When does large woody debris influence ancient rivers? Dendrochronology applications in the Permian and Triassic, Antarctica

When does large woody debris influence ancient rivers? Dendrochronology applications in the Permian and Triassic, Antarctica

Journal Pre-proof When does large woody debris influence ancient rivers? Dendrochronology applications in the Permian and Triassic, Antarctica Erik L...

27MB Sizes 0 Downloads 1 Views

Journal Pre-proof When does large woody debris influence ancient rivers? Dendrochronology applications in the Permian and Triassic, Antarctica

Erik L. Gulbranson, Gianluca Cornamusini, Patricia Ryberg, Valentina Corti PII:

S0031-0182(19)30400-6

DOI:

https://doi.org/10.1016/j.palaeo.2019.109544

Reference:

PALAEO 109544

To appear in:

Palaeogeography, Palaeoclimatology, Palaeoecology

Received date:

23 April 2019

Revised date:

16 December 2019

Accepted date:

16 December 2019

Please cite this article as: E.L. Gulbranson, G. Cornamusini, P. Ryberg, et al., When does large woody debris influence ancient rivers? Dendrochronology applications in the Permian and Triassic, Antarctica, Palaeogeography, Palaeoclimatology, Palaeoecology (2019), https://doi.org/10.1016/j.palaeo.2019.109544

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

© 2019 Published by Elsevier.

Journal Pre-proof When does large woody debris influence ancient rivers? Dendrochronology applications in the Permian and Triassic, Antarctica

Erik L. Gulbranson1, Gianluca Cornamusini2, Patricia Ryberg3, Valentina Corti2

Department of Geology, Gustavus Adolphus College, 56082, USA

2

Dipartimento di Scienze Fisiche, della Terra e dell’Ambiente, Università di Siena, Italy

3

Department of Natural and Physical Sciences, Park University, 64152, USA

ro

of

1

re

-p

corresponding author: [email protected]

Abstract

lP

Accumulations of woody debris and in situ forests in and near ancient river

na

systems are studied in Permian through Triassic strata of Antarctica. These Permian and Triassic fossil ecosystems represent paleo-polar and paleo-high-latitude environments,

ur

respectively, and therefore represent unique ecosystem end-members for assessing the

Jo

role of woody debris in influencing fluvial deposition and paleoecologic processes. A new application of dendrochronology, the cross-dating of ring widths from individual trees, is applied to these fossil woods to develop a chronology of wood growth in the transported woody debris relative to in situ stumps. Cross-matching of woody debris relative to in situ stumps for the Permian indicates that woody debris deposition in river channels was contemporaneous with the development of vegetated macroforms in inchannel fluvial deposits. In contrast, Triassic woody debris occurs without evidence of vegetated in-channel macroforms, yet the volume and areal distribution of woody debris

1

Journal Pre-proof is greater than that measured in the Permian. Floral diversity increased dramatically across the Permian–Triassic transition in Antarctica, in addition to an increase in tree density in Triassic forests relative to Permian forests. However, these observations do not clearly explain the difference between the woody debris influence on Permian river systems relative to Triassic river systems. Paleoclimate simulations, however, indicate that the Permian polar environment was in surplus of water during the austral summer

of

relative to the Triassic, consistent with paleosol morphology, suggesting that

-p

ro

paleoclimate variation exerted a prominent control on vegetation-river dynamics.

re

1. Introduction

Woody debris in rivers acts as a fundamental biogeomorphologic control on

lP

fluvial style, forest ecology, and riparian ecosystems (Abbe and Montgomery, 1996;

na

Nakayama et al., 2002; Rygel et al., 2004; Pettit et al., 2006; Collins et al., 2012). The Paleozoic marks the first occurrence of large woody debris accumulations and potential

ur

sedimentary feedbacks (Gibling et al., 2010; Davies et al., 2011; Davies and Gibling,

Jo

2013; Gibling et al., 2014). The influence of woody debris on fluvial environments is likely profound, yet given contemporary anthropogenic disturbance, deep time records of these processes are likely best suited to understand the scope and importance of these processes (Fielding and Alexander, 2001; Rygel et al., 2004; Gastaldo and Degges, 2007; Gibling et al., 2010; Ielpi et al., 2014). Broad syntheses over the relationships of plant ecology and fluvial style are emerging (Davies and Gibling, 2013; Gibling et al., 2014). These syntheses illustrate that vegetation-fluvial interactions occur on a spectrum of potential contradictory responses, such as: bank stabilization from rooting (Tal and Paola,

2

Journal Pre-proof 2007); enhancing avulsion through woody debris deposition (Jones and Schumm, 1999); and nucleating and stabilizing in-channel macroforms (Fielding et al., 1997; Collins et al., 2012). Thus, the occurrence of woody debris in fluvial environments is not a diagnosis of one set of processes, rather these unique deposits provide a new metric to more deeply explore alluvial environments and shed-light on flood basin to channel interactions. This study focuses on the sedimentology of fossil woody debris and in situ forests in paleo

of

polar ecosystems of the Permian and Triassic, Antarctica. The paleo polar setting of this

ro

study area provides for a paleoecologic end-member of fluvial-vegetation interactions

-p

given that no modern analogue exists for this biome in the modern. A new application of

re

dendrochronology is used to provide relative dating of woody debris entrained in fluvial deposits in order to determine whether deposition of woody debris influenced the

lP

development of in situ forests on in-channel fluvial macroforms. The influence of the

na

end-Permian biotic crisis on vegetation-fluvial interactions is examined due to the floral turnover that occurred in the region (Lindström and McLoughlin, 2007) area as well as

Jo

ur

paleoclimate variation during the Permian–Triassic transition (Fielding et al., 2019).

2. Geological setting

The Allan Hills outcrop area in southern Victoria Land, Antarctica (SVL, Fig. 1A) hosts a Permian through Jurassic succession of sedimentary and volcano-sedimentary strata. From the base of the succession and upwards, the Allan Hills outcrop area includes: the Permian Weller Coal Measures (WCM); the Permian–Triassic Feather Conglomerate; the Triassic Lashly Fm.; and the Jurassic Mawson Fm., intruded and sealed by the Jurassic Ferrar Dolerite and Kirkpatrick Basalt, respectively. Palynomorphs

3

Journal Pre-proof of the Protohaploxypinus biozone and Praecolpatites sinuosus are recovered from the middle and upper part of the WCM (Kyle and Schopf, 1982; Farabee et al., 1990; Askin, 1995). The Protohaploxypinus biozone is constrained to the APL4–APP1 Australian biozones, with recent radiometric ages placing this palynomorph group in the latest Carboniferous to Early Permian (Smith and Mantle, 2013; Laurie et al., 2016). The presence of P. sinuosus, however, is constrained to the APP3 biozone of Early to Middle

of

Permian age (Laurie et al., 2016). The uppermost Weller Coal Measures at the Allan

ro

Hills contains pollen and spore assemblages suggestive of an Early–Middle Permian age,

-p

but despite the lack of Dulhuntyispora spp. index fossils, these assemblages were

re

interpreted as a Late Permian age (Awatar et al., 2014). The Feather Conglomerate has not yielded plant micro- or macrofossils, hindering more precise age determinations than

lP

by stratigraphic positioning. The potential for diachroneity of the dispersed record of

na

fossil plants should be considered, however, in the accuracy of these palynologic comparisons across broad sectors of Gondwana (Barbolini et al., 2016). A

ur

disconformable contact between the WCM and the overlying Feather Conglomerate has

Jo

been proposed (Isbell and Cúneo, 1996), but this contact may be locally disconformable, grading into a conformable succession of fluvial sandstones and floodplain paleosols (Retallack and Krull, 1999; Tewari et al., 2015). The Triassic Lashly Formation, lying above the Feather Conglomerate, contains palynomorphs of the Alisporites biozone, conferring an Early Triassic to earliest Middle Triassic age for the Lashly A Member, and a Middle Triassic age for the Lashly B Member (Kyle and Schopf, 1982). The WCM contains volcaniclithic sandstones, numerous coal seams, organic-rich siltstone, and shale. The uppermost sandstone succession of the WCM has been

4

Journal Pre-proof interpreted as either low-sinuosity streams (Isbell and Cúneo, 1996) or high-sinuosity streams with minor sediment contribution from glaciated, or seasonal ice (Smith et al., 1998), environments (Francis et al., 1993). The Lashly Fm. is divided into four members in order from stratigraphically lowest to highest: A, B, C, and D (Barrett and Webb, 1973). The Lashly A Member is generally formed by sandstone-mudstone interbeds, with well-developed paleosols.

of

Bioturbation and compression fossils of leaves and wood occur towards the top of the

ro

Lashly A Member. The Lashly B Member is defined as generally sand-rich, and contains

-p

numerous laterally continuous sandstones with fluctuating abundances of fossil wood

re

(Liberato et al., 2017). The Lashly C Member contains intergraded sandstones and siltstones of varying lateral continuity; occurrences of fossils and bioturbation fluctuate

lP

throughout this interval. Coal and carbonaceous siltstones occur in the Lashly C Member,

na

where they are absent or minimally expressed in the lower two members of the Lashly

ur

Formation (Liberato et al., 2017).

Jo

3. Systematic sedimentology and stratal architecture The Permian WCM and Triassic Lashly Formation of the Allan Hills, South Victoria Land, Antarctica preserve abundant fossil wood in fluvial sandstones (Ballance, 1977; Retallack, 1995). The Feather Conglomerate, which occurs between the Weller Coal Measures and Lashly Formation, does not preserve plant macrofossils. Below, the depositional framework is presented on the basis of various scales of sedimentary deposits: the style and ranking of stratigraphic bounding surfaces; macroforms; and mesoforms. This scheme follows closely the approach used in Ielpi et al. (2014).

5

Journal Pre-proof

3.1 Hierarchy of stratigraphic bounding surfaces The identification of stratigraphic bounding surfaces, sensu Ielpi et al. (2014), creates a systematic architectural framework of alluvial stratigraphy. A hierarchy of six bounding surfaces are routinely expressed in the Permian and Triassic stratigraphy of the Allan Hills based on: 1) the lateral extent of the surfaces; 2) the scale of sedimentary

of

features of the bounded stratigraphy; and 3) the relationship of the bounding surface to

ro

over- and underlying stratigraphy, being erosional or depositional. The first two bounding

-p

surfaces are descriptive in nature, whereas the remaining four bounding surfaces include

re

interpretation of the association of stratigraphic units.

First order bounding surfaces are recognized as internal surfaces within

lP

mesoforms such as the boundary between discrete sets of cross-bedding. These surfaces

na

exist on a scale less than the scale of a mesoform unit, roughly on the meter-scale. Second order surfaces reflect the boundary of mesoforms, and therefore reflect bedsets,

ur

or the boundary between accumulations of multiple laminae-sets within a mesoform unit.

Jo

Third order surfaces occur within macroforms and reflect relationships interpreted as erosion or accretion of units bound by second order surfaces. Fourth order surfaces reflect preserved boundaries of discrete macroform elements, and are broadly interpreted to reflect the boundaries of in-channel deposition of low-sinuosity systems or of preserved discrete macroform elements in high-sinuosity systems. Fifth order surfaces are prominent erosional basal contacts, or depositional contacts at the tops of macroforms, within a given facies association, demarcating significant erosional relief from adjacent overbank facies as well as floodplain channel facies (Heller and Paola, 1996). Fifth order

6

Journal Pre-proof surfaces, therefore, delineate discrete channel forms. Sixth order surfaces reflect pronounced juxtaposition of alluvial facies patterns and generally reflects the lithostratigraphic contacts between the WCM, Feather Conglomerate, and Lashly Formation.

3.2 In-channel mesoform and macroform deposits

of

3.2.1 Ribbon sandstones, Weller Coal Measures

ro

Clastic deposits of the WCM are sandstone-dominated and preserve a variety of

-p

internal architecture; and vertical and lateral associations of in-channel deposits or

re

overbank deposits (Table 1). Sandy bedforms facies (SB) comprise the entirety of inchannel deposits. The SB units are preserved within sandsheet geometries and exhibit

lP

vertical stacking predominantly by depositional third order surfaces (Fig. 2). Erosional

na

third order contacts reflecting nested channel forms are a minor component of the vertical stacking relationships. Individual sandsheet units are bound commonly by erosional fifth

ur

order lower contacts and preserve flat or gradually dipping depositional fifth order upper

Jo

contacts. Fifth order contacts commonly form truncating relationships with adjacent and underlying OF facies (Table 1; Fig. 2A). Rare fourth order surfaces preserving the tops of macroforms occur in individual sand sheets, typically when woody debris is present. In situ fossil trees occur on two distinct fourth order surfaces (Figs. 2B, 3). Bound by second order surfaces are ripple cross-stratified bedsets, which in many cases are planar or gently inclined in the upstream direction. Trough cross-bedding is locally abundant and only in association with woody debris accumulations. Ripple cross-bedding is common throughout the sand units, whereas climbing ripple cross-bedding is rare.

7

Journal Pre-proof Third order depositional surfaces that exhibit downstream accretion (DA) orientations, upstream accretion orientations (UA), or laterally accretion (LA) orientations are locally abundant in specific sandsheet units (Table 1). LA units are common near the base of sandsheet deposition or within erosional third order surfaces. The relationship of DA and UA units in woody debris accumulations is complex along the paleoflow direction. The grain size and composition of WCM clastic deposits range

of

from very fine-grained to coarse grained and are volcaniclithic in composition.

ro

Autochthonous clasts of permineralized peat, coal, and woody debris are common in

-p

basal lags of individual macroform elements (Fig. 3).

re

The WCM exhibits meter-scale channelization of clastic material, bound by fifth order surfaces. These surfaces demarcate in-channel macroforms from overbank deposits,

lP

the latter forming channels predominantly on third order bounding surfaces. Internal

na

structure of the channel forms reveals rare preserved macroform tops, fourth order surfaces; common erosional third order surfaces; and numerous depositional third order

ur

surfaces often with tangential lower contacts in the downstream or upstream direction.

Jo

Channel forms have a length to height ratio consistent with “ribbon-sand” morphology in addition to the presence of significant volume of overbank material that occurs with erosional contacts to the channel features (Davies and Gibling, 2011). The grain size of in-channel macroform deposits is between factor of two or one-order of magnitude greater than that of adjacent overbank deposits.

3.2.2 Interpretation

8

Journal Pre-proof Based on the fifth order bounding surfaces that separate in-channel deposits from OF facies and the proportion of in-channel deposits to OF facies the clastic deposits of the WCM are consistent with low-sinuosity fluvial systems with diversely vegetated inchannel macroforms and floodplain environments (see section 3.3). The internal structure of channels suggests variable sinuosity of individual channel bodies and prevalence of flood stage conditions capable of re-working macroforms in the channel system resulting

of

in complex DA and LA relationships. The vegetation styles (sections 3.3 and 3.4) are

ro

consistent with forested ecosystems and wetland ecosystems. The combination of overall

-p

low-sinuosity sandsheet deposition, higher sinuosity in-channel deposition, and

re

vegetation spread throughout the in-channel to floodplain environments indicates that the WCM alluvial environment at the Allan Hills was similar to extant anastomosing stream

na

lP

systems.

3.2.3 Braidplain alluvial deposits, Feather Conglomerate

ur

The Feather Conglomerate preserves GB and predominantly SP facies throughout

Jo

this sand-rich depositional unit (Table 1; Fig. 2A). A dramatic decrease of preserved OF facies and fossil evidence or bioturbation is a key distinction of this depositional unit from over- and underlying stratigraphy in the Allan Hills (Fig. 4A). Overall, clastic units within the Feather Conglomerate consist of fining upward cycles of sandsheet style geometries. Small meter-scale third order erosional surfaces demarcate nested channels within sandsheet units. Individual sandsheet units are bound by erosional fifth order surfaces. Fourth order surfaces are relatively more common within sandsheet depositional units and are gently dipping (Figs. 2B; 4A). Paleocurrent data was not collected for the

9

Journal Pre-proof Feather Conglomerate, therefore the relationship of macroform elements to paleoflow is not known. However, Barret and Fitzgerald (1985) describe paleocurrents of the Feather Conglomerate south of the study area as characteristic of a low-sinuosity fluvial system. Macroforms of the Feather Conglomerate sandsheets have internal stratification by trough cross-bedding. The grain size and composition is distinct from over- and underlying units, where the Feather Conglomerate is predominantly coarse grained and sub-

of

angular arkosic sandstone.

ro

The Feather Conglomerate reflects channelization with a notable decrease in

-p

preserved overbank deposits. Fifth order surfaces bound tens of meters of macroform

re

deposits. Fourth order macroform tops are relatively higher in abundance than in the WCM, and erosional third order surfaces are relatively lower in abundance than the

lP

WCM. Locally abundant third-order surfaces occur near the base of distinct macroform

ur

3.2.4 Interpretation

na

units.

Jo

The change in proportion of OF to in-channel deposits, lateral continuity of inchannel deposits, and distinct fining upward trends of coarse sandstones suggest a lowsinuosity fluvial environment for the Feather Conglomerate. Paleocurrent data and additional observations from Barret and Fitzgerald (1985) confirm continuity of this fluvial style in paleo-downstream settings to the south of the study area.

3.2.5 High-sinuosity to braidplain alluvial deposits, Lashly Formation

10

Journal Pre-proof The Lashly Formation contains vertical patterns in SB and GB sandstone geometries and abundance of OF facies and woody debris (Table 1). These patterns are delineated broadly by sixth-order erosional and depositional surfaces (Fig. 4A). Fifthorder surfaces highlight laterally discontinuous channel forms that are bound laterally and vertically by OF facies. The geometry of fifth-order surfaces is variable in the Lashly Fm. The Lashly A and C members contains more abundant lenticular fifth-order surface

of

geometries (Fig. 4A); whereas the Lashly B Member is predominantly composed of

ro

sandsheets with planar fifth-order surfaces (Fig. 4B). Fourth-order surface preservation is

-p

rare in the Lashly Formation and generally occurs with the presence of woody debris

re

accumulations or preserved in situ fossil forests. In the Lashly A and C members, depositional third-order surfaces express predominantly LA orientation (Figs. 4A&C).

lP

GB facies are abundant as basal lags of sandsheets containing quartzite clasts and

na

paleosol clasts. Nested erosional third order surfaces more common in the Lashly B Member (Fig. 4B), which are in-filled by SB elements. Trough cross-bedding compose

ur

the dominant sedimentary structures of sandstones in the Lashly Formation. Siltstones

Jo

and mudstones are more abundant in the stratigraphy of the Lashly C Member (Fig. 5), as well as accumulations of carbonaceous leaf compressions. The sandstone lithofacies of the Lashly Fm. represents a clear compositional distinction from the underlying Feather Conglomerate where the Lashly Fm. has a volcaniclithic composition (Liberato et al., 2017), consistent with time-equivalent stratigraphy across the Transantarctic Mountains (Collinson et al., 1994).

3.2.6 Interpretation

11

Journal Pre-proof The Lashly Formation clastic deposits form three patterns of depositional facies. The Lashly A Member contains an equal abundance of in-channel sandstones with lenticular channel geometry (Fig. 4), erosive basal contacts and depositional to soilformed upper contacts. This depositional pattern in the Lashly A Member is consistent with a high-sinuosity fluvial environment characterized by discrete channel locations bounded by OF and floodplain facies, with channel abandonment or avulsion processes

of

promoting channel migration across the alluvial plain. The development of complex soil

ro

profiles in OF and floodplain facies indicates a broader spectrum of landscape ages in the

-p

alluvial environment, which further supports the interpretation of high-sinuosity fluvial

re

channels entrenched in discrete sectors of the alluvial plain environment for a period of time. The Lashly B Member in-channel facies illustrate a prominent shift in fluvial style

lP

towards laterally continuous sandsheet deposition, the vertical stacking of sandsheet

na

units, and the occurrence of woody debris in sandsheet macroforms. These facies patterns are consistent with a lower sinuosity stream system, however, the lack of clear evidence

ur

of forested macroform elements or a diversity of floodplain ecosystems makes it unclear

Jo

as to whether this environment is more similar to a braidplain setting or an anastomosing setting. The Lashly C Member facies patterns describe laterally discontinuous lenticular to sandsheet in-channel deposits with an abundance of OF and floodplain facies. Preserved in situ fossil forests in floodplain facies indicates establishment of long-lived stable floodplains at certain periods of alluvial deposition. These observations are consistent with a variable stream morphology on the alluvial plain setting, with highsinuosity stream facies in discrete stratigraphic positions, and relatively lower sinuosity

12

Journal Pre-proof sandsheet geometries towards the lower and upper sixth-order bounding surfaces of the Lashly C Member.

3.2.7 Woody debris accumulations Fossil wood in the WCM is preserved as: 1) axes with angular edges and/or root flare; 2) stems containing portions of roots; and 3) upright stumps that display roots

of

extending and tapering into the surrounding organic-rich silt sandstone (Figs. 6&7).

ro

Lower in the succession, these woody axes co-occur with fragments of siltstone materials

-p

(OF) as basal lags and are generally rare. Higher in the WCM (Figs., 3&6), woody debris

re

occurs over a range of sizes from small, <20 cm diameter axes of unknown length, to large ~1 m diameter axes that are up to 9 m in length and have roots or root flare

lP

attached. These accumulations of woody debris occur in two discrete SB macroforms and

na

contain at least 55 fossil wood axes per macroform (Figs. 7&8). By percentage, large woody debris, which includes axes >20 cm in diameter and may or may not include root

ur

flare, comprise 65% of the wood axes and therefore characterizes this site as a Large

Jo

Woody Debris deposit (LWD).

In the Lashly A Member, fossil wood rarely occurs, but when it is found it is predominantly as solitary axes, or more rarely as in situ stumps of small diameter (<10 cm). In contrast, the Lashly B Member preserves an overwhelming abundance of fossil wood primarily as axes; stems with root flare (Fig. 8C&D); or root mantles of stumps. The size and abundance of these wood axes qualifies these deposits as LWD, which occurs in a limited interval of sandstone macroforms of the Lashly B Member (Fig. 7). The Lashly C Member preserves lower abundances of transported woody debris, but

13

Journal Pre-proof contains a locally abundant assemblage of in situ stumps in a single bedding plane of floodplain facies (Fig. 4F).

3.3 Overbank deposits 3.3.1 Permian overbank and floodplain deposits Outcrop exposures characterized as floodplain facies are plane parallel bedded

of

carbonaceous sandy siltstones (OF) (Fig. 9). These OF intervals are laterally continuous

ro

at the scale of several meters to 50 m. Numerous centimeter-scale third-order scour

-p

surfaces occur locally within OF facies. Additional observations of overbank fine facies

re

contemporaneous with channel deposition derive from autochthonous cobble-size clasts of peat and bioturbated siltstones. Bedded in-situ representations of these organic-rich OF

lP

facies occur above fifth-order bounding surfaces as laterally continuous beds. Thus,

na

organic paleosols (Histosols) of permineralized peat and coal deposits are characterized as contemporaneous flood basin units (Fig. 9B). The areal distribution of clastic and

ur

organic-rich OF facies suggests spatial variability in the occurrence of each OF

Jo

constituent at a given stratigraphic level. Carbonaceous silty sandstones occur with two styles of stratigraphic relationships to in-channel macroforms, either truncated by overlying SP deposits, or as capping units that rest conformably on top of finer grained SP facies (Figs. 2&4A). Histosols occur primarily as capping units to sandstone macroforms and are present as laterally continuous beds at the decimeter-scale.

3.3.2 Interpretation

14

Journal Pre-proof Clastic OF and floodplain facies characterize intermittent deposition during the highest flood stages of the WCM low-sinuosity fluvial system. Minor in-channel deposits within OF facies indicate episodes of breached levees and a transition from unconfined flow deposition characterizing most floodplain deposits to confined flow condition in discrete channels. Gently tapering and fining upward clastic units are interpreted as discrete crevasse splay (CS) deposition. These deposits suggest intermittent deposition in

of

areas where sediment deposition rates outpaced soil formation rates. Organic flood plain

ro

facies such as the permineralized peat and Histosol units indicate the occurrence of

-p

wetland ecosystems within the margins of the fluvial system. The occurrence of

re

autochthonous clasts of these organic units in channel deposits indicate reworking of the floodplain environment as the fluvial system evolved over time. In situ fossil stumps on

lP

macroform tops indicates the establishment of forested islands within channels of the

na

fluvial system, resulting in a complex mosaic of riparian ecologic communities.

ur

3.3.3 Feather Conglomerate overbank and floodplain deposits

Jo

Fine-grained laterally continuous units depicting a breakdown of primary sedimentary fabric and incipient soil horizonation occur at the sixth-order boundary between the WCM and Feather Conglomerate (Fig. 2A). These weakly developed paleosols persist into the lower sections of the Feather Conglomerate. Although, at the Allan Hills, fine-grained facies are rare in most of the Feather Conglomerate. Bioturbation, while also rare, occurs in the form of vertical burrows in very-fine grained silty sandstones (Fitzgerald and Barret, 1986).

15

Journal Pre-proof 3.3.4 Interpretation The laterally continuous but rare occurrence of weakly developed paleosols at the sixth order bounding surface between the WCM and Feather Conglomerate is suggestive of a change in soil-forming processes from the predominantly hydromorphic paleosols of WCM, to more well-drained mineral paleosols observed in the Triassic. The lack of soil structure in these paleosols could be related to climate, ecologic, or landscape changes.

of

This multifaceted interaction can be summarized as the soils reflecting: 1) a rate of

ro

sedimentation similar to the rate of soil-formation; or 2) lower productivity of soil micro-

-p

and macroflora and fauna. The specific cause(s) are unclear, but this dichotomous

re

framework would provide hypotheses capable of potentially differentiating more specific causes for the change in soil development observed in the Feather Conglomerate. The

lP

lack of abundance of these paleosols within the Feather Conglomerate suggests that the

na

low-sinuosity fluvial system of the Feather Conglomerate shifted laterally to the Allan

ur

Hills outcrop area and persisted in this region for a considerable timeframe.

Jo

3.3.5 Lashly Formation overbank and floodplain deposits The Lashly Formation contains a variety of fine-grained sedimentary facies characterized as overbank and floodplain deposits. Bioturbated siltstones occur in situ as meter-thick Protosols and as autochthonous clasts within gravel bar and bedform facies (GB). Sandstones are tabular plane-parallel to ripple cross-bedded with erosional lower contacts and flat upper contacts (OF). Siltstones contain plane parallel bedding with flat upper and lower contacts (OF). Laterally discontinuous tapering sandstones with ripple cross-bedding to mm-scale massive bedding occur throughout (CS) as discrete coarser-

16

Journal Pre-proof grained sandstone units. Carbonaceous siltstones containing leaf compressions occur in the Lashly A and C members. In the Lashly C Member carbonaceous siltstones also host 37 in situ stumps near Roscolyn Tor (Fig. 4F). Bioturbated Protosols (described in detail below) occur in situ as the tops of fining upward intervals of laminated and channelized overbank sediments. As autochthonous sediment, the bioturbated Protosols occur as lags within GB facies.

of

Tabular sandstone bodies (OF) separate in-channel depositional elements and the laterally

-p

fining upward grain observed in OF successions.

ro

discontinuous sandstones (CS) comprise the coarser-grained elements in the overall

re

Several well-developed paleosol horizons are observed in the Lashly A Member (described in detail below) at the transition of GB and SP deposits of laterally

lP

discontinuous channels to the OF plane parallel bedded deposits above. These paleosols

Jo

3.3.6 Interpretation

ur

accumulations.

na

preserve evidence of clay accumulation; and Fe-oxide and Mn-oxide mineral

The diversity of overbank and floodplain deposits throughout the Lashly Formation is suggestive of variations in landscape stability and/or landscape age. Crossbedded sandstones and siltstones, with third-order erosional surfaces delineating channels are interpreted as floodplain deposits in environments where the rate of sediment deposition outpaced the rate of soil formation. Overbank channels are interpreted as forming during high flood stage conditions in the floodplain environment following

17

Journal Pre-proof breaching of levees. Laterally discontinuous ripple cross-bedded sandstones are interpreted as crevasse splay deposition on the floodplain environment. Bioturbated Protosols reflect environments where sediment accumulation rates were more sporadic over time and/or the locus of sediment deposition shifted spatially from the area where the Protosols occur. A third interpretation of the bioturbated Protosols is that they may reflect distinct environmental conditions in the floodplain

of

environment supportive of burrowing arthropods.

ro

The well-developed clay-rich paleosols are interpreted as environments where the

-p

rate of soil formation greatly outpaced the rate of sediment deposition. The association of

re

these paleosols with laterally discontinuous GB and SP channels suggests soil development may have taken place after avulsion or channel abandonment.

lP

Carbonaceous and fossiliferous siltstones in the Lashly A Member are found in a laterally

na

discontinuous geometry bound by lenticular third-order surfaces, consistent with an

ur

abandoned-channel in-fill.

Jo

3.4 Paleoecology of woody debris 3.4.1 Permian paleoecology Based on the predominance of Glossopteris leaf compressions in carbonaceous siltstones of the Weller Coal Measures, the arborescent taxa is exclusively related to the glossopterids, whereas non-arborescent taxa is known to only include the Equistales (Tewari et al., 2015). While the true diversity of glossopterids in this locality is not known, regional Antarctic wood anatomy of the glossopterids reveals three distinct morphogenera (Decombeix et al., 2011). Glossopteris leaf “species” at the Allan Hills are

18

Journal Pre-proof estimated to be on the order of 20 or more (Tewari et al., 2015), along with <5 “species” of Gangamopteris and a single representative of the Surangephyllum morphogenera. The true diversity of glossopterids is a long-standing issue for late Paleozoic paleoecologic reconstructions (McLoughlin, 2011), with conservative viewpoints that there are far too many leaf morphogenera to account for the order of magnitude lower diversity of other plant organs related to the glossopterids (fructifications, wood). Moreover, on the basis of

of

carbon isotopic evidence, two functional groups of glossopterids are identified in the

ro

Permian of Antarctica that are either deciduous or evergreen (Gulbranson et al., 2012;

-p

2014). Analysis of leaf compressions in the Allan Hills suggest three generic level

re

putative occurrences of glossopterid reproductive axes (Retallack et al., 2009). Therefore, while the woody debris is most plausibly glossopterid in origin, the paleoecology of

lP

plants grouped within the glossopterids is unknown for the study area. Forest density of

na

glossopterid-dominated ecosystems ranges from 100–2500 trees ha-1, with co-variation of deciduous and evergreen glossopterid functional groups (Gulbranson et al., 2014).

ur

Root axes of the Vertebraria morphogenus are present throughout the Weller

Jo

Coal Measures either as in situ roots with vertical to subvertical orientation relative to bedding, attached to transported woody axes, as fragments within autochthonous peat fragments; or as root compression fossils within sandstone mesoforms. Vertebraria is taxonomically associated with the Glossopteris flora. Two of the classic structural styles of Vertebraria are present, so-called open-cylinder and closed-cylinder (Decombeix et al., 2009). These styles reflect the occurrence or absence of large pore spaces within the root axis. The association of these root morphologies with paleoecology is unclear as both rooting styles occur in the various facies which preserve these floral elements.

19

Journal Pre-proof

3.4.2 Triassic paleoecology Triassic fossil plants in the Allan Hills show a remarkable increase in diversity in comparison to the Permian. While not exclusive to the Allan Hills, the Lashly A Member arborescent taxa include 3 species of Dicroidium (Escapa et al., 2010). In the Lashly B Member a new discovery of a highly diverse floral assemblage was discovered by Dr.

of

Ignacio Escapa in 2014, a brief summary of some of the arborescent morphogenera is

ro

listed here: several species of Dicroidium, Ginkgoales, and conifers. The Lashly C

-p

Member contains six species of Dicroidium, and a Heidiphyllum conifer (Escapa et al.,

re

2010). Fossil wood at Roscolyn Tor are associated only with leaf mats of the coniferous Heidiphyllum leaf morphogenera. Forest density of Triassic fossil forests ranges from

lP

200–4000 trees ha-1.

na

Non-arborescent taxa are likewise more diverse in both function and morphology in the Triassic than the Permian. Thalloid fossils interpreted as freshwater macroalgae

ur

have been recovered from the Lashly A and C members (Bomfluer et al., 2009).

Jo

Neocalamites compressions occur as discrete axes with one or more nodes preserved in tabular plane parallel stratified siltstones (OF) facies in the Lashly A–B Members. Fern genera include Osmunda in the Lashly C Member. Taeniopteris has also been recovered from the Lashly C Member of the Allan Hills (Townrow, 1967).

4. Methods 4.1 Sedimentology and field techniques

20

Journal Pre-proof Permian fossil woods were mapped along two bedding plane surfaces in the Allan Hills outcrop area, near Trudge Valley, along a transect (Fig. 1A), where the transect start (76º 42.479’ S, 159º 42.286’ E) and end (76º 42.498’ S, 159º 42.513’ E) positions were measured with GPS. The area covered in this transect is 0.02 km2. Triassic fossil woods were similarly mapped along a broad sandstone bedding plane at the Feather Bay locality (Figs. 1A&7) over an area of 2.88 km2. The distance of trees were determined using a

of

Brunton™ compass for orientation, and by pacing out the distance between tree stumps

ro

from arbitrary sampling points along the transect. The orientation of the stumps was

D bedform surfaces throughout the transect.

-p

measured using a Brunton™ compass, and paleocurrent orientations were measured on 3-

re

The in situ fossil forest (76º 41.826 S, 159º 48.647’ E, 1901 masl) discovered at

lP

Roscolyn Tor, Lashly C Member, during the 2014-2015 austral summer season was

na

mapped using the point-center-quarter method (Cúneo et al., 1993; Gulbranson et al., 2012). The point-center-quarter method establishes arbitrarily defined sampling points

ur

along a transect and the distance of the nearest tree to a given sampling point is measured

Jo

both as horizontal distance and bearing from the sampling station. The diameter of each tree stump is also measured for computing the basal area of wood. Wood samples were collected for anatomical and geochemical study.

4.2 Dendrochronology Dendrochronology is a statistical comparison of tree growth rings, where the goal is to provide an absolute match of a growth rings produced in the same time interval between two or more trees. Whereas contemporary dendrochronology seeks to provide

21

Journal Pre-proof absolute chronologies with annual precision, the extreme antiquity of the fossils examined here means that fossil wood specimens can only be cross-dated, or not. Thus, cross-dating of ancient fossil wood can only provide a relative dating of wood growth. The potential advantage of this technique is that multiple decades or potentially centuries of continuous wood growth for a given study area can be reconstructed, providing unparalleled time-resolution to study paleoclimate or paleoecologic processes.

of

Fossil wood was prepared for dendrochronologic measurements of ring widths by

ro

first creating thin sections of a given sample to identify ring boundaries as the transition

-p

of latewood to earlywood cells, where latewood cells are identified as having thicker

re

tracheid cell walls and smaller diameter cell lumen than earlywood cells. The thin section determined ring boundaries were compared to the expression of these ring boundaries in

lP

hand sample in order to validate the ring width measurements to the anatomically defined

na

wood growth increment. This technique was done to avoid erroneous ring boundary determination in hand sample as weathering and secondary mineralization can create

ur

features that resemble wood growth rings, but have no connection to the anatomical

Jo

transition from latewood to earlywood (Garland et al., 2007). Ring width measurements were made at the University of Wisconsin-Milwaukee using a LINTAB™ linear table (Rinntech®, Heidelberg, Germany) with attached Leica microscope. Measurements were made at 1/1,000 mm resolution using TSAP-Win™ software. For a given sample, at least four radial transects were measured (see cross matching below). Specimens with more than 20 growth rings were selected for dendrochronologic analysis in order to produce statistically significant ring width series for a t-distribution.

22

Journal Pre-proof

4.2.1 Cross-matching fossil wood Cross-matching of the fossil wood specimens consists of two phases: 1) internal cross-matching of the four radial ring-width transects for an individual sample, and 2) cross-matching of different samples based on the average ring widths of internally-cross matched samples. The overall aim of these two phases of analysis is to statistically assess

of

the replication of trends in ring width variation within an individual sample for quality

ro

control of ring width data; and replication of ring width variation is assessed between

-p

multiple samples in order to produce a statistically significant cross match. Cross-

re

matching statistical analyses were performed using Past5™ software. We use the following statistical metrics to assess the quality and confidence of cross matches:

lP

correlation coefficient, gleichlaufigkeit (percent parallel covariation), and the Student’s t-

na

statistic following ring width normalization of Baillie and Pilcher (1973). After internal cross-matching was performed on an individual sample, and that the cross-match was

ur

successful, an average of the four radial transects was made and it is the average radial

Jo

ring width transect that was used to cross match different samples.

4.2.2 Detrending and analysis Growth related trends occur in all living trees. Given adequate preservation of growth rings in fossils, these growth-related trends should also be apparent in fossil wood material. Therefore, the average (of n=4 radial transects) ring width measurements for a given specimen were detrended, after cross-matching, using a smoothing spline, where each spline was created for the measured time series of a given specimen. The resulting

23

Journal Pre-proof mathematical expression of the spline is used to compute the Ring Width Index (RWI), which is a time series of the observed growth (actual ring width measurement) divided by the expected growth (computed from the spline equation). As an index, there are three possible outcomes: 1) an RWI value of 1 indicates the measured ring width is equivalent to the predicted ring width for that year; 2) an RWI value greater than 1 indicates that the measured ring width exceeded the predicted ring width for that year; and 3) an RWI

of

value less than indicates that the measured ring width was less than predicted for that

ro

year. RWI is a useful technique in evaluating growth trends over multiple individuals,

-p

given the potential for state factor changes in the paleo landscape that could limit or

re

enhance plant growth, which would result in some trees showing wider/narrower rings than others due to differing growth rates. The ring width measurements and RWI are

lP

recorded for each studied specimen and are presented as the master chronology, based on

na

the successful cross matches, for the Permian specimens from the Weller Coal Measures, and for the Triassic specimens from the Lashly B Member. The RWI curves are used in

5. Results

Jo

ur

subsequent analysis and interpretation.

5.1 Paleosols of the Permian-Triassic strata Sedimentary rocks meeting the criteria for paleosols: 1) relict or lack of sedimentary structures; 2) development of soil structure; 3) evidence of colonization by plants; 4) accumulation of soil-formed material (e.g., organic matter, minerals: clay films, Fe-oxides); 5) reddening or color variation consistent with redoxymorphy (Vepraskas, 1994); and 6) horizonation were described in the Weller Coal Measures (WCM), Feather

24

Journal Pre-proof Conglomerate, and Lashly Formation overbank fine facies (OF). We describe: O horizons based on the abundance of organic material in a soil horizon (having less than 20% noncementing mineral material); B horizons, or subsoil horizons, based on a horizon being composed on mineral material and satisfying one or more of the above criteria; and C horizons, or parent material, as reflecting some alteration of the primary rock or sediment, but lacking sufficient alteration to satisfy the criteria for a B horizon.

of

Appropriate designators for B and C horizons are included to denote specific types of

ro

mineral accumulations and/or redox conditions experienced during soil formation.

-p

The WCM contains a predominance of paleosols rich with organic matter. These

re

paleosols display layering of organic materials with either distinct anatomical preservation or lacking anatomical preservation. The organic matter-rich paleosols may

lP

either be preserved as coal or as a silica-cemented rock with abundant plant fragments

na

and less than 20% non-cementing mineral material (Fig. 4A). The field observations for these paleosols meet the taxonomic criteria of the Histosols (Mack et al., 1993).

ur

In situ tree stumps found in sandstone macroforms indicate that these macroform

Jo

tops represent a distinct paleosol type, where soil structure is crude or absent, horizonation is absent or comprised of a very thin silt- to clay-rich layer overlying sandstone (Fig. 6C&D), sedimentary structures are absent or weakly defined, and there is a lack of soil formed minerals. The colonization of these sediments by plants is the primary distinguishing criteria of these rocks as paleosols (Fig. 4C). Based on these observations these paleosols meet the taxonomic criteria of the Protosols (Mack et al., 1993). The disruption of primary sedimentary fabric and colonization of these paleosols by plants indicates that these subsoil horizons are Bw horizons.

25

Journal Pre-proof In contrast, the Feather Conglomerate has a lack of organic matter-rich paleosols, and instead displays two prominent paleosol types. The first paleosol type is a massive to bioturbated green muddy siltstone with crude angular blocky structure. This paleosol meets the field descriptions of the Dolores pedotype of Retallack and Krull (1999), and similarly this paleosol type is observed to be abundant in the Feather Conglomerate. As defined by Retallack and Krull (1999) the Dolores pedotype meets the taxonomic criteria

of

of the Protosols (Mack et al., 1993). The Protosol classification of Retallack and Krull

ro

(1999) is based upon the observation that these paleosols display evidence of soil

-p

formation in the form of: 1) destruction of primary sedimentary features; 2) the initiation

re

of horizonation; 3) the development of berthierine nodules that are interpreted to have been soil-formed (Retallack and Krull, 1999), although berthierine is not typically a

lP

stable mineral phase in soil environments and is more commonly formed in metamorphic

na

environments (Iijima and Matsumoto, 1982); and 4) the lack of additional criteria that meets higher-level taxonomy.

ur

A second paleosol type was discovered near Dennes Point at the stratigraphic

Jo

contact of the Feather Conglomerate and the Lashly A Formation (Fig. 10A). This paleosol type is comprised of mineral soil material and is broken down into 5 horizons. The lowest horizon, Cg2, is a green massive sandstone with conspicuous black manganese concretions that occlude pore spaces between quartz grains (Fig. 10B). The second horizon, Cg1, alternates between green and maroon, contains burrows and relict bedding, minor manganese concretions occur along strike in a lenticular sandstone body that intergrades to the horizon below (Fig. 10C). The third horizon, Btg, is dominantly maroon with minor green redoximorphic spots and contains abundant clay and clay films

26

Journal Pre-proof on the surfaces of angular blocky peds (Fig. 10D). The fourth horizon, Btv, is predominantly maroon with minor green redoximorphic spots, contains abundant clay, clay films on angular blocky peds, and tubiform nodules of iron oxides (Fig. 10E). The fifth and uppermost horizon, Bt, is maroon and contains abundant clay, clay films, and aggregates of iron oxide nodules and displays platy structure (Fig. 10F). This second paleosol type is laterally discontinuous along strike, intergrading over an, along strike,

of

distance of 10m with a lenticular sandstone at its base. This paleosol meets multiple

ro

taxonomic criteria resulting in a refined taxonomic identification, using the Mack et al.

-p

(1993) scheme, as a gleyed ferritic Argillisol. The most important observation of this

re

paleosol is the presence of soil-formed clay (Bt horizons) evidenced by clay films in horizons 4 and 5, giving the Argillisol classification. The occurrence of iron oxide

lP

nodules in horizon 4 (Btv horizons) indicates that pedogenic iron accumulated in this soil,

na

in which the term ferritic is used. And the occurrence of redoximorphic features (Btg and Cg1 horizons) as well as accumulations of redox-sensitive Mn as Mn-oxide minerals in

ur

horizon 1 (Cg2 horizon) signifies that gleying, or alternating reducing and oxidizing

Jo

conditions, were important on this landscape. Paleosols recognized at higher stratigraphic positions in the Lashly Fm. display weak soil structure development and horizonation. These paleosols are primarily silty to silty sandstone in texture, contain crude angular blocky structure, overlie sandstones that display massive texture, and contain abundant trace fossils of either burrows and/or roots. In situ tree stumps found in sandstones indicate that these sandstones represent a distinct paleosol type, where soil structure is crude or absent, sedimentary structures are absent, yet there is a lack of soil formed minerals. The colonization of these sediments by plants

27

Journal Pre-proof is the primary distinguishing criteria of these rocks as paleosols. These paleosols meet the taxonomic criteria of the Protosols (Mack et al., 1993). The disruption of primary sedimentary fabric, development of soil structure, and evidence of plant or subsurface faunal activity indicate the presence of Bw horizons, and without additional subsoil horizons present, this results in the Protosol classification.

of

5.2 Field relationships of fossil wood

ro

The spatial location and orientation of Permian LWD deposited in two sandstone

-p

beds of the Weller Coal Measures were mapped at the Trudge Valley locality in the Allan

re

Hills (Fig. 1). Fifty-five fossil wood axes were mapped, and of these axes nine are found as in situ stumps. The in situ stumps are rooted into sandstone or a thin veneer of siltstone

lP

overlying the sandstone, with roots extending laterally by ~1 m, and have a mean ring

na

count of 93 rings. All of the in situ stumps occur in the same sandstone bedding plane. The mean orientation of the long axis of LWD on a bedding plane is 202º  79º (Fig. 7A).

ur

The mean paleocurrent orientation from ripple cross-bedding is 323º and those of dune

Jo

cross-bedding of 82º, thus a subset of the mapped population of fossil wood is oriented subparallel to paleoflow directions of ripple cross-bedding in the sandstones and nearly orthogonal to 3-D dune cross-bedding in the sandstones. Three horizontal LWD axes exhibited orientation with the root flare portion of the trunk pointing downstream, relative to the paleocurrent orientation. These three LWD axes occur directly upstream of three of the nine in situ stumps found along this bedding plane. Triassic fossil wood in the Lashly B Member (Figs. 7B; 8C&D) at the Feather Bay (S76º 40.524’, E159º 52.203’, 1753 masl), while originally an extensive deposit of

28

Journal Pre-proof fossils, has been vigorously mined for souvenirs (Bradshaw, 1978). Thus, the present-day exposure of these fossils likely reflects a combination of the geologic preservation bias with subsequent anthropogenic disturbance marring the relation of the fossils to their original depositional and ecologic nature. In contrast to the LWD in the Lashly B Member an in situ fossil forest discovered during the 2014-2015 austral summer season near the summit of Roscolyn Tor, Lashly C Member, was mapped and described using

of

the point-center quarter method. However, the trees in this fossil forest display too few

-p

ro

rings (n<20) to provide reliable dendrochronologic information.

re

5.3 Roscolyn Tor forest structure

Thirty-seven fossil tree stumps in upright positions with attached roots were

lP

discovered along a single sandstone bedding plane below the summit of the western side

na

of Roscolyn Tor, within the Lashly C Member. Stumps range in diameter from 2.5 cm to 32 cm, and displayed a maximum of about 20 potential growth increments in the largest

ur

diameter stumps. Horizontal and sub-horizontal oriented wood axes are interspersed

Jo

along the sandstone bedding plane containing the upright stumps. Roots of the trees stumps extend into a gray siltstone, displaying massive texture and relict sedimentary bedding. Fine (1 mm) to coarse (>3 mm) root haloes permeate the siltstone. Rooting is more vertically oriented than with the Permian in situ stumps. The tree density as measured by the point-center-quarter method is 420 trees ha-1, and the basal area of wood for the observed area of the fossil forest is 12 m2 ha-1, which is consistent with early succession forests in modern boreal ecosystems (Viereck et al., 1993). Heidiphyllum leaf compressions are observed in the gray siltstone that underlies the sandstone bedding

29

Journal Pre-proof plane as well as in the overlying sandstone. The overlying sandstone displays LA elements with dune-scale trough cross-bedding at the base of the sandstone and ripplescale trough cross-bedding at the top of the sandstone. The overlying sandstone is lenticular and is laterally discontinuous, with conformable upper contacts with siltstone, with an exposed length along strike of 30m.

of

5.4 Dendrochronology of Permian and Triassic fossil wood

ro

Permian wood samples did not result in successful cross matches for every

-p

specimen. Of the 11 specimens studied, six did not yield successful cross matches

re

between specimens. The five specimens that did yield successful cross matches result in a master chronology of 86 successive growth increments (Fig. 11A). The ring width

lP

variation in this master chronology is supported by an averaging of between two or four

na

specimens per growth increment, indicating the extent of replication of these ring width signals in the population of fossil wood. Statistics on cross matches are reported in Table

ur

1. An ensemble of Gleilaughikeit (Gl) (percent parallel co-variation in ring widths

Jo

between two samples), the correlation coefficient (cc), and t-statistics (Tho=t-statistic with a Hollstein method of ring width normalization, TBP=t-statistic with the Baillie/Pilcher ring width normalization, after Baillie and Pilcher, 1973) are used to assess quality of the cross-match (Table 2). For Permian samples Gl ranged from 57.7 to 82.1 (Gl has values from 0 to 100, where 100 is perfect covariation in ring width trends), correlation coefficient (cc) ranged from 0.06 to 0.96, Tho ranged from 4.1 to 6.4, and TBP ranged from 3.9 to 4.3 (Table 2). Maximum overlap between the five specimens was for 43 growth increments. The specimen with the most overlapping growth increments

30

Journal Pre-proof was one of the in situ stumps (AH-47; Fig. 11A), which therefore provides a means to assess the relation of transported wood to trees growing in the riparian environment represented by the Weller Coal Measures. Ring widths varied from 0.3 mm to 4.8 mm for 1,518 ring measurements of the 11 specimens. Mean sensitivity ranges from 0.23 to 0.49, averaging 0.37  0.09 for all specimens, where mean sensitivity is regarded as a metric to generally assess plant stress

of

(Fritts, 1976). Mean sensitivity values >0.3 indicate generally stressed growth conditions,

ro

and mean sensitivity <0.3 indicate relatively stable growth conditions.

-p

The ring width index for Permian specimens represents the average of cross-

re

matched RWIs for each specimen, reflecting the replication of RWI signals between multiple individual specimens. This chronology depicts 86 growth increments, or years if

lP

the tree rings represent annual periods of wood growth. The RWI chronology for the

na

Permian reveals a periodicity of ~40 growth increments between RWI intervals less than 1, where growth is below the expected growth based on the smoothing spline fitted to

ur

each specimen (Fig. 11A). The frequency of growth increments where the RWI is greater

Jo

than 1 is 52. Moreover, the cross matched average RWI curve can be compared to the RWI curve for an individual specimen revealing that some specimens have different trends in RWI for segments of the cross matched interval that they represent. Triassic fossil wood from Feather Bay (Lashly B Member) yielded successful cross-matches for all seven of the studied specimens (Fig. 11B). The resulting crossmatched chronology represents 238 growth increments, representing averages of two to five specimens per growth increment. For Triassic specimens Gl ranged from 62.5 to 93.8, cc ranged from -0.13 to 0.84, Tho ranged from 1.31 to 5.2, and TBP ranged from

31

Journal Pre-proof 1.1 to 4.1 (Table 2). Maximum overlap between the eight specimens was for 78 growth increments, ranging between eight and 78 overlapping growth increments. While no evidence for in situ stumps was found in these specimens, the immense range of growth increments produced by this cross-match allows for the analysis of climatically and ecologically relevant growth trends during the Middle Triassic. Ring widths varied from 0.8mm to 7.2 mm for 1,342 measured ring widths. Mean sensitivity ranges from 0.34 to

of

0.67, with a mean of 0.50  0.12.

ro

The Triassic ring width index represents 238 successive growth increments, or

-p

years if the growth increments represent annual periods of wood growth. The average

re

RWI curve indicates a periodicity of 23 growth increments between RWI values less than one, approximately half of the periodicity observed for the Permian (Fig. 11B). The

lP

frequency of growth increments with an RWI greater than 1 is 137, more than double the

na

frequency of positive RWIs observed in the Permian specimens (Fig. 11B). Moreover, most of the individual RWI curves display concordant trends compared to the average

ur

RWI curve (Fig. 11B), where a single specimen shows out of phase RWI trends for a

Jo

segment of the cross matched interval.

6. Discussion 6.1 Fluvial architecture through time The occurrence of large woody debris does not co-vary with fluvial style. The low-sinuosity fluvial systems of the WCM preserve an abundance of LWD in association with in situ trees. The low-sinuosity braidplain facies of the Feather Conglomerate contain no evidence of LWD, similar to the high-sinuosity facies of the Lashly A

32

Journal Pre-proof Member. The low-sinuosity braidplain facies of the Lashly B Member documents the reoccurrence of LWD in the Mesozoic following the demise of glossopterid ecosystems. Whereas the high-sinuosity facies of the Lashly C formation shows a decreases in the occurrence of LWD. While these patterns do not suggest that one fluvial style is associated with or causally linked to LWD, it does illustrate an important time series of changes in fluvial style and the occurrence of fossil plant remains. Time equivalent

of

successions in the Karoo Basin indicate a change in fluvial style concomitant with

ro

vertebrate fossil turnover (Smith, 1995). A tectonic control on this change in fluvial style

-p

is inferred from provenance analysis (Smith, 1995), however, emerging results suggest

re

that this regional tectonism likely affected terrestrial biota in the Karoo Basin via aridification (Rey et al., 2018). A sequence, nearly identical to the Antarctic stratigraphy

lP

studied herein, of a low sinuosity fluvial system in Permian coal measures transitioning

na

abruptly to low sinuosity braidplain fluvial facies, lacking plant macrofossils, occurs in: 1) the Prince Charles Mountains, East Antarctica (Lindström and McLoughlin, 2007);

ur

and 2) the Bowen Basin, eastern Australia (Fielding et al., 1995; Michaelsen, 2002). In

Jo

the Prince Charles Mountains, roughly 1/3 of the Permian spore and pollen assemblages are lost between the uppermost Permian coal and the lowest Triassic pollen-yielding sample (Lindström and McLoughlin, 2007), suggesting a co-variation of floral assemblage and sedimentation style. The Bowen Basin provenance during this transition is not suggestive of a change in tectonic regime, and this facies transition was interpreted to reflect a reorganization of fluvial regime following the collapse of high-latitude terrestrial ecosystems (Michaelsen, 2002). The Sydney Basin, further to the south in eastern Australia, however, shows no clear shift in fluvial style or sandstone abundance

33

Journal Pre-proof that co-varies with observed changes in plant macro- and microfossils (Fielding et al., 2019). Low paleolatitude basins display similar alterations of fluvial style and vegetation regime surrounding the transition from the Permian to the Triassic (Arche and LópezGómez, 2005; Diéguez et al., 2007). Changes in sediment provenance can be ruled out for this Permian–Triassic succession in South Victoria Land as invariant temporal trends in elements that are

of

conserved during chemical weathering processes (Ti, Al, U, Th, La, Ce) indicate a

ro

common provenance for Permian and Triassic sandstones (Sheldon et al., 2014). Epsilon

-p

Nd values, a robust isotopic tracer for sediment provenance, from the coeval

re

Transantarctic Basin (south of the study area), however, suggest a shift in sediment source coinciding with the uplift of a fold-and-thrust belt during the Late Permian (Elliot

lP

and Fanning, 2008). Therefore, on a lithostratigraphic and sediment provenance basis the

na

occurrence of LWD and fluvial style described herein is most likely related to the extinction and re-emergence of vegetation in these alluvial environments.

ur

The stratigraphic pattern of LWD and fluvial style during the Permian–Triassic

Jo

transition overlaps in coarser resolution with emerging data supported by high-precision radiometric ages that the glossopterid ecosystems of Gondwana collapsed prior to the end-Permian biotic crisis (Mays et al., 2019; Vajda et al., 2020). Based on high-precision U-Pb ages (Metcalfe et al., 2015), the demise of glossopterids in eastern Australia occurred ~410 ky prior to end-Permian biotic crisis (Fielding et al., 2019). The available time control for these Paleozoic and Mesozoic strata of Antarctica, updated to current Eastern Australian biozones (Smith and Mantle, 2013), indicate that the WCM most likely represents an Early Permian, perhaps up to a Late Permian, time interval. The

34

Journal Pre-proof Lashly A Member palynology indicates a more consistent early Middle Triassic age, resulting in the Feather Conglomerate representing an unknown time interval that falls between the Permian and Middle Triassic. Uncertainties in these age assessments are likely profound given the broad time range of indicator pollen such as P. haploxypinus as well as the potential for diachroneity of the dispersed record of fossil plants (Barbolini et al., 2016). Thus, future efforts towards radiometric age determination for this stratigraphy

of

would be vital towards better understanding how these polar environments and

-p

ro

ecosystems changed during the end-Permian biotic crisis.

re

6.2 Taphonomy

The sedimentologic interpretations and occurrence of macrofossils are interpreted

lP

here at face value to reflect time intervals where processes that contribute LWD were

na

active and time intervals where such processes were absent or outcompeted by processes inconsistent with the formation of LWD. Taphonomic bias is a distinct possibility that

ur

could confound these interpretations (Gastaldo et al., 2005). So-called “isotaphonomic”

Jo

assemblages are suggested to be used to assess the validity of stratigraphic interpretations. While such isotaphonomic assemblages exist between the study area in South Victoria Land, Antarctica and the Bowen Basin of eastern Australia; time equivalent sedimentary facies with differing taphonomy are present in the Feather Conglomerate. In the study area, this depositional unit lacks abundant OF facies and reflects a vertical accumulation of third order surfaces within discrete sheet-sandstone units, and the amalgamation of those sheet-sandstones, separated by fifth-order surfaces. The taphonomic potential of this depositional environment is interpreted to be minimal,

35

Journal Pre-proof and indeed there is a dearth of macrofossil evidence found in the Feather Conglomerate in the study area. Down-dip to the south, the Feather Conglomerate facies association changes to include greater thicknesses of parallel bedded siltstones muddy siltstones, representing more abundant OF facies preservation (Fitzgerald and Barrett, 1986). The OF facies here are interpreted as having greater taphonomic potential, and yield Skolithos trace fossils, however, there is a lack of plant macrofossils recovered from these units.

of

Time equivalent higher-resolution dispersed plant microfossil assemblages in the Prince

ro

Charles Mountains indicate floral turnover, separated by a gap in plant microflora

-p

preservation. The loss of Late Permian flora occurred in a stepwise fashion, with roughly

re

1/3 of the Late Permian flora lost per sampling step into Early Triassic stratigraphy (Lindström and McLoughlin, 2007). The taphonomic trends, of deposits with similar or

lP

different taphonomic potential, yield similar trends as is inferred from this study area

na

alone. Therefore, it is likely that the change in LWD occurrence and change in plant

Jo

and space.

ur

macroflora in the study area reflect real changes in terrestrial ecosystems through time

6.3 Large woody debris accumulations The large woody debris (LWD) accumulations are preserved primarily within inchannel facies of the Permian WCM and Triassic Lashly B Member. The LWD facies is separated by a gap of fossiliferous deposits in the Feather Conglomerate, and fossiliferous but non-LWD fluvial facies of the Lashly A Member. The chronologies of tree rings in LWD and the fluvial facies are analyzed here to better understand the paleo biogeomorphologic processes that may have contributed to LWD in these environments.

36

Journal Pre-proof

6.3.1 Dendrochronology for relative dating of wood deposition The relative age of the in situ stumps of the WCM, having an average of 93 growth increments, indicates that these macroform tops persisted as emerged/stabilized surfaces for nearly a century or longer. However, the timing of wood transport, forest growth, and the sedimentary build-up of bars is unclear from field relationships alone.

of

Dendrochronology, applied to trees that are disjointed from continuous chronologies

ro

because of extreme age or transport, has been successful in identifying the provenance

-p

and relative timing of woody debris deposition (Giddings, 1941; 1952; Eggertsson, 1994;

re

Eggertsson and Lavendecker, 1995). The cross-matching of fossil wood in the WCM can thus act to elucidate the timing of woody debris deposition when transported wood

lP

chronologies are referenced against an in situ tree that represents a fixed point in space.

na

Of the five samples that successfully cross-matched, one of them is in situ (AH-31), and this stump will be used as a reference point for the four additional cross-matched LWD

ur

(Fig. 12). Two of the cross-matched LWD samples (AH-47 and AH-35) occur as LWD

Jo

downstream of the paleoflow direction of the in situ tree. Sample AH-47 has a nearly complete overlap in cross-matched tree rings, but may be relatively older than the in situ tree, and occurs proximal to upstream accretion surfaces of an adjacent macroform (Figs. 7A and 12). Sample AH-35 was deposited farther upstream from this macroform, and displays less overlap of tree rings with the in situ tree, potentially suggesting that this LWD sample is relatively younger than the in situ tree. The remaining two cross-matched LWD (AH-27 and AH-29) occur in a densely spaced accumulation of woody debris upstream of the in situ tree and display a similar

37

Journal Pre-proof pattern of deposition as the downstream LWD. Sample AH-29 has a complete overlap in cross matched rings to the in situ stump and this sample was deposited upstream of the bar form where the in situ stump is located (Figs. 7A and 12). Sample AH-27 has fewer overlapping rings than the in situ stump and therefore may be relatively younger, and this wood fragment was deposited upstream of AH-29. The relative dating of LWD to in situ stumps suggest that macroforms in this

of

fluvial system are influenced by the deposition of woody debris, where the trees closest

ro

in relative age to the reference point occur on the fringes of the exposed bar surfaces.

-p

Subsequent woody debris deposition occurs adjacent to the previously deposited woody

re

debris, younging upstream. The relative ages of the LWD samples suggest that the deposited trees were locally sourced from this paleoecologic environment, which

lP

establishes a cycle of process. We interpret this cycle to involve: 1) the emergence of

na

macroforms in the river system, enhanced by the variable flow conditions; 2) colonization of these bars by trees and development of forests; 3) erosion, death, and

ur

subsequent transport of eroded or dead trees into the river system; 4) deposition of the

Jo

trees during waning flow conditions adjacent to down-stream macroform tops; and 5) sediment trapping and bar propagation during waning flow stages following enhanced bed movement during high flow stages. Six of the eleven LWD samples mentioned herein do not cross match with the in situ stump. The lack of cross-match likely indicates that these transported trees died and were transported prior to, or postdating, the growth of some of the in situ forests, rather than representing errors in ring width measurement or cross-matching (Brown et al.,

38

Journal Pre-proof 2014). Therefore, these results suggest that such a cycle of wood transport, deposition, and bar nucleation was long-lived for this fluvial system. The Triassic dendrochronology on transported trees of the Feather Bay site, which produces a chronology of 238 growth years, yields important insights that the crossmatched trees were not deposited in a chronologic fashion relative to sedimentary strata, as in the Permian. Thus, the establishment of stabilized in-channel macroforms was not

of

likely not a dominant process during the deposition of the Lashly B Member in the

ro

Middle Triassic. Floodplain forests, however, must have been a dominant source of LWD

-p

to the Triassic alluvial environment given the propensity of LWD in the Lashly B

re

Member and immense overlapping chronologies of tree rings in the LWD.

lP

6.4.1 Active channels or abandoned channels

na

Undercutting of banks likely contributes to the supply of woody debris to the Permian and Triassic fluvial systems. This connection is derived from observation of re-

ur

worked paleosol components–peat in the Permian, and mineral paleosols in the Triassic–

Jo

in LWD accumulations. However, the establishment of forests through incision and isolation of floodplain elements does not appear to be a dominant process for the Permian fluvial environments. Coeval OF facies throughout the Permian stratigraphy provide clear evidence of the soil landscape adjacent to the active fluvial environments. The in situ forest developed on a fluvial macroform is not developed on these OF paleosols, nor is it feasible to have been further altered by soil-forming processes as the in situ forest paleosol has an immature morphologic development, whereas OF paleosols are predominantly organic Histosols.

39

Journal Pre-proof

6.4.2 Channel abandonment The forest density at Roscolyn Tor in the Lashly C Member strongly suggests an early succession forest environment on point bars that prograde into a shallow meandering stream. These inferences are based on tree density as compared to Permian fossil forests in the central Transantarctic Mountains and comparison to modern boreal

of

forest succession patterns (Viereck et al., 1993; Gulbranson et al., 2012), as well as the

ro

laterally prograding fluvial sands that buried the Roscolyn Tor forest. This fossil forest

-p

also lacks sedimentary features associated with variable current direction and bedload

re

transport described by Fielding et al. (1997) or the unique forest growth patterns from

na

6.5 Paleoclimate implications

lP

Fielding and Alexander (2001).

Paleoclimate for the polar regions during the Permian and Triassic has been

ur

elusive to quantify with conventional paleoclimate proxies, such as soil-formed minerals.

Jo

Global paleoclimate simulations in particular struggle to create agreement with modeled biomes versus the fossil record in the paleo polar latitudes (Rees et al., 2002; Kiehl and Shields, 2005). A new regional paleoclimate model, including monthly normals for precipitation and temperatures, for the high- to polar-latitudes (Fielding et al., 2019) is used to investigate the paleo-moisture regime of the polar environment during the Late Permian (Fig. 13). Following the seasonality approach of Gulbranson et al. (2011), the model results for monthly temperature at 12.7X pre-industrial CO2 at 65ºS were adjusted for a temperature lapse rate with latitude of 0.63K/ºlatitude for the southern hemisphere

40

Journal Pre-proof to 80ºS and used to calculate evapotranspiration (Etp) using the Thornthwaite equation (Thornthwaite, 1948) and net primary productivity using the relationship of Leith (1975). The precipitation and Etp curves for the Late Permian indicate an overabundance of water during the polar summer season (Dec.–Feb.), with modeled precipitation vastly outpacing Etp for each summer month in addition to a likely surplus of water from melted snow/ice that accumulated during the polar winter. These moisture input/output

of

conditions are highly consistent with interpretations of hydromorphic soils (e.g.,

ro

Histosols) as well as the weakly developed nature of Protosols of the in-channel forested

-p

ecosystems of the WCM. The sandy nature of the Protosols and the relatively thin

re

profiles of these soils would likely be inundated with water during the non-freezing months of the polar summer. The uncertainties in this modeled approach are due to

lP

extrapolating precipitation conditions from 65ºS latitude across the isolating polar

na

cyclones, thus the true pattern of seasonal rainfall could differ from the modeled results at lower latitudes. Secondly, the temperature lapse rate most likely underestimates polar

ur

summer temperature conditions during the 24 h illuminated photoperiod that lasts

Jo

between 4–6 months depending on latitude (Berge et al., 2015). However, despite these uncertainties, it is likely that the overall trend of greater rainfall than Etp for the polar summer condition holds true given the low incident angle of solar radiation at the polar latitudes. The paleoclimate model of Fielding et al. (2019) results in Etp and precipitation seasonality in the expected range from observed soil morphologies in the Triassic Lashly Fm., showing Etp nearly equivalent with precipitation during one month of the austral summer and a lengthened growing season. As a stratigraphic bounding condition for

41

Journal Pre-proof assessing these modeled results, the presence of gleyed ferritic Argillisols in the Lashly Fm. necessitates that two hydrologic conditions be met: 1) a static water table that does not intersect the land surface; and 2) seasonal fluctuation in groundwater to promote the movement of redox sensitive Fe and Mn ions into and out of the vadose. Nearby field areas result in similar reductions in accumulations of redox sensitive elements in Triassic paleosols (Sheldon et al., 2014), underscoring that well-drained soil conditions were

of

likely more ubiquitous in the Triassic than the Permian (Fig. 14). Given these conditions,

ro

it would be expected that there would be a period during the austral summer of utilization

-p

of soil water, which would explain a persistent pattern of water table fluctuation.

re

Moreover, an enhanced monthly rate of Etp would likely contribute to less available water as compared to the Permian polar climate estimation provided here. An increase in

lP

monthly summer temperatures would be consistent in creating these scenarios, therefore

na

it is likely that the Triassic climate in Antarctica was significantly warmer than the Late Permian. Given the equatorward migration of the study area away from the Polar Circle

ur

by the Middle Triassic (Scotese, 2014) it is likely that the study area achieved such a

Jo

seasonal temperature increase. The Lashly Fm., however, is a relatively thick accumulation of strata (Fig. 5), and based on the change in fluvial facies and plant diversity in the Lashly Fm. it is likely that paleoclimate also varied during the time interval recorded by these strata. If the balance of precipitation versus evapotranspiration shifted during this time interval, it would be expected that fluvial processes and ecologic stresses would change in kind, which would be consistent with the observation of LWD occurring in one narrow stratigraphic interval in the Lashly B Member, rather than throughout the entire Lashly Fm. (Fig. 14).

42

Journal Pre-proof

7. Conclusions The Permian through Triassic strata of the Allan Hills, Antarctica records major reorganizations in fluvial style. The Middle Permian is characterized by low-sinuosity river systems with vegetated islands and floodplain environments, with a distinct surplus of available water based on comparison of paleosol morphology and paleoclimate

of

simulations. The Permian–Triassic transition records a shift to low-sinuosity braidplain

ro

environments with a lack of plant macrofossil evidence, suggesting a severe change in

-p

terrestrial ecology in the Late Permian. The Middle Triassic is marked by high-sinuosity

re

fluvial environments with sparsely vegetated floodplains, increasing in floral diversity upsection. An alternation between low-sinuosity braidplain to high-sinuosity fluvial

lP

environments occurs at the top of the succession. The occurrence of Large Woody Debris

na

(LWD) does not correlate with a specific fluvial facies, rather LWD is correlated with the density of forested ecosystems and the specific style of fluvial style, such that low-

ur

sinuosity systems without vegetation lack LWD; and high-sinuosity fluvial systems for

Jo

all observed vegetation densities also lack LWD. The diversity of plants in these ecosystems does not have a correlation with the occurrence of LWD or influence on fluvial style. Paleoclimate simulations compared to paleosol morphologies in the study area, however, suggest that the duration of the growing season and amount of available water as indicated by the balance of precipitation versus evapotranspiration exerted the largest impact on both fluvial style as well as vegetation and likely acted to correlate the interaction of vegetation with fluvial bedforms under specific paleoclimate scenarios.

43

Journal Pre-proof Acknowledgements: We thank reviewers Drs. Martin Gibling and Chris Fielding for their time and highly constructive comments. We would also like to thank the 2012 field team members: Drs. Anne-Laure Decombeix, Benjamin Bomfluer, Rudy Serbet, and John Isbell; and Drew Brown, Dave Buchanan, Rebekah Davis, Anna Zajicek, the staff of the Berg Field Center, PHI, and Ken Borek Air Ltd. for invaluable field support, unwavering positivity,

of

sound judgement, and for being a part of what made these field teams work so well. This

ro

research was funded in part by the National Science Foundation [grants OPP 1142495 to

re

-p

PER, and OPP 1443557 and OPP 1142749 to ELG].

References

lP

Abbe, T.B., Montgomery, D.R., 1996. Large woody debris jams, channel hydraulics and

ur

201-221.

na

habitat formation in large rivers. Regulated Rivers: Research and Management, v. 12, p.

Jo

Arche, A., López-Gómez, J., 2005. Sudden changes in fluvial style across the Permian– Triassic boundary in the eastern Iberian Ranges, Spain: Analysis of possible causes. Palaeogeography, Palaeoclimatology, Palaeoecology, v. 229, p. 104–126.

Askin, R.A., 1995. Permian palynomorphs from southern Victoria Land, Antarctica. Antarctic Journal of the U.S., v. 30, p. 47-48.

44

Journal Pre-proof Awatar, R., Tewari, R., Agnihotri, D., Chatterjee, S., Pillai, S.S.K., Meena, K.L., 2014. Later Permian and Triassic palynomorphs from the Allan Hills, central Transantarctic Mountains, South Victoria Land, Antarctica. Current Science, v. 106, p. 988-996.

Ballance, P.F., 1977. The Beacon Supergroup in the Allan Hills, central Victoria Land,

of

Antarctica. New Zealand Journal of Geology and Geophysics, v. 20, no. 6, p. 1003-1116.

ro

Baillie, M.G.L., Pilcher, J.R., 1973. A simple cross-dating program for tree-ring research.

re

-p

Tree-Ring Bulletin, v. 33, p. 7-14.

Barbolini, N., Bamford, M.K., Rubidge, B., 2016. Radiometric dating demonstrates that

na

Research, v. 37, p. 241–251.

lP

Permian spore-pollen zones of Australia and South Africa are diachronous. Gondwana

ur

Berge, J., Renaud, P.E., Darnis, G., Cottier, F., Last, K., Gabrielsen, T.M., Johnsen, G.,

Jo

Seuthe, L., Weslawski, J.M., Leu, E., Moline, M., Nahrgang, J., Søreide, J., Varpe, Ø., Lønne, O.J., Daase, M., Falk-Petersen, S., 2015. In the dark: A review of ecosystem processes during the Arctic polar night. Progress in Oceanography, v. 139, p. 258–271.

Bomfleur, B., Krings, M., Kastovsky, M., Kerp, H., 2009. An enigmatic non-marine thalloid organism from the Triassic of East Antarctica. Review of Paleobotany and Palynology, v. 157, p. 317–325.

45

Journal Pre-proof Bradshaw, M.A., 1978. Has exploitation begun in Antarctica? Newsletter of the Geological Society of New Zealand, v. 45, p. 7-9.

Braudrick, C.A., Grant, G.E., 2000. When do logs move in rivers? Water Resources Research, v. 36, p. 571-583.

of

Brown, P.M., Nash, S.E., Kline, D., 2014. Identification and dendrochronology of wood

ro

found at the Ziegler Reservoir fossil site, Colorado, USA. Quaternary Research, v. 82, p.

re

-p

575-579.

Collins, B.D., Montgomery, D.R., Fetherston, K.L., Abbe, T.B., 2012. The floodplain

lP

large-wood cycle hypothesis: a mechanism for the physical and biotic structuring of

na

temperate forested alluvial valleys in the North Pacific coastal ecoregion.

ur

Geomorphology, v. 139-140, p. 460-470.

Jo

Collinson, J.W., Isbell, J.L., Elliot, D.H., Miller, M.F., Miller, J.M., Veevers, J.J., 1994. Permian–Triassic Transantarctic Basin. In Permian–Triassic Pangean Basins and Foldbelts Along the Panthalassan Margin of Gondwanaland: Geological Society of America Memoir, 184, p. 173-222.

Cúneo, N.R., Isbell, J., Taylor, E.L., Taylor, T.N., 1993. The Glossopteris flora from Antarctica: Taphonomy and paleoecology. In: 12 Congrés International de Géologie du Carbonifére-Permian, Buenos Aires, Compte Rendu, v. 2, Buenos Aires, p. 13-40.

46

Journal Pre-proof

Davies, N.S., Gibling, M.R., 2011. Evolution of fixed-channel alluvial plains in response to Carboniferous vegetation. Nature Geoscience, v. 4, p. 629–633.

Davies, N.S., Gibling, M.R., 2013. The sedimentary record of Carboniferous rivers:

ro

ecosystems. Earth-Science Reviews, v. 120, p. 40–79.

of

Continuing influence of land plant evolution on alluvial processes and Palaeozoic

-p

Decombeix, A.-L., Taylor, E.L., Taylor, T.N., 2009. Secondary growth in Vertebraria

re

roots from the Late Permian of Antarctica: a change in developmental timing.

lP

International Journal of Plant Sciences, v. 170, p. 644-656.

na

Decombeix, A.-L., Taylor, E.L., Taylor, T.N., 2012. Gymnosperm trees from the Permian of Antarctica: An anatomically preserved trunk of Kaokoxylon sp. Comptes Rendus

Jo

ur

Palevol., v. 11, p. 21-29.

Diéguez, C., de la Horra, R., López-Gómez, J., Benito, M.I., Barrenechea, J., Arche, A., Luque, J., 2007. Late Permian plant remains in the SE Iberian Ranges, Spain: Biodiversity and palaeovegetational significance. Comptes Rendus Palevol, v. 6, p. 403– 411.

Eggertsson, O., 1994. Mackenzie River driftwood – a dendrochronological study. Arctic, v. 47, p. 128-236.

47

Journal Pre-proof

Eggertsson, O., Layendecker, D., 1995. Dendrochronological study of the origin of driftwood in Frobisher Bay, Baffin Island, NWT, Canada. Arctic Alpine Research, v. 27, p. 180-186.

Elliot, D.H., Fanning, C.M., 2008. Detrital zircons from upper Permian and lower

of

Triassic Victoria Group sandstones, Shackleton Glacier region, Antarctica: Evidence for

ro

multiple sources along the Gondwana plate margin. Gondwana Research, v. 13, p. 259–

re

-p

274.

Escapa, I.H., Taylor, E.L., Cúneo, R.N., Bomfleur, B., Bergene, J., Serbet, R., Taylor,

lP

T.N., 2011. Triassic floras of Antarctica: plant diversity and distribution in high

na

paleolatitude communities. Palaios, v. 26, p. 522-544.

ur

Farabee, M.J., Taylor, E.L., and Taylor, T.N., 1990. Correlation of Permian and Triassic

Jo

palynomorph assemblages from the central Transantarctic Mountains, Antarctica. Review of Palaeobotany and Palynology, v. 65, p. 257-265.

Fielding, C.R., Alexander, J., Newman-Sutherland, E., 1997. Preservation of in situ, arborescent and fluvial construction in the Burdekin River of north Queensland, Australia. Palaeogeography, Palaeoclimatology, Palaeoecology, v. 135, p. 123-144.

48

Journal Pre-proof Fielding, C.R., Alexander, J., 2001. Fossil trees in ancient fluvial channel deposits: evidence of seasonal and longer-term climatic variability. Palaeogeography, Palaeoclimatology, Palaeoecology, v. 170, p. 59-80.

Fielding, C.R., Falkner, A.J., Scott, S.G., 1993. Fluvial response to foreland basin overfilling; the Late Permian Rangal Coal Measures in the Bowen Basin, Queensland,

ro

of

Australia. Sedimentology, v. 85, p. 475–497.

-p

Fielding, C.R., Frank, T.D., McLoughlin, S., Vajda, V., Mays, C., Tevyaw, A.P.,

re

Winguth, A., Winguth, C., Nicoll, R.S., Bocking, M., Crowley, J.L., 2019. Age and pattern of the southern high-latitude continental end-Permian extinction constrained by

na

lP

multiproxy analysis. Nature Communications v. 10, p. 385

Fitzgerald, P.G., Barrett, P.J., 1986. Skolithos in a Permian braided river deposit, southern

Jo

237–247.

ur

Victoria Land, Antarctica. Palaeogeography, Palaeoclimatology, Palaeoecology, v. 52, p.

Francis, J.E., Woolfe, K.J., Arnot, M.J., Barrett, P.J., 1993. Permian forests of Allan Hills, Antarctica: the palaeoclimate of Gondwanan high latitudes. Special Papers in Palaeontology, no. 49, p. 75-83.

Fritts, H.C., 1976. Tree Rings and Climate. Academic Press, London.

49

Journal Pre-proof Garland, M.J., Bannister, J.M., Lee, D.E., White, J.D.L., 2007. A coniferous tree stump of late Early Jurassic age from the Ferrar Basalt, Coombs Hills, southern Victoria Land, Antarctica. New Zealand Journal of Geology and Geophysics, v. 50, p. 263-269.

Gastaldo, R.A., Degges, C.W., 2007. Sedimentology and paleontology of a

of

Carboniferous logjam. International Journal of Coal Geology, v. 69 (1–2), p. 104-118.

ro

Gastaldo, R.A., Adendorff, R., Bamford, M., Labandeira, C.C., Neveling, J., Sims, H.,

-p

2005. Taphonomic trends macrofloral assemblages across the Permian–Triassic

re

Boundary, Karoo Basin, South Africa. Palaios, v. 20, p. 479–497.

lP

Gibling, M.R., Bashforth, A.R., Falcon-Lang, H.J., Allen, J.P., Fielding, C.R., 2010. Log

na

jams and flood sediment buildup caused channel abandonment and avulsion in the

ur

Pennsylvanian of Atlantic Canada. Journal of Sedimentary Research, 80 (3), 268-287.

Jo

Gibling, M.R., Davies, N.S., Falcon-Lang, H.J., Bashforth, A.R., DiMichele, W.A., Rygel, M.C., Ielpi, A., 2014. Palaeozoic co-evolution of rivers and vegetation: a synthesis of current knowledge. Proceedings of the Geologist’s Association, v. 125, p. 524–533.

Giddings, J.L., 1941. Dendrochronology in northern Alaska. University of Alaska Bulletin IV, 107 pp.

50

Journal Pre-proof Giddings, J.L., 1952. Driftwood patterns and problems of the Arctic Sea currents. Am. Philos. Soc., v. 96, p. 129-142.

Gulbranson, E.L., Montañez, I.P., Tabor, N.J., 2011. A proxy for humidity and floral province from paleosols. The Journal of Geology, v. 119, p. 559–573.

of

Gulbranson, E.L., Isbell, J.L., Taylor, E.L., Ryberg, P.E., Taylor, T.N., Flaig, P.P., 2012.

ro

Permian polar forests: deciduousness and environmental variation. Geobiology, v. 10, p.

re

-p

479-495.

Gulbranson, E.L., Ryberg, P.E., Decombeix, A.-L., Taylor, E.L., Taylor, T.N., Isbell,

lP

J.L., 2014. Leaf habit of Late Permian Glossopteris trees from high-palaeolatitude forests.

na

Journal of the Geological Society, v. 171, p. 493-507.

ur

Heller, P.C., Paola, C., 1996. Downstream changes in alluvial architecture: an exploration

306.

Jo

of controls on channel stacking patterns. Journal of Sedimentary Research, v. 66, p. 297–

Hupp, C.R., Simon, A., 1991. Bank accretion and the development of vegetated depositional surfaces along modified alluvial channels. Geomorphology, v. 4, p. 111-124.

Hyatt, T.L., Naiman, R.J., 2001. The residence time of large woody debris in the Queets River, Washington, USA. Ecological Applications, v. 11, p. 191-202.

51

Journal Pre-proof

Ielpi, A., Gibling, M.R., Bashforth, A.R., Lally, C., Rygel, M.C., Al-Silwadi, S., 2014. Role of vegetation in shaping Early Pennsylvanian braided rivers: Architecture of the Boss Point Formation, Atlantic Canada. Sedimentology, v. 61, p. 1659–1700.

Iijima, A., Matsumoto, R., 1982. Berthierine and chamosite in coal measures of Japan.

ro

of

Clays and Clay Minerals, v. 30, no. 4, p. 264-274.

-p

Isbell, J.L., Cúneo, N.R., 1996. Depositional framework of Permian coal-bearing strata,

re

southern Victoria Land Antarctica.Palaeogeography, Palaeoclimatology, Palaeoecology,

lP

v. 125, p. 217-238.

na

Jansen, J.D., Nanson, G.C., 2010. Functional relationships between vegetation, channel morphology, and flow efficiency in an alluvial (anabranching) river. Journal of

Jo

ur

Geophysical Research, v. 115, F04030, doi:10.1029/2010JF001657.

Jones, L.S., Schumm, S.A., 1999. Causes of avulsion: an overview. International Association of Sedimentology Special Publication 28, p. 171–178.

Kiehl, J.T., Shields, C.A., 2005. Climate simulation of the latest Permian: Implications for mass extinction. Geology, v. 33, p. 757–760.

52

Journal Pre-proof Kyle, R.A., Schopf, J.M., 1982. Permian and Triassic palynostratigraphy of the Victoria Group, Transantarctic Mountains.In: Craddock, C. (Ed.), Antarctic Geosciences. University of Wisconsin Press, Madison, p. 649-659.

Latterell, J.T., Naiman, R.J., 2007. Sources and dynamics of large logs in a temperate

of

floodplain river. Ecological Applications, v. 17, p. 1127-1141.

ro

Laurie, J.R., Bodorkos, S., Nicoll, R.S., Crowley, J.L., Mantle, D.J., Mory, A.J., Wood,

-p

G.R., Backhouse, J., Holmes, E.K., Smith, T.E., Champion, D.C., 2016. Calibrating the

re

middle and late Permian palynostratigraphy of Australia to the geologic time-scale via U-

lP

Pb zircon CA-IDTIMS dating. Australian Journal of Earth Sciences, v. 63, p. 701–730.

na

Lieth, H., 1975. Primary production of the major vegetation units of the world. In Lieth,

Jo

p. 203–215.

ur

H., and Whitaker, R.H., eds. Primary productivity of the biosphere. New York, Springer,

Liberato, G.P., Cornamusini, G., Perotti, M., Sandroni, S., Talarico, F.M., 2017. Stratigraphy of a Permian-Triassic fluvial-dominated succession in Southern Victoria Land (Antarctica): preliminary data. Journal of Mediterranean Earth Sciences, 9, 167171.

Lindström, S., McLoughlin, S., 2007. Synchronous palynofloristic extinction and recover after the end-Permian event in the Prince Charles Mountains, Antarctica: Implications for

53

Journal Pre-proof palynofloristic turnover across Gondwana. Review of Palaeobotany and Palynology, v. 145, p. 89–122.

Mantle, D.J., Kelman, A.P., Nicoll, R.S., Laurie, J.R., 2010. Australian Biozonation Chart. Geoscience Australia, Canberra, 2 pl.

of

Mays, C., Vajda, V., Frank, T.D., Fielding, C.R., Nicoll, R.S., Tevyaw, A.P.,

ro

McLoughlin, S., 2019. Refined Permian–Triassic floristic timeline reveals early collapse

-p

and delayed recovery of south polar terrestrial ecosystems. GSA Bulletin,

re

https://doi.org/10.1130/B35355.1

lP

McLoughlin, S., 2011. Glossopteris – insights into the architecture and relationships of

ur

1–14.

na

an iconic Permian Gondwanan plant. Journal of the Botanical Society of Bengal, v. 65, p.

Jo

Metcalfe, I., Crowley, J.L., Nicoll, R.S., Schmitz, M., 2015. High-precision U-Pb CATIMS calibration of Middle Permian to Lower Triassic sequences, mass extinction and extreme climate-change in eastern Australian Gondwana. Gondwana Research, v. 28. P. 61–81.

Miall, A.D., 2013. The Geology of Fluvial Deposits: Sedimentary Facies, Basin Analysis, and Petroleum Geology. Springer, 582 pp.

54

Journal Pre-proof Michaelsen, P., 2002. Mass extinction of peat-forming plants and the effect on fluvial styles across the Permian–Triassic boundary, northern Bowen Basin, Australia. Palaeogeography, Palaeoclimatology, Palaeoecology, v. 179, p. 173–188.

Nakayama, K., Fielding, C.R., Alexander, J., 2002. Variations in character and preservation potential of vegetation-induced obstacle marks in the variable discharge

of

Burdekin River of north Queensland, Australia. Sedimentary Geology, v. 149, p. 199–

-p

ro

218.

re

Pettit, N.E., Latterell, J.J., Naiman, R.J., 2006. Formation, distribution and ecological consequences of flood-related wood debris piles in a bedrock confined river in semi-arid

na

lP

South Africa. River Research and Applications, v. 22, p. 1097-1110.

Price, P.L., Filatoff, J., Williams, A.J., Pickering, S.A., Wood, G.R., 1985.Late

ur

Palaeozoic and Mesozoic palynostratigraphic units. In CSR Oil and Gas Division

Report.

Jo

Palynology Facility Report, Queensland Department of Resource Industries Open File

Rees, P.M., Ziegler, A.M., Gibbs, M.T., Kutzbach, J.E., Behling, P.J., Rowley, D.B., 2002. Permian phytogeographic patterns and climate data/model comparisons. Journal of Geology, v. 110, p. 1–31.

55

Journal Pre-proof Retallack, G.J., 1995. Permian and Triassic driftwood from the Allan Hills, Antarctica. Antarctic Journal of the United States, v. 30, p. 37-39.

Retallack, G.J., Krull, E.S., 1999. Landscape ecological shift at the Permian-Triassic boundary in Antarctica. Australian Journal of Earth Sciences, v. 46, p. 785-812.

of

Rey, K., Day, M.O., Amiot, R., Goedert, J., Lécuyer, C., Sealy, J., Rubidge, B.S., 2018.

ro

Stable isotope record implicates aridification without warming during the late Capitanian

re

-p

mass extinction. Gondwana Research, v. 59, p. 1–8.

Rygel, M.C., Gibling, M.R., Calder, J.H., 2004. Vegetation-induced sedimentary

lP

structures from fossil forests in the Pennsylvanian Joggins Formation, Nova Scotia.

na

Sedimentology, v. 51, no. 3, p. 531-552.

ur

Serbet, R., Decombeix, A.-L., Escapa, I.H., Harper, C.J., Gulbranson, E.L., Taylor, E.L.,

Jo

Taylor, T.N., 2016. A diverse Late Triassic flora from the Allan Hills (Lashly Formation), southern Victoria Land, Antarctica. Botanical Society of America, Savannah, GA.

Smith, R.M.H., 1995. Changing fluvial environments across the Permian-Triassic boundary in the Karoo Basin, South Africa and possible causes of tetrapod extinctions. Palaeogeography, Palaeoclimatology, Palaeoecology, v. 117, p. 81-104.

56

Journal Pre-proof Smith, N.D., Barrett, P.J., Woolfe, K.J., 1998. Glacier-fed(?) sandstone sheets in the Weller Coal Measures (Permian), Allan Hills, Antarctica. Palaeogeography, Palaeoclimatology, Palaeoecology, v. 141, p. 35-51.

Smith, T.E., Mantle, D., 2013. Late Permian palynozones and associated CA-IDTIMS dated tuffs from the Bowen Basin, Australia. Record 2013/46. Geoscience Australia:

ro

of

Canberra.

-p

Tal, M., Paola, C., 2007. Dynamic single-thread channels maintained by the interaction

re

of flow and vegetation. Geology, v. 35, p. 347–350.

lP

Tewari, R., Chatterjee, S., Agnihotri, D., Pandita, S.K., 2015. Glossopteris flora in the

na

Permian Weller Formation of Allan Hills, South Victoria Land, Antarctica: Implications for Paleogeography, Paleoclimatology, and Biostratigraphic Correlation. Gondwana

Jo

ur

Research, v. 28, p. 905-932.

Thomas, G.S.P., Connell, R.J., 1985. Iceberg drop, dump, and grounding structures from Pleistocene glacio-lacustrine sediments, Scotland. Journal of Sedimentary Petrology, v. 55, p. 243-249.

Thornthwaite, C.W., 1948. An approach toward a rational classification of climate. Geographical Review, v. 38, p. 55–94.

57

Journal Pre-proof Townrow, J.A., 1967. Fossil plants from Allan and Carapace Nunataks and from the upper Mill and Shackleton Glaciers, Antarctica. New Zealand Journal of Geology and Geophysics, v. 10, p. 456–473.

Vajda, V., McLoughlin, S., Mays, C., Frank, T.D., Fielding, C.R., Tevyaw, A., Lehsten, V., Bocking, M., Nicoll, R.S., 2020. End-Permian (252 Mya) deforestation, wildfires and

of

flooding–An ancient biotic crisis with lessons for the present. Earth and Planetary

-p

ro

Science Letters, v. 529, p. 115875.

re

Vepraskas, M.J., 1994. Redoximorphic features for identifying aquic conditions. North

lP

Carolina Agricultural Research Service Technical Bulletin, no. 301.

na

Viereck, L.A., Dyrness, C.T., Foote, M.J., 1993. An overview of the vegetation and soils of the floodplain ecosystems of the Tanana River, interior Alaska. Canadian Journal of

Jo

ur

Forest Research, v. 23, p. 889-898.

Figure Captions

Figure 1. Geography and paleogeography of study locations. A) Map of the Allan Hills outcrop area denoting the sites studied herein and transects used for mapping measured sections. B) Relative location of the Allan Hills on Antarctica. C) Paleogeographic reconstruction of Antarctica during the late Paleozoic and earliest Mesozoic (from Gulbranson et al., 2014).

58

Journal Pre-proof Figure 2. Field photographs and sketches of the Weller Coal Measures and Feather Conglomerate. A) Weller Coal Measures and Feather Conglomerate near Trudge Valley, Allan Hills, Antarctica. B) Sketch of the photograph in A. Stratigraphic bounding surfaces are drawn on the outcrop. Overbank fines and histosols are denoted by the shaded areas due to their thickness, other paleosol units are drawn as red dashed lines. The Dolores pedotype is the uppermost paleosol in the photograph. C) The Weller Coal

of

Measures and Feather Conglomerate near Manhaul Bay. D) Sketch of the photograph in

ro

C. The Dolores pedotype occurs at the sixth-order contact between the Weller Coal

re

-p

Measures and Feather Conglomerate.

Figure 3. Stratigraphic section of the Permian Weller Coal Measures at the Allan Hills.

lP

Two woody debris zones (zones 1 and 2) are denoted around meters 12 and 15

na

respectively. The stratigraphic position of the Weller Coal Measures section is approximately the upper portion of this formation for South Victoria Land, as 6

ur

prominent coal seams were observed in the Weller Coal Measures at the Allan Hills, and

Jo

the upper two of those six are represented here.

Figure 4. Field photographs and sketches of the Weller Coal Measures, Feather Conglomerate, and Lashly Formation. A) Aerial photograph near Manhaul Bay, Allan Hills, Antarctica, of the Weller Coal Measures, Feather Conglomerate, and Lashly Fm. Note the widespread occurrence of darkbrown dolerite sills in the foreground and multiple dolerite dikes. B) Photograph of the Lashly B Member in the large woody debris zone, Feather Bay, Allan Hills, Antarctica. C) Photograph of the Lashly C Member, near

59

Journal Pre-proof Roscolyn Tor, Allan Hills, Antarctica. D) Sketch of photograph A, only the Lashly A Member is visible above the Feather Conglomerate. E) Sketch of photograph B, note the prevalence of erosive third-order surfaces forming a nested channel. Dips and contacts of second-order surfaces are shown to illustrate erosive versus aggradational third-order surfaces. F) Sketch of photograph C, the in situ forest occurs on the left side of image

of

indicated by the in situ stump symbols.

ro

Figure 5. Stratigraphic section of the uppermost Feather Conglomerate and Triassic

-p

Lashly Formation at the Allan Hills. The prominent woody debris zone at Feather Bay,

re

Allan Hills, Antarctica (Figs. 1 and 4B) is represented around meter 100. The fossil forest near Roscolyn Tor occurs at meter 147. Paleosols shown here represent stacked

lP

accumulations of individual paleosol profiles as opposed to very thick individual paleosol

na

profiles.

ur

Figure 6. Field images of the woody debris zone of the Weller Coal Measures. A)

Jo

Photograph of the upper woody debris zone, in situ stump in the foreground, showing the exposure of bedding planes in this locality. In the distance a coal seam crops out in a cliff, beneath the coal seam are thicker and more continuous beds of carbonaceous and micaceous silty sandstone with transported axes and peat material. B) Line drawing of A showing the locations of in situ stumps in the background. Thin weight lines denote bedding contacts interpreted as upstream accretion surfaces. Thick dashed line denotes a zone of exposed trough cross-bedding, paleocurrent direction indicated by the arrow. Intervals of coeval silty sandstone carbonaceous and micaceous Protosols are indicated

60

Journal Pre-proof by the gray areas. C) In situ stump in the upper woody debris zone (zone 2 of Fig. 3), hammer for scale. D) A second in situ stump in the upper woody debris zone, note the geometry of bed surfaces in proximity to the root flare of the stump, field notebook for scale.

Figure 7. Maps of fossil wood axes and in situ stumps on sandstone bedding planes. A)

of

Permian Weller Coal Measures, Trudge Valley, Allan Hills, Antarctica. The shaded

ro

ellipses denote the position of woody debris and the orientation of the long-axis of the

-p

wood axes. The red ellipses correspond to woody debris and the in situ stump used in the

re

dendrochronologic analysis. Circle symbols represent the location of in situ upright stumps. B) Lashly B Member, Feather Bay, Allan Hills, Antarctica. Triassic woody

lP

debris orientations are indicated by the shaded ellipses. Jurassic dolerite dikes of the

na

Ferrar Group intruding the sandstone, are also shown.

ur

Figure 8. Woody debris accumulations in the Weller Coal Measures and Lashly Fm. A)

Jo

Basal lag of small woody debris fragments above an erosive third-order surface that truncates overbank fine facies, Weller Coal Measures, Trudge Valley. B) Large wood debris, a single wood axis in the upper woody debris zone of the Weller Coal Measures, Trudge Valley. C) Large woody debris in the Lashly B Member, Feather Bay. Note trough shape scour trending towards the person in the background is an erosive remnant of a wood axis. D) Large woody debris in the Lashly B Member, with root flare near the bottom of the photograph, Feather Bay.

61

Journal Pre-proof Figure 9. Overbank fine facies. A) Carbonaceous siltstones of the Weller Coal Measures, Trudge Valley, Antarctica. B) Coal (Histosol) and permineralized peat (Histosol) adjacent to the woody debris zone of the Weller Coal Measures, Trudge Valley.

Figure 10. Triassic paleosol, Lashly A Member, Allan Hills near Dennes Point. A) Photograph of one complete profile of this paleosol, along strike variation is noted in the

of

text. Horizon labels Cg2, Cg1 Btg, Btv, and Bt are on the left side, horizon boundaries

ro

are indicated by the overlain lines, and approximate positions of the photographs in B–F

-p

are indicated in boxes. In B–F stratigraphic up is indicated by the arrow. B) Horizon 1,

re

Cg2, soil formed Mn is shown here as the dark coloration in the green sandstone. C) Horizon 2, Cg1, the redoximorphic variation is shown in this photograph. D) Horizon 3,

lP

Btg, the transition of variegated redoximorphic conditions of Cg1 into the maroon and

na

clay-rich Btg horizon is shown here. Faint subvertical redoximorphic features suggestive of root haloes or bioturbation are seen in the Btg horizon. E) Horizon 4, Btv, red iron-

ur

oxide nodules and tubiform concretions are seen just below the thin sandstone fragment

Jo

(sandstone is not in place). F) Horizon 5, Bt, lack of redoximorphic features, overall maroon color, and clay-films, indicating illuviated clay, are defining characteristics of this horizon.

Figure 11. Ring width indices (RWI) for cross-matched Permian and Triassic fossil wood. A) Cross matched RWI time series for Permian fossil wood from the upper woody debris zone (zone 2 of Fig. 2). Dark line indicates the RWI chronology for the 5 crossmatched trees. The colored lines indicate individual RWI time series for each specimen in

62

Journal Pre-proof their relative dating chronologic order. B) Cross-matched RWI chronology for Triassic fossil wood, Feather Bay, Allan Hills. The dark line indicates the RWI chronology for all eight of the cross-matched trees. Colored lines indicate individual RWIs in their relative chronologic order.

Figure 12. Individual ring width measurements for the five cross-matched Permian fossil

of

wood samples, Trudge Valley, Allan Hills. The ring width series are shown in a stacked

ro

arrangement to illustrate how the individual wood axes are arranged in space and in their

-p

relative chronologies. Wood axes deposited downstream of the in situ stump (AH-31) are

re

shown above the ring width series for AH-31; whereas wood axes deposited upstream of

lP

the in situ stump are shown below AH-31.

na

Figure 13. Water balance diagrams for the Late Permian and Early Triassic based on the climatographs produced by climate modelling in Fielding et al. (2019). A) Late Permian

ur

water balance diagram, evapotranspiration (Etp) produced by temperature correction as a

Jo

function of latitude by 0.63K/ºlatitude from 65ºS to 80ºS. The austral winter has a 5–6 month period of 24 h of darkness indicated by the dark shaded “polar night” region. Monthly average temperatures below freezing outside of the polar night are indicated with dashed blue lines. The accumulated precipitation during the austral winter is interpreted to act as snow/ice melt during the austral spring and contribute to the modelled precipitation during that month. B) Early Triassic water balance diagram. Based on paleogeographic reconstructions (Scotese, 2014), the field area was approximately 10º equatorward of the Late Permian position, resulting in reduced length

63

Journal Pre-proof of the polar night and temperature correction from the 65ºS paleoclimate simulation. A snow/ice melt effect was estimated for this time interval, resulting in 40 cm of meltwater contributing to the monthly precipitation influx to the land surface during the austral spring.

Figure 14. Evolution of landscapes through the Permian and Triassic, Allan Hills,

of

Antarctica. Fossil forest maps representing in situ tree positions for the Permian Weller

ro

Coal Measures and Triassic Roscolyn Tor fossil forests are shown at the bottom and top

-p

of the figure, respectively. Representative paleosol profiles and landscape position are

re

shown for each block diagram. Block diagrams reflect the synthesis of sedimentary facies identified for each lithostratigraphic unit. Plant macrofossils are illustrated with their

lP

relative density and diversity based on compression floras. Based on reconstructed tree

na

density and wood basal area the Triassic Roscolyn Tor fossil forest is more representative of early succession forests, similar to extant early succession boreal forests (Viereck et

ur

al., 1993). In contrast, the Permian Trudge Valley fossil forest has tree density and basal

Jo

area more closely aligned with intermediate succession floodplain forests.

64

Journal Pre-proof Table 1. Compilation of architectural elements (Miall, 2013; Ielpi et al., 2014) in the Weller Coal Measures, Feather Conglomerate, and Lashly Formation. Weller Coal Measures Position Element Description In-Channel

Overbank

SB

Sandy bedforms: volcaniclithic cross-bedded units bound by second-order surfaces. Sets of crossbedded units, bound by third-order surfaces can be aggradational or erosive.

DA, LA, UA

Downstream-, lateral-, and upstream-accretion macroforms: Sandstone bodies inclined relative to paleo-flow in the direction indicated.

LWD

Large Woody Debris: Accumulations of permineralized fossil wood ranging in length from 0.5 m to 9 m in length. Depositional units containing LWD are either cross-bedded or exhibit planeparallel bedding. LWD occurs as lag deposits; or as the basal units of UA/DA surfaces adjacent to macroform tops. Overbank fines: Either tabular carbonaceous siltstones with parallel bedding, or as organic paleosols (Histosols) that are either coalified or permineralized. Crevasse splay: Tabular siltstones and sandy siltstones that exhibit tapering thickness and fining upward grain size.

OF CS

of

Feather Conglomerate Element

Description

In-Channel

GB

Gravel bedforms: arkosic to subarkosic cross-bedded units bound by second-order surfaces. Typically forms the basal deposit of a stratigraphic unit bound by fifth-order surfaces. Sandy bedforms: arkosic cross-bedded units bound by second-order surfaces. Sets of cross-bedded units, bound by third-order surfaces can be aggradational or erosive. Preserved macroforms of SB facies are bound by fourth-order surfaces and display a mounded surface geometry. Overbank fines: mineral paleosols (e.g., Dolores pedotype) that are dark brown to drab green in color and display weak horizonation. OF facies is predominant near the base of the succession and decreases in abundance upsection. OF facies occurs more predominantly throughout the succession down-dip of the study area.

OF

-p

Overbank

re

SB

ro

Position

Lashly Formation Element

Description

In-Channel

SB

Sandy bedforms: volcaniclithic cross-bedded units bound by second-order surfaces.

GB

Gravel bedforms: quartzite and paleosol clast lag deposits in coarse- to medium-grained volcaniclithic sandstone matrix. Occur as basal units of discrete channel forms or directly overlying erosive third-order surfaces of sheet sand units. Predominant in the Lashly A Member

LA

Lateral accretion macroforms: Sandstone bodies inclined roughly 90º to paleoflow direction, occur within laterally discontinuous and tapering sandstone units bound by third-order surfaces. LA facies occur in the Lashly A and C members. Large Woody Debris: Accumulations of permineralized fossil wood, primarily in the Lashly B Member, and to a minor extent as lag deposits in the Lashly A Member. Depositional units containing LWD are cross-bedded to plane-parallel bedded volcaniclithic sandstone with minimal OF facies. Overbank fines: plane-parallel to cross-bedded sandstone and silstone, bioturbated silty sandstone, and minor carbonaceous siltstone with leaf compression fossils. Leaf compression fossils in carbonaceous siltstone occur in the upper part of the Lashly A Member. Bioturbated siltstone is abundant throughout the Lashly A Member. Crevasse splay: Cross-bedded to massive-bedded sandstone displaying a tapering thickness.

Overbank

na

ur

Jo

LWD

lP

Position

OF

CS

65

Journal Pre-proof

Jo

ur

na

lP

re

-p

ro

of

Triassic

Permian

Table 1. Statistics on cross-matches for Permian fossil wood and Triassic fossil wood Reference Test Gl CC Tho TBP Overlap AH-31 AH-27 66.7 0.35 3.44 2.55 8 AH-31 AH-29 60.9 0.07 2.72 2.77 45 AH-31 AH-35 80.0 0.13 3.71 3.45 21 AH-31 AH-47 68.1 0.37 3.41 2.59 37 AH-29 AH-47 57.7 0.04 2.93 2.98 30 AH-29 AH-35 82.1 -0.03 3.28 1.76 16 AH-27 AH-35 77.8 0.22 2.5 2.36 14 AH-Tr-4 AH-Tr-1 75 -0.13 3.66 3.87 26 AH-Tr-4 AH-Tr-2 93.8 0.41 1.31 1.07 8 AH-Tr-4 AH-Tr-3 75 0.45 3.08 2.76 36 AH-Tr-4 AH-Tr-5 82.1 0.84 5.15 4.13 14 AH-Tr-4 AH-Tr-7 62.5 0.69 2.63 2.76 28 AH-Tr-4 AH-Tr-8 69.2 0.06 4.46 4.04 78 Notes: GL is gleichlaufigkeit, CC is correlation coefficient, Tho and TBP are t-statistics after Baillie and Pilcher (1973), and overlap indicates the number of rings overlapping between the test specimen and reference chronology.

66

Journal Pre-proof Declaration of interests ☒ The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Jo

ur

na

lP

re

-p

ro

of

☐The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:

67

Journal Pre-proof Highlights

of ro -p re lP



na



ur



Application of dendrochronology to Permian and Triassic fossil wood from Antarctica Dendrochronology reveals relationship of transported wood to fluvial bar formation Paleoclimate and/or paleoecology differences are distinct in tree ring chronologies Synthesis of paleosols, sedimentology, and tree ring records to interpret paleoecology

Jo



68

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14