P.
G. Blackwell, G. Boswijk, C. Bronk Ramsey, C. Pearson, F. Petchey, P. Reimer, R.
Reimer, L. Wacker, SHCal20 Southern Hemisphere calibration, 0–55,000 years cal BP.
Radiocarbon 62, 759–778 (2020). doi:10.1017/RDC.2020.59
69. J. Southon, A. L. Noronha, H. Cheng, R. L. Edwards, Y. Wang, A high-resolution record of
atmospheric 14C based on Hulu Cave speleothem H82. Quat. Sci. Rev. 33, 32–41 (2012).
doi:10.1016/j.quascirev.2011.11.022
70. K. A. Hughen, J. Southon, S. Lehman, C. Bertrand, J. Turnbull, Marine-derived 14C
calibration and activity record for the past 50,000 years updated from the Cariaco Basin.
Quat. Sci. Rev. 25, 3216–3227 (2006). doi:10.1016/j.quascirev.2006.03.014
71. G. Wagner, J. Beer, J. Masarik, R. Muscheler, P. W. Kubik, W. Mende, C. Laj, G. M.
Raisbeck, F. Yiou, Presence of the solar de Vries Cycle (~205 years) during the last ice
age. Geophys. Res. Lett. 28, 303–306 (2001). doi:10.1029/2000GL006116
72. C. Bronk Ramsey, S. Lee, Recent and planned developments of the program OxCal.
Radiocarbon 55, 720–730 (2013). doi:10.1017/S0033822200057878
73. C. Bronk Ramsey, Dealing with outliers and offsets in radiocarbon dating. Radiocarbon 51,
1023–1045 (2009). doi:10.1017/S0033822200034093
74. J. Masarik, J. Beer, Simulation of particle fluxes and cosmogenic nuclide production in the
Earth’s atmosphere. J. Geophys. Res. 104, 12099–12111 (1999).
doi:10.1029/1998JD200091
75. U. Siegenthaler, M. Heimann, H. Oeschger, 14C variations caused by changes in the global
carbon cycle. Radiocarbon 22, 177–191 (1980). doi:10.1017/S0033822200009449
76. R. Muscheler, J. Beer, G. Wagner, C. Laj, C. Kissel, G. M. Raisbeck, F. Yiou, P. W. Kubik,
Changes in the carbon cycle during the last deglaciation as indicated by the comparison
of 10Be and 14C records. Earth Planet. Sci. Lett. 219, 325–340 (2004).
doi:10.1016/S0012-821X(03)00722-2
77. N. R. Nowaczyk, H. W. Arz, U. Frank, J. Kind, B. Plessen, Dynamics of the Laschamp
geomagnetic excursion from Black Sea sediments. Earth Planet. Sci. Lett. 351-352, 54–
69 (2012). doi:10.1016/j.epsl.2012.06.050
78. A. C. Cummings, E. C. Stone, B. C. Heikkila, N. Lal, W. R. Webber, G. Jóhannesson, I. V.
Moskalenko, E. Orlando, T. A. Porter, Galactic cosmic rays in the local interstellar
medium: Voyager 1 observations and model results. Astrophys. J. 831, 18 (2016).
doi:10.3847/0004-637X/831/1/18
79. K. G. McCracken, H. Moraal, M. A. Shea, The high-energy impulsive ground-level
enhancement. Astrophys. J. 761, 101 (2012). doi:10.1088/0004-637X/761/2/101
80. A. Mishev, Short- and medium-term induced ionization in the Earth atmosphere by galactic
and solar cosmic rays. Int. J. Atmos. Sci. 2013, 184508 (2013). doi:10.1155/2013/184508
81. T. Sukhodolov, I. Usoskin, E. Rozanov, E. Asvestari, W. T. Ball, M. A. J. Curran, H.
Fischer, G. Kovaltsov, F. Miyake, T. Peter, C. Plummer, W. Schmutz, M. Severi, R.
Traversi, Atmospheric impacts of the strongest known solar particle storm of 775 AD.
Sci. Rep. 7, 45257 (2017). doi:10.1038/srep45257 Medline
82. F. Mekhaldi, R. Muscheler, F. Adolphi, A. Aldahan, J. Beer, J. R. McConnell, G. Possnert,
M. Sigl, A. Svensson, H.-A. Synal, K. C. Welten, T. E. Woodruff, Multiradionuclide
evidence for the solar origin of the cosmic-ray events of ᴀᴅ 774/5 and 993/4. Nat.
Commun. 6, 8611 (2015). doi:10.1038/ncomms9611 Medline
83. K. McCracken, J. Beer, F. Steinhilber, J. Abreu, A phenomenological study of the cosmic ray
variations over the past 9400 years, and their implications regarding solar activity and the
solar dynamo. Sol. Phys. 286, 609–627 (2013). doi:10.1007/s11207-013-0265-0
84. K. G. McCracken, F. B. McDonald, J. Beer, G. Raisbeck, F. Yiou, A phenomenological
study of the long-term cosmic ray modulation, 850–1958 AD. J. Geophys. Res. 109,
A12103 (2004). doi:10.1029/2004JA010685
85. K. G. McCracken, J. Beer, The annual cosmic-radiation intensities 1391 – 2014; The annual
heliospheric magnetic field strengths 1391 – 1983, and identification of solar cosmic-ray
events in the cosmogenic record 1800 – 1983. Sol. Phys. 290, 3051–3069 (2015).
doi:10.1007/s11207-015-0777-x
86. F. Miyake, A. J. T. Jull, I. P. Panyushkina, L. Wacker, M. Salzer, C. H. Baisan, T. Lange, R.
Cruz, K. Masuda, T. Nakamura, Large 14C excursion in 5480 BC indicates an abnormal
sun in the mid-Holocene. Proc. Natl. Acad. Sci. U.S.A. 114, 881–884 (2017).
doi:10.1073/pnas.1613144114 Medline
87. R. G. Harrison, M. Lockwood, Rapid indirect solar responses observed in the lower
atmosphere. Proc. R. Soc. London Ser. A 476, 20200164 (2020).
doi:10.1098/rspa.2020.0164
88. J. Kirkby, Cosmic rays and climate. Surv. Geophys. 28, 333–375 (2007).
doi:10.1007/s10712-008-9030-6
89. L. Zhang, B. Tinsley, L. Zhou, Low Latitude Lightning Activity Responses to Cosmic Ray
Forbush Decreases. Geophys. Res. Lett. 47, e2020GL087024 (2020).
doi:10.1029/2020GL087024
90. M. Sinnhuber, J. P. Burrows, M. P. Chipperfield, C. H. Jackman, M.-B. Kallenrode, K. F.
Künzi, M. Quack, A model study of the impact of magnetic field structure on
atmospheric composition during solar proton events. Geophys. Res. Lett. 30, 1818 (2003).
doi:10.1029/2003GL017265
91. IACETH, “Documentation for the SOCOL model” (Institute for Atmospheric and Climate
Science, ETH, 2009).
92. E. Roeckner, R. Brokopf, M. Esch, M. Giorgetta, S. Hagemann, L. Kornblueh, E. Manzini,
U. Schlese, U. Schulzweida, Sensitivity of simulated climate to horizontal and vertical
resolution in the ECHAM5 atmosphere model. J. Clim. 19, 3771–3791 (2006).
doi:10.1175/JCLI3824.1
93. E. Manzini, M. Giorgetta, M. Esch, L. Kornblueh, E. Roeckner, The influence of sea surface
temperatures on the northern winter stratosphere: Ensemble simulations with the
MAECHAM5 model. J. Clim. 19, 3863–3881 (2006). doi:10.1175/JCLI3826.1
94. E. Rozanov, M. E. Schlesinger, V. Zubov, The University of Illinois, Urbana-Champaign
three-dimensional stratosphere-troposphere general circulation model with interactive
ozone photochemistry: Fifteen-year control run climatology. J. Geophys. Res. D 106,
27233–27254 (2001). doi:10.1029/2000JD000058
95. T. Egorova, E. Rozanov, V. Zubov, I. Karol, Model for investigating ozone trends
(MEZON). Izv. Atmos. Ocean. Phys. 39, 277–292 (2003).
96. C. R. Hoyle, “Three dimensional chemical transport model study of ozone and related gases
1960-2000,” thesis, ETH Zurich (2005).
97. S. J. Marsland, H. Haak, J. H. Jungclaus, M. Latif, F. Röske, The Max-Planck-Institute
global ocean/sea ice model with orthogonal curvilinear coordinates. Ocean Model. 5, 91–
127 (2003). doi:10.1016/S1463-5003(02)00015-X
98. T. Egorova, E. Rozanov, V. Zubov, E. Manzini, W. Schmutz, T. Peter, Chemistry-climate
model SOCOL: A validation of the present-day climatology. Atmos. Chem. Phys. 5,
1557–1576 (2005). doi:10.5194/acp-5-1557-2005
99. M. Schraner, E. Rozanov, C. Schnadt Poberaj, P. Kenzelmann, A. M. Fischer, V. Zubov, B.
P. Luo, C. R. Hoyle, T. Egorova, S. Fueglistaler, S. Brönnimann, W. Schmutz, T. Peter,
Chemistry-climate model SOCOL: Version 2.0 with improved transport and
chemistry/microphysics schemes. Atmos. Chem. Phys. 8, 5957–5974 (2008).
doi:10.5194/acp-8-5957-2008
100. M. Calisto, I. Usoskin, E. Rozanov, T. Peter, Influence of galactic cosmic rays on
atmospheric composition and dynamics. Atmos. Chem. Phys. 11, 4547–4556 (2011).
doi:10.5194/acp-11-4547-2011
101. I. G. Usoskin, G. A. Kovaltsov, Cosmic ray induced ionization in the atmosphere: Full
modeling and practical applications. J. Geophys. Res. 111, D21206 (2006).
doi:10.1029/2006JD007150
102. I. G. Usoskin, G. A. Kovaltsov, I. A. Mironova, Cosmic ray induced ionization model
CRAC:CRII: An extension to the upper atmosphere. J. Geophys. Res. 115, D10302
(2010). doi:10.1029/2009JD013142
103. H. S. Porter, C. H. Jackman, A. E. S. Green, Efficiencies for production of atomic nitrogen
and oxygen by relativistic proton impact in air. J. Chem. Phys. 65, 154–167 (1976).
doi:10.1063/1.432812
104. S. Solomon, D. W. Rusch, J.-C. Gérard, G. C. Reid, P. J. Crutzen, The effect of particle
precipitation events on the neutral and ion chemistry of the middle atmosphere: II. Odd
hydrogen. Planet. Space Sci. 29, 885–893 (1981). doi:10.1016/0032-0633(81)90078-7
105. J. Kirkby, J. Curtius, J. Almeida, E. Dunne, J. Duplissy, S. Ehrhart, A. Franchin, S. Gagné,
L. Ickes, A. Kürten, A. Kupc, A. Metzger, F. Riccobono, L. Rondo, S. Schobesberger, G.
Tsagkogeorgas, D. Wimmer, A. Amorim, F. Bianchi, M. Breitenlechner, A. David, J.
Dommen, A. Downard, M. Ehn, R. C. Flagan, S. Haider, A. Hansel, D. Hauser, W. Jud,
H. Junninen, F. Kreissl, A. Kvashin, A. Laaksonen, K. Lehtipalo, J. Lima, E. R. Lovejoy,
V. Makhmutov, S. Mathot, J. Mikkilä, P. Minginette, S. Mogo, T. Nieminen, A. Onnela,
P. Pereira, T. Petäjä, R. Schnitzhofer, J. H. Seinfeld, M. Sipilä, Y. Stozhkov, F.
Stratmann, A. Tomé, J. Vanhanen, Y. Viisanen, A. Vrtala, P. E. Wagner, H. Walther, E.
Weingartner, H. Wex, P. M. Winkler, K. S. Carslaw, D. R. Worsnop, U. Baltensperger,
M. Kulmala, Role of sulphuric acid, ammonia and galactic cosmic rays in atmospheric
aerosol nucleation. Nature 476, 429–433 (2011). doi:10.1038/nature10343 Medline
106. S. Muthers, J. G. Anet, A. Stenke, C. C. Raible, E. Rozanov, S. Brönnimann, T. Peter, F. X.
Arfeuille, A. I. Shapiro, J. Beer, F. Steinhilber, Y. Brugnara, W. Schmutz, The coupled
atmosphere–chemistry–ocean model SOCOL-MPIOM. Geosci. Model Dev. 7, 2157–
2179 (2014). doi:10.5194/gmd-7-2157-2014
107. P. Arsenovic, E. Rozanov, J. Anet, A. Stenke, W. Schmutz, T. Peter, Implications of
potential future grand solar minimum for ozone layer and climate. Atmos. Chem. Phys.
18, 3469–3483 (2018). doi:10.5194/acp-18-3469-2018
108. N. Butchart, The Brewer-Dobson circulation. Rev. Geophys. 52, 157–184 (2014).
doi:10.1002/2013RG000448
109. Y. Kuroda, K. Yamazaki, K. Shibata, Role of ozone in the solar cycle modulation of the
North Atlantic Oscillation. J. Geophys. Res. 113, D14122 (2008).
doi:10.1029/2007JD009336
110. D. Budikova, Northern Hemisphere climate variability: Character, forcing mechanisms, and
significance of the North Atlantic/Arctic Oscillation. Geogr. Compass 6, 401–422 (2012).
doi:10.1111/j.1749-8198.2012.00498.x
111. D. W. Thompson, S. Lee, M. P. Baldwin, “Atmospheric processes governing the Northern
Hemisphere Annular Mode/North Atlantic Oscillation” in The North Atlantic Oscillation:
Climatic Significance and Environmental Impact, J. W. Hurrell, Y. Kushnir, G. Ottersen,
M. Visbeck, Eds., vol. 134 of Geophysical Monograph Series (American Geophysical
Union, 2003), pp. 81–112.
112. K. Kodera, Solar cycle modulation of the North Atlantic Oscillation: Implication in the
spatial structure of the NAO. Geophys. Res. Lett. 29, 59-1–59-4 (2002).
doi:10.1029/2001GL014557
113. PAGES2k Consortium, A global multiproxy database for temperature reconstructions of the
Common Era. Sci. Data 4, 170088 (2017). doi:10.1038/sdata.2017.88 Medline
114. PAGES 2k Consortium, Continental-scale temperature variability during the past two
millennia. Nat. Geosci. 6, 339–346 (2013). doi:10.1038/ngeo1797
115. G. A. Meehl, J. M. Arblaster, K. Matthes, F. Sassi, H. van Loon, Amplifying the Pacific
climate system response to a small 11-year solar cycle forcing. Science 325, 1114–1118
(2009). doi:10.1126/science.1172872 Medline
116. R. Checa-Garcia, M. I. Hegglin, D. Kinnison, D. A. Plummer, K. P. Shine, Historical
tropospheric and stratospheric ozone radiative forcing using the CMIP6 database.
Geophys. Res. Lett. 45, 3264–3273 (2018). doi:10.1002/2017GL076770
117. K. B. Tokarska, M. B. Stolpe, S. Sippel, E. M. Fischer, C. J. Smith, F. Lehner, R. Knutti,
Past warming trend constrains future warming in CMIP6 models. Sci. Adv. 6, eaaz9549
(2020). doi:10.1126/sciadv.aaz9549 Medline
118. C. Harrison, J. Prospero, Reversals of the Earth’s magnetic field and climatic changes.
Nature 250, 563–565 (1974). doi:10.1038/250563a0
119. R. Olson, W. Roberts, C. Zerefos, Short term relationships between solar flares,
geomagnetic storms and tropospheric vorticity patterns. Nature 257, 113–115 (1975).
doi:10.1038/257113a0
120. R. W. Fairbridge, Global climate change during the 13,500-b.p. Gothenburg geomagnetic
excursion. Nature 265, 430–431 (1977). doi:10.1038/265430a0
121. A. E. Carlson, L. Tarasov, T. Pico, Rapid Laurentide ice-sheet advance towards southern
last glacial maximum limit during marine isotope stage 3. Quat. Sci. Rev. 196, 118–123
(2018). doi:10.1016/j.quascirev.2018.07.039
122. R. P. Duncan, P. Fenwick, J. G. Palmer, M. S. McGlone, C. S. M. Turney, Non-uniform
interhemispheric temperature trends over the past 550 years. Clim. Dyn. 35, 1429–1438
(2010). doi:10.1007/s00382-010-0794-2
123. J. Stevenson, G. Hope, A comparison of late Quaternary forest changes in New Caledonia
and northeastern Australia. Quat. Res. 64, 372–383 (2005).
doi:10.1016/j.yqres.2005.08.011
124. L. Becerra-Valdivia, T. Higham, The timing and effect of the earliest human arrivals in
North America. Nature 584, 93–97 (2020). doi:10.1038/s41586-020-2491-6 Medline
125. J. Hansford, P. C. Wright, A. Rasoamiaramanana, V. R. Pérez, L. R. Godfrey, D. Errickson,
T. Thompson, S. T. Turvey, Early Holocene human presence in Madagascar evidenced
by exploitation of avian megafauna. Sci. Adv. 4, eaat6925 (2018).
doi:10.1126/sciadv.aat6925 Medline
126. R. Tobler, A. Rohrlach, J. Soubrier, P. Bover, B. Llamas, J. Tuke, N. Bean, A. Abdullah-
Highfold, S. Agius, A. O’Donoghue, I. O’Loughlin, P. Sutton, F. Zilio, K. Walshe, A. N.
Williams, C. S. M. Turney, M. Williams, S. M. Richards, R. J. Mitchell, E. Kowal, J. R.
Stephen, L. Williams, W. Haak, A. Cooper, Aboriginal mitogenomes reveal 50,000 years
of regionalism in Australia. Nature 544, 180–184 (2017). doi:10.1038/nature21416
Medline
127. F. Saltré, J. Chadoeuf, K. J. Peters, M. C. McDowell, T. Friedrich, A. Timmermann, S.
Ulm, C. J. A. Bradshaw, Climate-human interaction associated with southeast Australian
megafauna extinction patterns. Nat. Commun. 10, 5311 (2019). doi:10.1038/s41467-019-
13277-0 Medline
128. S. D. Mooney, S. P. Harrison, P. J. Bartlein, A.-L. Daniau, J. Stevenson, K. C. Brownlie, S.
Buckman, M. Cupper, J. Luly, M. Black, E. Colhoun, D. D’Costa, J. Dodson, S. Haberle,
G. S. Hope, P. Kershaw, C. Kenyon, M. McKenzie, N. Williams, Late Quaternary fire
 regimes of Australasia. Quat. Sci. Rev. 30, 28–46 (2011).
doi:10.1016/j.quascirev.2010.10.010
129. L. C. White, K. J. Mitchell, J. J. Austin, Ancient mitochondrial genomes reveal the
demographic history and phylogeography of the extinct, enigmatic thylacine (Thylacinus
cynocephalus). J. Biogeogr. 45, 1–13 (2018). doi:10.1111/jbi.13101
130. T. W. Horton, A. N. Zerbini, A. Andriolo, D. Danilewicz, F. Sucunza, Multi-Decadal
Humpback Whale Migratory Route Fidelity Despite Oceanographic and Geomagnetic
Change. Front. Mar. Sci. 7, 414 (2020). doi:10.3389/fmars.2020.00414
131. T. Higham, T. Compton, C. Stringer, R. Jacobi, B. Shapiro, E. Trinkaus, B. Chandler, F.
Gröning, C. Collins, S. Hillson, P. O’Higgins, C. FitzGerald, M. Fagan, The earliest
evidence for anatomically modern humans in northwestern Europe. Nature 479, 521–524
(2011). doi:10.1038/nature10484 Medline
132. T. Higham, K. Douka, R. Wood, C. B. Ramsey, F. Brock, L. Basell, M. Camps, A.
Arrizabalaga, J. Baena, C. Barroso-Ruíz, C. Bergman, C. Boitard, P. Boscato, M.
Caparrós, N. J. Conard, C. Draily, A. Froment, B. Galván, P. Gambassini, A. Garcia-
Moreno, S. Grimaldi, P. Haesaerts, B. Holt, M.-J. Iriarte-Chiapusso, A. Jelinek, J. F.
Jordá Pardo, J.-M. Maíllo-Fernández, A. Marom, J. Maroto, M. Menéndez, L. Metz, E.
Morin, A. Moroni, F. Negrino, E. Panagopoulou, M. Peresani, S. Pirson, M. de la Rasilla,
J. Riel-Salvatore, A. Ronchitelli, D. Santamaria, P. Semal, L. Slimak, J. Soler, N. Soler,
A. Villaluenga, R. Pinhasi, R. Jacobi, The timing and spatiotemporal patterning of
Neanderthal disappearance. Nature 512, 306–309 (2014). doi:10.1038/nature13621
Medline
133. C. Lalueza-Fox, H. Römpler, D. Caramelli, C. Stäubert, G. Catalano, D. Hughes, N.
Rohland, E. Pilli, L. Longo, S. Condemi, M. de la Rasilla, J. Fortea, A. Rosas, M.
Stoneking, T. Schöneberg, J. Bertranpetit, M. Hofreiter, A melanocortin 1 receptor allele
suggests varying pigmentation among Neanderthals. Science 318, 1453–1455 (2007).
doi:10.1126/science.1147417 Medline
134. J. A. Haws, M. M. Benedetti, S. Talamo, N. Bicho, J. Cascalheira, M. G. Ellis, M. M.
Carvalho, L. Friedl, T. Pereira, B. K. Zinsious, The early Aurignacian dispersal of
modern humans into westernmost Eurasia. Proc. Natl. Acad. Sci. U.S.A. 117, 25414–
25422 (2020). doi:10.1073/pnas.2016062117 Medline
135. M. Cortés-Sánchez, F. J. Jiménez-Espejo, M. D. Simón-Vallejo, C. Stringer, M. C. Lozano
Francisco, A. García-Alix, J. L. Vera Peláez, C. P. Odriozola, J. A. Riquelme-Cantal, R.
Parrilla Giráldez, A. Maestro González, N. Ohkouchi, A. Morales-Muñiz, An early
Aurignacian arrival in southwestern Europe. Nat. Ecol. Evol. 3, 207–212 (2019).
doi:10.1038/s41559-018-0753-6 Medline
136. J. I. Svendsen, H. P. Heggen, A. K. Hufthammer, J. Mangerud, P. Pavlov, W. Roebroeks,
Geo-archaeological investigations of Palaeolithic sites along the Ural Mountains – On the
northern presence of humans during the last Ice Age. Quat. Sci. Rev. 29, 3138–3156
(2010). doi:10.1016/j.quascirev.2010.06.043
137. M. C. Langley, S. O’Connor, “Exploring red ochre use in Timor-Leste and surrounds:
Headhunting, burials, and beads” in The Archaeology of Portable Art: Southeast Asian,
 Pacific, and Australian Perspectives, M. C. Langley, M. Litster, D. Wright, S. K. May,
Eds. (Routledge, 2018), pp. 25–36.
138. B. M. Chase, M. Chevalier, A. Boom, A. S. Carr, The dynamic relationship between
temperate and tropical circulation systems across South Africa since the last glacial
maximum. Quat. Sci. Rev. 174, 54–62 (2017). doi:10.1016/j.quascirev.2017.08.011
139. H. J. Deacon, Excavations at Boomplaas cave - a sequence through the Upper Pleistocene
and Holocene in South Africa. World Archaeol. 10, 241–257 (1979).
doi:10.1080/00438243.1979.9979735
140. J. Pargeter, E. Loftus, A. Mackay, P. Mitchell, B. Stewart, New ages from Boomplaas Cave,
South Africa, provide increased resolution on late/terminal Pleistocene human behavioral
variability. Azania 53, 156–184 (2018). doi:10.1080/0067270X.2018.1436740
141. J. T. Faith, R. L. Lyman, Paleozoology and Paleoenvironments: Fundamentals,
Assumptions, Techniques (Cambridge Univ. Press, 2019).
142. J. Sealy, J. Lee-Thorp, E. Loftus, J. T. Faith, C. W. Marean, Late Quaternary environmental
change in the Southern Cape, South Africa, from stable carbon and oxygen isotopes in
faunal tooth enamel from Boomplaas Cave. J. Quaternary Sci. 31, 919–927 (2016).
doi:10.1002/jqs.2916
143. A. S. Talma, J. C. Vogel, Late Quaternary palaeotemperatures derived from a speleothem
from Cango Caves, Cape Province, South Africa. Quat. Res. 37, 203–213 (1992).
doi:10.1016/0033-5894(92)90082-T
144. B. M. Chase, A. Boom, A. S. Carr, M. E. Meadows, P. J. Reimer, Holocene climate change
in southernmost South Africa: Rock hyrax middens record shifts in the southern
westerlies. Quat. Sci. Rev. 82, 199–205 (2013). doi:10.1016/j.quascirev.2013.10.018
145. F. d’Errico, L. Backwell, P. Villa, I. Degano, J. J. Lucejko, M. K. Bamford, T. F. G.
Higham, M. P. Colombini, P. B. Beaumont, Early evidence of San material culture
represented by organic artifacts from Border Cave, South Africa. Proc. Natl. Acad. Sci.
U.S.A. 109, 13214–13219 (2012). doi:10.1073/pnas.1204213109 Medline
146. M. Aubert, A. Brumm, M. Ramli, T. Sutikna, E. W. Saptomo, B. Hakim, M. J. Morwood,
G. D. van den Bergh, L. Kinsley, A. Dosseto, Pleistocene cave art from Sulawesi,
Indonesia. Nature 514, 223–227 (2014). doi:10.1038/nature13422 Medline
147. R. F. Rifkin, L. Dayet, A. Queffelec, B. Summers, M. Lategan, F. d’Errico, Evaluating the
photoprotective effects of ochre on human skin by in vivo SPF assessment: Implications
for human evolution, adaptation and dispersal. PLOS ONE 10, e0136090 (2015).
doi:10.1371/journal.pone.0136090 Medline
148. A. Svensson, K. K. Andersen, M. Bigler, H. B. Clausen, D. Dahl-Jensen, S. M. Davies, S. J.
Johnsen, R. Muscheler, F. Parrenin, S. O. Rasmussen, R. Röthlisberger, I. Seierstad, J. P.
Steffensen, B. M. Vinther, A 60 000 year Greenland stratigraphic ice core chronology.
Clim. Past 4, 47–57 (2008). doi:10.5194/cp-4-47-2008
149. D. E. Nelson, G. Chaloupka, C. Chippindale, M. S. Alderson, J. R. Southon, Radiocarbon
dates for beeswax figures in the prehistoric rock art of northern Australia. Archaeometry
37, 151–156 (1995). doi:10.1111/j.1475-4754.1995.tb00733.x
150. S. O’Connor, B. L. Fankhauser, in Histories of Old Ages: Essays in Honour of Rhys Jones
(Pandanus Books, 2001).
151. N. J. Conard, A female figurine from the basal Aurignacian of Hohle Fels Cave in
southwestern Germany. Nature 459, 248–252 (2009). doi:10.1038/nature07995 Medline
152. R. Vogelsang, Middle-Stone-Age-Fundstellen in Südwest-Namibia, vol. 11 (Heinrich-Barth-
Institut, 1998).
153. J. Clottes, B. Gély, C. Ghemis, E. Kaltinecker, V. T. Lascu, C. Moreau, M. Philippe, F.
Prud’Homme, H. Valladas, Un art très ancien en Roumanie. Les dates de Coliboaia.
International Newsletter on Rock Art 61, 1–3 (2011).
154. H. Floss, N. Rouquerol, in Éditions Musée-forum Aurignac, Aurignac (2007).
155. H. Valladas, Direct radiocarbon dating of prehistoric cave paintings by accelerator mass
spectrometry. Meas. Sci. Technol. 14, 1487–1492 (2003). doi:10.1088/0957-
0233/14/9/301
156. T. Jones, V. A. Levchenko, P. L. King, U. Troitzsch, D. Wesley, A. A. Williams, A.
Nayingull, Radiocarbon age constraints for a Pleistocene–Holocene transition rock art
style: The Northern Running Figures of the East Alligator River region, western Arnhem
Land, Australia. J. Archaeol. Sci. Rep. 11, 80–89 (2017).
doi:10.1016/j.jasrep.2016.11.016
157. J. Ross, K. Westaway, M. Travers, M. J. Morwood, J. Hayward, Into the Past: A Step
Towards a Robust Kimberley Rock Art Chronology. PLOS ONE 11, e0161726 (2016).
doi:10.1371/journal.pone.0161726 Medline
158. P. S. Tacon, T. Huisheng, M. Aubert, Naturalistic animals and hand stencils in the rock art
of Xinjiang Uyghur Autonomous Region, northwest China. Rock Art Res. 33, 19–31
(2016).
159. S. di Lernia, M. A. Tafuri, M. Gallinaro, F. Alhaique, M. Balasse, L. Cavorsi, P. D.
Fullagar, A. M. Mercuri, A. Monaco, A. Perego, A. Zerboni, Inside the “African cattle
complex”: Animal burials in the holocene central Sahara. PLOS ONE 8, e56879 (2013).
doi:10.1371/journal.pone.0056879 Medline
160. B. David, B. Barker, F. Petchey, J.-J. Delannoy, J.-M. Geneste, C. Rowe, M. Eccleston, L.
Lamb, R. Whear, A 28,000 year old excavated painted rock from Nawarla Gabarnmang,
northern Australia. J. Archaeol. Sci. 40, 2493–2501 (2013).
doi:10.1016/j.jas.2012.08.015
161. L. G. Straus, M. R. González Morales, The Magdalenian settlement of the Cantabrian
region (Northern Spain): The view from El Miron Cave. Quat. Int. 272–273, 111–124
(2012). doi:10.1016/j.quaint.2012.03.053
162. W. A. Neves, A. G. M. Araujo, D. V. Bernardo, R. Kipnis, J. K. Feathers, Rock art at the
pleistocene/holocene boundary in Eastern South America. PLOS ONE 7, e32228 (2012).
doi:10.1371/journal.pone.0032228 Medline
163. E. Malotki, H. D. Wallace, Columbian mammoth petroglyphs from the San Juan River near
Bluff, Utah, United States. Rock Art Res. 28, 143–152 (2011).
164. D. Huyge, D. A. G. Vandenberghe, M. De Dapper, F. Mees, W. Claes, J. C. Darnell, First
evidence of Pleistocene rock art in North Africa: Securing the age of the Qurta
petroglyphs (Egypt) through OSL dating. Antiquity 85, 1184–1193 (2011).
doi:10.1017/S0003598X00061998
165. S. O’Connor, K. Aplin, E. St Pierre, Y. Feng, Faces of the ancestors revealed: Discovery
and dating of a Pleistocene-age petroglyph in Lene Hara Cave, East Timor. Antiquity 84,
649–665 (2010). doi:10.1017/S0003598X00100146
166. M. Smith, A. Watchman, J. Ross, Direct dating indicates a Mid-Holocene age for archaic
rock engravings in arid Central Australia. Geoarchaeology 24, 191–203 (2009).
doi:10.1002/gea.20262
167. M. Dietzel, H. Kolmer, P. Pölt, S. Simic, Desert varnish and petroglyphs on sandstone–
geochemical composition and climate changes from Pleistocene to Holocene (Libya).
Geochemistry 68, 31–43 (2008). doi:10.1016/j.chemer.2007.03.001
168. A. W. Pike, M. Gilmour, P. Pettitt, R. Jacobi, S. Ripoll, P. Bahn, F. Muñoz, Verification of
the age of the Palaeolithic cave art at Creswell Crags, UK. J. Archaeol. Sci. 32, 1649–
1655 (2005). doi:10.1016/j.jas.2005.05.002
169. N. Cole, A. Watchman, AMS dating of rock art in the Laura Region, Cape York Peninsula,
Australia–protocols and results of recent research. Antiquity 79, 661–678 (2005).
doi:10.1017/S0003598X00114590
170. V. Plagnes, C. Causse, M. Fontugne, H. Valladas, J.-M. Chazine, L.-H. Fage, Cross dating
(Th/U-14 C) of calcite covering prehistoric paintings in Borneo. Quat. Res. 60, 172–179
(2003). doi:10.1016/S0033-5894(03)00064-4
171. R. G. Bednarik, About the age of Pilbara rock art. Anthropos 97, 201–215 (2002).
172. K. Steelman, M. Rowe, V. Shirokov, J. Southon, Radiocarbon dates for pictographs in
Ignatievskaya Cave, Russia: Holocene age for supposed Pleistocene fauna. Antiquity 76,
341–348 (2002). doi:10.1017/S0003598X00090426
173. D. S. Whitley, R. I. Dorn, Rock art chronology in eastern California. World Archaeol. 19,
150–164 (1987). doi:10.1080/00438243.1987.9980031
174. D. L. Hoffmann, C. D. Standish, M. García-Diez, P. B. Pettitt, J. A. Milton, J. Zilhão, J. J.
Alcolea-González, P. Cantalejo-Duarte, H. Collado, R. de Balbín, M. Lorblanchet, J.
Ramos-Muñoz, G.-C. Weniger, A. W. G. Pike, U-Th dating of carbonate crusts reveals
Neandertal origin of Iberian cave art. Science 359, 912–915 (2018).
doi:10.1126/science.aap7778 Medline
175. M. Aubert, R. Lebe, A. A. Oktaviana, M. Tang, B. Burhan, A. Hamrullah, A. Jusdi, B.
Abdullah, B. Hakim, J. X. Zhao, I. M. Geria, P. H. Sulistyarto, R. Sardi, A. Brumm,
Earliest hunting scene in prehistoric art. Nature 576, 442–445 (2019).
doi:10.1038/s41586-019-1806-y Medline