Paleozoic landscapes shaped by plant evolution

Abstract

Fluvial landscapes diversified markedly over the 250 million years between the Cambrian and Pennsylvanian periods. The diversification occurred in tandem with the evolution of vascular plants and expanding vegetation cover. In the absence of widespread vegetation, landscapes during the Cambrian and Ordovican periods were dominated by rivers with wide sand-beds and aeolian tracts. During the late Silurian and Devonian periods, the appearance of vascular plants with root systems was associated with the development of channelled sand-bed rivers, meandering rivers and muddy floodplains. The widespread expansion of trees by the Early Pennsylvanian marks the appearance of narrow fixed channels, some representing anabranching systems, and braided rivers with vegetated islands. We conclude that the development of roots stabilized the banks of rivers and streams. The subsequent appearance of woody debris led to log jams that promoted the rapid formation of new river channels. Our contention is supported by studies of modern fluvial systems and laboratory experiments. In turn, fluvial styles influenced plant evolution as new ecological settings developed along the fluvial systems. We suggest that terrestrial plant and landscape evolution allowed colonization by an increasingly diverse array of organisms.

Full text

NATURE GEOSCIENCE | VOL 5 | FEBRUARY 2012 | www.nature.com/naturegeoscience 99
Terrestrial landscapes are dominated by a remarkable assem-
blage of vascular plants, mosses and liverworts, lichen, fungi,
algae and microbial mats. Vegetation cover is so perva-
sive that we take it for granted, and some 84% of the land surface
experiences geomorphological change associated with organisms1.
Paradoxically, studies of ancient sedimentary systems have oen
undervalued the role of plants in shaping uvial systems, partly
because human activity has profoundly inuenced vegetation along
many modern rivers with which we are familiar. e science of
uvial geomorphology, for example, was based largely on channel
systems that lacked a substantial load of woody debris2. e fun-
damental interaction between vegetation and landscape through
the Palaeozoic era, when land plants rst colonized the Earth, is
receiving increased attention. is Review draws together recent
contributions from the disciplines of sedimentology, palaeontology,
geomorphology, ecology, engineering and experimental science.
e Palaeozoic ‘greening’ of terrestrial landscapes heralded a
fundamental and irreversible change in sedimentation patterns,
one of the most remarkable in the planet’s history, and resulted
in far-reaching modications to the atmosphere, oceans and bio-
logical ecosystems (Fig. 1). ese changes were set against the
background of continental collision and the progressive assem-
bly of the Pangaean supercontinent with its Himalayan-scale
mountainranges.
Biogeomorphology and the Devonian plant hypothesis
e expanding discipline of biogeomorphology owes its origin
to Charles Darwin’s far-sighted treatise on earthworms published
late in his life, which highlighted the importance of these animals
in promoting landscape denudation over geological timescales3.
e discipline emphasizes the two-way coupling between organ-
isms and landscapes: organisms serve as geomorphic engineers to
shape landscapes to their own advantage or as an indirect conse-
quence of their activity, and landscapes in turn inuence organic
evolution1,4–8. On a global scale, such geomorphic engineering may
reect the activity of many taxonomic groups over tens to hun-
dreds of millions of years.
How did the Palaeozoic evolution of plants inuence terres-
trial landscapes? e ‘Devonian plant hypothesis’9,10 draws on
known events in plant evolution, the marine record and the evo-
lution of the atmosphere11, inferring that the spread of vascular
Palaeozoic landscapes shaped by plant evolution
Martin R. Gibling* and Neil S. Davies†
Fluvial landscapes diversified markedly over the 250 million years between the Cambrian and Pennsylvanian periods. The
diversification occurred in tandem with the evolution of vascular plants and expanding vegetation cover. In the absence of
widespread vegetation, landscapes during the Cambrian and Ordovican periods were dominated by rivers with wide sand-beds
and aeolian tracts. During the late Silurian and Devonian periods, the appearance of vascular plants with root systems was
associated with the development of channelled sand-bed rivers, meandering rivers and muddy floodplains. The widespread
expansion of trees by the Early Pennsylvanian marks the appearance of narrow fixed channels, some representing anabranching
systems, and braided rivers with vegetated islands. We conclude that the development of roots stabilized the banks of rivers and
streams. The subsequent appearance of woody debris led to log jams that promoted the rapid formation of new river channels.
Our contention is supported by studies of modern fluvial systems and laboratory experiments. In turn, fluvial styles influenced
plant evolution as new ecological settings developed along the fluvial systems. We suggest that terrestrial plant and landscape
evolution allowed colonization by an increasingly diverse array of organisms.
plants greatly inuenced the Earth system because plants medi-
ate weathering intensity through their eect on soils. As increased
plant cover and root penetration enhanced bedrock weathering,
nutrient runo promoted plankton blooms in shallow seas and
the extinction of goniatites and other marine fauna. Marine and
coastal strata stored large amounts of carbon12 and, as photosyn-
thetic activity increased, the rise of atmospheric O2 and extrac-
tion of CO2 (Fig.1) plunged the Earth into a major ice age by the
Late Devonian. However, until recently, research on the global
inuence of Palaeozoic vegetation has provided little information
about the terrestrial landscapes where land plants grew.
Fluvial style and plant evolution throughout the Palaeozoic
Before vegetation cover, rivers were broad bed-load systems with
unstable banks13,14. is inference is borne out by observations
of Precambrian and earliest Palaeozoic uvial-channel deposits
that demonstrate a ‘sheet-braided’ style (Figs1, 2a, 3) generated
by wide, shallow channels with low relief margins, and with little
evidence of muddy oodplains15–18. Aeolian deposits were wide-
spread, and ne sediment was largely deated and blown out to
sea19–22. At basin margins, some Cambro–Ordovician alluvial-
fan deposits show little sand and mud, suggesting that adjoining
uplands were only slightly weathered in the absence of vascular-
plant cover23. However, by the late Proterozoic eon, increased for-
mation of pedogenic clay minerals implies enhanced biotic soil
activity and the inception of the ‘clay mineral factory’24.
Embryophytes (land plants) rst appeared by the Middle
Ordovician period, on the basis of discoveries of palynomorphs
and plant fragments25–28 (Fig.1). Mud then became prominent in
alluvial settings from basin-margin alluvial fans down to the coastal
zone, signifying enhanced upland mud production, reduced dea-
tion and the rise of oodplains14. e universal sheet-braided style
gave way progressively to a ‘channelled-braided’ style (Fig.3) of
thick sandstone sheets composed of amalgamated lenses, suggest-
ing that channel sediments were more cohesive and channel forms
more readily preserved18.
rough the Late Silurian and Early Devonian, many funda-
mental evolutionary traits appeared in plants for the rst time29,
and vascular plants with water-transmission and supporting struc-
tures became abundant, although rooting structures were mod-
est30,31 (Fig.2b). e enigmatic Prototaxites appeared, interpreted
Department of Earth Sciences, Dalhousie University, PO Box15000, Halifax, Nova Scotia, Canada B3H 4R2. †Present address: Department of Geology and
Soil Sciences, Krijgslaan 281, S8, University of Ghent, 9000 Ghent, Belgium. *e-mail: mgibling@dal.ca
FOCUS | REVIEW ARTICLE
PUBLISHED ONLINE: 1 FEBRUARY 2012 | DOI: 10.1038/NGEO1376
© 2012 Macmillan Publishers Limited. All rights reserved

100 NATURE GEOSCIENCE | VOL 5 | FEBRUARY 2012 | www.nature.com/naturegeoscience
as a gigantic fungus32,33. Isotopic excursions documented in marine
carbonates (Fig.1) are consistent with greatly increased vegetation
cover at the Siluro–Devonian boundary34. By the Early Devonian,
the presence of charcoal testies to the earliest forest res35, the
rst wood appeared36, and vegetation was suciently abundant to
accumulate locally as peat14.
By the latest Silurian, single-thread meandering systems
appeared, identied by the presence of lateral-accretion (point
bar) surfaces37. Initially restricted to small creeks, meandering
systems are present in more than 30% of uvial rock units of Late
Devonian age, forming trunk channels (Fig.3). eir presence sig-
nies a considerable increase in avulsive behaviour through chute
and neck cuto and the relocation of entire channel belts. e par-
allel evolution of meandering rivers and rooted vegetation pro-
vides circumstantial evidence that vegetation stabilized banks and
promoted systematic channel migration. In Early Devonian strata,
plant fragments are found in braided-river and alluvial-fan depos-
its, supporting the isotopic evidence for increased plant cover38.
By the Middle to Late Devonian, lowlands were colonized by
archaeopteridalean and cladoxylalean trees, some up to 25 m
tall with large roots, which formed the Earth’s rst dense, shady
forests39,40. Isotopic excursions suggest that vegetation cover may
have increased from 10% to 30% at the Middle–Upper Devonian
boundary41. e Late Devonian advent of the seed habit allowed
trees to colonize drier alluvial plains42. Mississippian plants also
greatly diversied (Fig.1), with some developing tolerance of sea-
sonal growth and water stress (as indicated by tree rings)43. Late
in the Mississippian, conifer pollen appeared for the rst time,
and the related group of cordaitalean trees, some nearly 50m tall,
became prominent44,45. ese gymnosperms diversied early in
the Pennsylvanian (Fig.1), colonizing dry alluvial plains and even
evaporitic sabkhas46,47. By the Pennsylvanian, vegetation was capa-
ble of growing within channels48.
Although the prolic coal-swamp ora (Fig.2c) is oen featured
as the archetypal Pennsylvanian vegetation in museum dioramas,
the gymnosperm oras of alluvial plains were arguably the domi-
nant biome, and their rise probably marked a threshold in land-
scape evolution. Modern conifers have a high tracheid content,
which provides eective support and water transport, explaining
in part their drought resistance and successful colonization of
drier terrain49. Many modern conifers and other plants have roots
that reach depths of tens of metres, allowing direct and ecient
access to the water table50. Although Palaeozoic roots are incom-
pletely preserved, some reached at least 4m below the surface of
well-drained alluvial plains45. Log jams are prominent in some
Early Pennsylvanian uvial deposits51,52, highlighting the increased
abundance of large woody debris as riparian trees grew higher
and became more common, causing channel blockage and pro-
moting break-out (avulsion). By the Mississippian, more frequent
occurrences of charcoal-bearing strata indicate the importance of
300
Ma
350
400
450
500
550
Permian
Penn.
Miss.
Devonian
Silurian
Ordovician
Cambrian
Protero.
E
M
L
E
M
L
Llan.
Wen.
Lud.
Prid.
E
M
L
Tour.
Vis.
Serp.
E
M
L
Fluvial style
Large log jams;
Expansion into
drylands
Gymnosperms
Seed plants
Forests
Carbonaceous
compressions
Definitive
cryptospores
Coal
Definitive roots
Root-like forms,
charcoal
Diverse spores
Origin of
most
tracheophyte
nutritional
traits
Atmosphere
(%)
(x present
level)
Events of
vegetation
expansion
(isotopic
evidence)
(Oldest known events; key events)
Island-
braided
Channelled-
braided
Small meandering
channels
Trunk meandering
channels
Sand-filled
fixed
channels
(Approximate onset)
Carbon
burial
Families
Origination rate
(families / Ma)
Plant evolution Animal
evolution
O2CO2
(x 1018 mol
Myr–1)
1625200020050
Reptile fossils
Tetrapod fossils
Tetrapod tracks
Hexapods
arachnids
Freshwater fish,
expansion of
terrestrial
trace fossils
Te rrestrial
arthropod
tracks
Sheet-braided
(from earlier
Precambrian)
Figure 1 | Palaeozoic events of fluvial and landscape development, in relation to plant evolution and atmospheric change. Data sources: atmospheric
CO2 and O2 curves11, employing volcanic weathering factor in CO2 curve, and carbon burial curve12 (curves have wide error bars); fluvial styles and vascular-
plant events14,18,37,55; numbers of plant families in time intervals and their origination rates96,97; plant nutritional traits29; events of rapid vegetation expansion,
inferred from isotopic excursions at the Siluro–Devonian boundary34 and at the Famennian–Frasnian boundary (Late Devonian)41; key events in animal
evolution88,90,93,94,98,99; geological timescale100. Protero., Proterozoic; Miss., Mississippian; Penn., Pennsylvanian; Serp., Serpukhovian; Vis., Visean; Tour.,
Tournaisian; Prid., Pridolian; Lud., Ludfordian; Wen., Wenlockian; Llan., Llandovery.
REVIEW ARTICLE | FOCUS NATURE GEOSCIENCE DOI: 10.1038/NGEO1376
© 2012 Macmillan Publishers Limited. All rights reserved

NATURE GEOSCIENCE | VOL 5 | FEBRUARY 2012 | www.nature.com/naturegeoscience 101
wildre as a geomorphic agent, with strong eects on landscape
stability and sedimentation53.
Two distinctive uvial styles appear prominently in the
Pennsylvanian for the rst time (Fig. 3). First, suites of channel
sandstones encased in oodplain mudstones are suciently narrow
for one or both margins to be visible in outcrop (Fig.2d). ese
sand-lled xed channels have commonly been attributed to anas-
tomosing or anabranching rivers with several coexisting threads,
although they may also represent single-thread channels that
avulsed frequently. Although some mud-lled channels in Siluro–
Devonian rocks may have been part of anabranching networks54,
sand-lled narrow channels rst appear commonly early in the
Pennsylvanian55. Second, within braided-river deposits, the abun-
dance of logs lling deep channels52 suggests comparison to modern
island-braided systems that have discrete channels with abundant
woody debris between vegetated islands56. ese two uvial styles
imply the further expansion of riparian corridors and avulsive
strategies (Fig.3). ey correspond broadly with the Mississippian
and Pennsylvanian diversication of plants, the incoming of coni-
fers, the expansion of vegetation across dry alluvial plains, and the
appearance of large log jams (Fig.1). More directly, the abundance
of sedimentary features formed around upright trees and the pres-
ervation of upright trees on channel banks55,57 suggest that vegeta-
tion inuenced erosion and deposition on the scale of bedforms and
channel belts. Although dicult to assess, vegetation had probably
spread to upland bedrock tracts, or at least to upland valleys44,52,58.
e study of molecular clocks has yielded evidence that many
key evolutionary traits originated long before their rst appear-
ance in the fossil record. Although the earliest undisputed embry-
ophyte fossils are Ordovician, embryophytes may have originated
in the Precambrian, perhaps more than one billion years ago59.
However, the inuence of plants on landscape evolution is closely
tied to the extent and density of vegetation, which should lag well
behind the rst appearances of taxa14. As noted above, the general
accord between changes in uvial style and macrofossil abundance
suggests that the latter is a reasonable proxy for ‘critical mass’ in
landscape evolution.
The impact of vegetation on modern landscapes
e profound inuence of vegetation (largely angiosperm-domi-
nated) on uvial style is supported by a wealth of evidence from
modern landscapes. In Australian anabranching rivers (Fig.2e),
vegetation exerts an inuence on all scales from scours and vegeta-
tion shadows associated with individual trees to the sculpturing
of whole channel systems60–62. ese studies show that vegetation
promotes channel stability by increasing bank cohesion and reduc-
ing erosion, principally through the binding power of roots that
commonly extend below bank base or form surface mats. During
dry periods, trees that become established within the channels
buttress the banks and generate bars and ridges that separate the
anabranches. Studies elsewhere conrm the inuence of vegeta-
tion in a wide range of uvial settings63,64.
On the basis of eld and laboratory engineering studies, the
eect of root strength in stabilizing landscapes has been progres-
sively quantied65–68. e root systems of individual plants are
impressive (Fig.2f ), and roots play a vital role in soil and slope
stability because they impart tensile strength. In some cases, roots
provide virtually all the cohesive strength of the soil, with which
they are frictionally coupled.
Within many channels, large woody debris accumulates in su-
cient quantity to create log jams and promote channel avulsion63,69–70,
and woody debris has been cited as a crucial driver of landscape
evolution71. In sandy braided settings, woody debris promotes bar
development and island formation, especially through the regrowth
of transported live wood and the germination of seeds stranded
along banks during oods56,72. Some buried logs resist decay for
thousands of years and may be reworked into modern streams, con-
tributing to blockage73.
a
def
bc
10 m
Coal
Coal
A
R
Channel
50 m wide
Fixed-channel
body
6 m
Figure 2 | Plants and fluvial systems in ancient and modern settings. a, Braided-fluvial sheets, Alderney Sandstone, Cambrian, Channel Islands.
b, Rhynia (R) and Aglaophyton (A) in growth position in hotspring silica, Rhynie Chert, Pragian, Scotland. Pound coin is 2.3cm in diameter (See Fig.6a,b
of ref. 31 for more detail). Image courtesy of N. Trewin. c, Lycopsid trees, arrowed, rooted in former peat (coal seam), Sydney Mines Formation, Middle
Pennsylvanian, Canada. Hammer (centre, right) is 30 cm long. d, Fixed-channel body of sandstoneand mudstone, possibly deposited by an anastomosing river,
with red mudstone and crevasse-splay sandstone, Joggins Formation, Canada. e, Anastomosing Diamantina River, Queensland, Australia, with eucalyptus
trees (Eucalyptus microtheca) along riparian zones. f, Extensive eucalyptus roots exposed by bank erosion, Thomson River, Queensland, Australia.
FOCUS | REVIEW ARTICLE
NATURE GEOSCIENCE DOI: 10.1038/NGEO1376
© 2012 Macmillan Publishers Limited. All rights reserved

102 NATURE GEOSCIENCE | VOL 5 | FEBRUARY 2012 | www.nature.com/naturegeoscience
Unfortunately, we know little about alluvial tracts that have
escaped human interference, and we have probably underesti-
mated the inuence of vegetation on unmodied river systems.
Comparison of a pristine Australian watershed with a neighbour-
ing one aected by agricultural activity, logging and the removal
of woody ‘snags’ from the channels yielded arresting results74.
e pristine channel was a narrow, winding creek, beset by fallen
trees, whereas the modied channel was a straight, wide system
with caving banks and rapid sediment transport. It is clear that
few modern rivers are suitable analogues for Late Palaeozoic riv-
ers. In northwest Europe, oodplain deforestation was underway
by about 6,000  and many lowland rivers may have been ana-
branching before human modication75,76.
Fluvial analogue and numerical modelling has recently made
great strides in adding vegetation to the models77–81. Although
braided rivers are only part of the planform range, earlier ume
studies created virtually nothing else, suggesting that a key experi-
mental parameter was missing. Recent experiments have modied
experimental sediment surfaces by seeding them with rapidly ger-
minating alfalfa, demonstrating that an important missing ingre-
dient was vegetation, which suppresses the basic instability that
causes braiding78. In all the experimental set-ups with vegetation,
bank mobility and the number of active channels were reduced,
leading to lower migration rates, narrower and deeper channels,
and a decreased width/depth ratio. Even with a relatively sparse
alfalfa cover, ows began to split around stable islands, generating
an island-braided style77,79. By seeding alfalfa during simulated low-
ow stages, researchers generated a meandering channel that, for
the rst time in experiments, migrated and maintained its form by
balancing erosion and deposition, resulting in alluvial tracts with
stable oodplains80 (Fig.4). Furthermore, when ne sediment is
deposited in chute channels of vegetated systems, the chutes do
not develop into full-scale cutos; this prevents the channel belts
from being destroyed too rapidly and maintains a meandering
style81. us, the experiments illustrate how vegetation and clay
deposition promote island-braided and single-thread channels,
and shi the braided planform towards an anabranchingpattern79.
Experiments have also shown that ood disturbance selec-
tively uproots seedlings with particular root and stem character-
istics, until the remaining plants are suciently robust to survive
later oods as they are out of scale with ows that might uproot
them82,83. ese experiments highlight the ability of roots to grow
suciently rapidly to survive uprooting during intermittent oods,
and conrm the dynamic interaction between hydrology and veg-
etation growth over a range of timescales.
Evolving landscapes and biological evolution
Shaped themselves by plants, how may evolving landscapes have
in turn inuenced organic evolution? From the Silurian onwards,
Flow
Point
bar
2 m
Alfalfa
300
Ma
350
400
450
500
550
Penn.
Miss.
Sheet-braided
Channelled-
braided
Fixed
channel
Island-
braided
% of fluvial rock units
Avulsive
strategies
Riparian
corridors
Levees
& splays
Floodplain
area
Increase
with
time
050 100
Dev.
Sil.
Ord.
Camb.
Protero.
Meandering
Figure 3 | Palaeozoic diversification of fluvial style. Proportions of
rock units with braided, meandering and fixed-channel styles based on
assessment of 330 rock units14,55. Meandering rivers were identified by
heterolithic lateral-accretion sets, and fixed channels by ribbons to narrow
sheets with vertically aggraded fill. The proportion of channelled-braided
and island-braided styles is estimated, as few literature descriptions
provide sucient detail for assessment. Penn., Pennsylvanian; Miss.,
Mississipian; Dev., Devonian; Sil., Silurian; Ord., Ordovician; Camb.,
Cambrian; Protero.; Proterozoic.
Figure 4 | Experimental study of eects of vegetation on channels.
Meandering channel created in flume at St Anthony Falls Laboratory,
Minneapolis. Following initial set-up of a braided channel, seeding with
alfalfa (green) during low-discharge periods stabilized the banks and
resulted in restriction of flow to a self-maintaining single-thread channel,
shown with red dye. The channel migrated systematically, and was
bordered by a stable floodplain80. A flume length of 10m is shown, viewed
during low flow. Image courtesy of M. Tal.
REVIEW ARTICLE | FOCUS NATURE GEOSCIENCE DOI: 10.1038/NGEO1376
© 2012 Macmillan Publishers Limited. All rights reserved

NATURE GEOSCIENCE | VOL 5 | FEBRUARY 2012 | www.nature.com/naturegeoscience 103
plant colonization ushered in a fundamental global coupling
between geomorphic and biological processes, with particular
importance for faunal diversication associated with standing veg-
etation and leaf litter84. New ecosystems appeared during this inter-
val in a series of major palaeoecological events85.
We identify three key changes in Palaeozoic uvial systems
(Fig. 3), forced by plant evolution, that would have led to posi-
tive feedbacks for both plant and animal evolution14,55. First, as a
consequence of enhanced upland weathering, a general increase in
mud from the Middle Ordovician onward produced varied low-
land substrates and more accessible nutrients. Second, the develop-
ment of meandering rivers with strengthened banks through the
Late Silurian and Devonian promoted stable muddy oodplains
partially protected by levees and highly suitable for vegetation
growth and soil development, including carbonate-rich soils14,86.
ird, Early Pennsylvanian suites of narrow xed channels on
muddy plains and of island-braided styles in sand-bed rivers imply
an increased length of channel margins and riparian corridors
with their varied plant communities and subsurface water prisms,
which are of crucial importance for many animal species87. e
Devonian to Pennsylvanian development of avulsive channel sys-
tems generated abandoned channels suitable for organic occupa-
tion, especially during dry periods. Towards the sea, concomitant
expansion of muddy coastal plains and deltas would have also had
many biological consequences. Feedback loops are likely to have
been complex.
It is probably no coincidence that many key plant develop-
ments29 took place within this diversifying alluvial framework,
and these changes must collectively have inuenced the evolution
of animals. Although animals made intermittent passage across
subaerial substrates during the Cambro–Ordovician88, uncontro-
versial fossil evidence for robust terrestrial faunas is absent until
the Siluro–Devonian, when vegetated muddy oodplains and plant
litter became available89. Trace-fossil suites testify to a signicant
Silurian to Pennsylvanian invasion of continental ecospace, with
distinctive assemblages recorded in active and abandoned chan-
nels, desiccated overbank areas and poorly drained swamps, lakes
and soils90,91. is invasion may be linked to plant diversication, as
postulated for the diversication of brackish traces85 within evolv-
ing coastal and deltaic landscapes. Trophically modern ecosys-
tems appeared during the Late Silurian to Middle Devonian when
arthropods were feeding on sporangia, stems and fungal thalli;
by the late Mississippian to Pennsylvanian, they were also target-
ing roots, leaves, woods and seeds92. For vertebrates, sh began to
move into terrestrial settings during the Middle to Late Silurian93,
and there is an expanding Devonian tetrapod record with tracks
known from the early Middle Devonian94, although the earliest
herbivorous tetrapods date to the Pennsylvanian95.
e body of research reviewed here records not only the
Palaeozoic invasion of the land by plants and animals over a period
of some 250 million years, but also an intricate interplay between
organisms and physical environments, represented in the growing
discipline of biogeomorphology. Plants in particular acted as geo-
morphic engineers, but the diversied uvial realm that they engi-
neered in turn altered the framework for organic evolution.
References
1. Corenblit, D. etal. Feedbacks between geomorphology and biota controlling
Earth surface processes and landforms: A review of foundation concepts and
current understandings. Earth Sci. Rev. 106, 307–331(2011).
2. Montgomery, D.R. & Piégay, H. Wood in rivers: interactions with channel
morphology and processes. Geomorphology 51, 1–5 (2003).
3. Darwin, C. e Formation of Vegetated Mould through the Action of Worms
with Observation of their Habitats (John Murray, 1883).
4. Corenblit, D., Steiger, J., Gurnell, A.M. & Tabacchi, E. Darwinian origin of
landforms. Earth Surf. Proc. Land. 32, 2070–2073 (2007).
5. Fisher, S.G., Heernan, J.B., Sponseller, R.A. & Welter, J.R. Functional
ecomorphology: Feedbacks between form and function in uvial landscape
ecosystems. Geomorphology 89, 84–96 (2007).
6. Murray, A.B., Knaapen, M.A. F., Tal, M. & Kirwan, M.L.
Biomorphodynamics: Physical-biological feedbacks that shape landscapes.
Water Resour. Res. 44, W11301 (2008).
7. Osterkamp, W.R., Hupp, C.R. & Stoel, M. e interactions between vegetation
and erosion: new directions for research at the interface of ecology and
geomorpholog y. Earth Surf. Proc. Land. http://dx.doi.org/101002/esp.2173 (2011).
8. Jones, C.G., Lawton, J.H. & Shachak, M. Organisms as ecosystem engineers.
Oikos 69, 373–386 (1994).
9. Algeo, T.J., Berner, R.A., Maynard, J.B. & Scheckler, S.E. Late Devonian
oceanic anoxic events and biotic crises: “Rooted” in the evolution of vascular
land plants? GSA Today 5, 64–66 (1995).
10. Algeo, T.J. & Scheckler, S.E. Terrestrial-marine teleconnections in the
Devonian: links between the evolution of land plants, weathering processes,
and marine anoxic events. Philos. Trans. R.Soc. Lond. B 353, 113–130 (1998).
11. Berner, R.A. GEOCARBSULF: A combined model for Phanerozoic
atmospheric O2 and CO2. Geochim. Cosmochim. Acta 70, 5653–5664 (2006).
12. Berner, R.A. e long-term carbon cycle, fossil fuels and atmospheric
composition. Nature 426, 323–326 (2003).
13. Schumm, S.A. Speculations concerning paleohydrologic controls of terrestrial
sedimentation. Geol. Soc. Am. Bull. 79, 1573–1588 (1968).
14. Davies, N.S. & Gibling, M.R. Cambrian to Devonian evolution of alluvial
systems: e sedimentological impact of the earliest land plants. Earth Sci.
Re v. 98, 171–200 (2010).
15. Cotter, E. in Fluvial Sedimentology (ed. Miall, A.D.) 361–383 (Canadian
Society of Petroleum Geologists Memoir 5, 1978).
16. Long, D.G. F. in e Precambrian Earth: Tempos and Events (eds
Eriksson,P.G., Altermann, W., Nelson, D.R., Mueller, W.U. & Catuneanu, O.)
660–663 (Elsevier, 2004).
17. Long, D.G.F. in From River to Rock Record: e Preservation of Fluvial
Sediments and their subsequent Interpretation (eds Davidson, S., Leleu, S. &
North, C.P.) 37–61 (SEPM, 2011).
18. Davies, N.S., Gibling, M.R. & Rygel, M.C. Alluvial facies evolution during
the Palaeozoic greening of the continents: case studies, conceptual models and
modern analogues. Sedimentology 58, 220–258 (2011).
19. Dott, R.H. Jr & Byers, C.W. SEPM research conference on modern shelf and
ancient cratonic sedimentation - the orthoquartzite-carbonate suite revisited.
J.Sedim. Petrol. 51, 329–347 (1981).
20. Dott, R.H. Jr, Byers, C.W., Fielder, G.W., Stenzel, S.R. & Winfree, K.E.
Aeolian to marine transition in Cambro-Ordovician cratonic sheet sandstones
of the northern Mississippi valley, U.S.A. Sedimentology 33, 345–367 (1986).
21. Dott, R.H. Jr e importance of eolian abrasion in supermature quartz
sandstones and the paradox of weathering on vegetation-free landscapes.
J.Geol. 111, 387–405 (2003).
22. Dalrymple, R.W., Narbonne, G.M. & Smith, L. Eolian action and the
distribution of Cambrian shales in North America. Geology 13, 607–610 (1985).
23. Went, D.J. Pre-vegetation alluvial fan facies and processes: an example from
the Cambro-Ordovician Rozel Conglomerate Formation, Jersey, Channel
Islands. Sedimentology 52, 693–713 (2005).
24. Kennedy, M., Droser, M., Mayer, L.M., Pevear, D. & Mroa, D. Late
Precambrian oxygenation; inception of the clay mineral factory. Science
311, 1446–1449 (2006).
25. Taylor, W.A. & Strother, P.K. Ultrastructure of some Cambrian
palynomorphs from the Bright Angel Shale, Arizona, USA. Rev. Palaeobot.
Palyno. 151, 41–50 (2008).
26. Steemans, P. etal. Origin and radiation of the earliest vascular land plants.
Science 324, 353 (2009).
27. Tomescu, A.M. F., Pratt, L.M., Rothwell, G.W., Strother, P.K. &
Nadon,G.C. Carbon isotopes support the presence of extensive land
oras pre-dating the origin of vascular plants. Palaeogeogr. Palaeoclimatol.
Palaeoecol. 283, 46–59 (2009).
28. Rubinstein, C.V., Gerrienne, P., de la Puente, G.S. l., Astini, R.A. &
Steemans,P. Early Middle Ordovician evidence for land plants in Argentina
(eastern Gondwana). New Phytol. 188, 365–369 (2010).
29. Raven, J.A. & Andrews, M. Evolution of tree nutrition. Tree Physiol.
30, 1050–1071 (2010).
30. Kidston, R. & Lang, W.H. On Old Red Sandstone plants showing structure,
from the Rhynie Chert Bed, Aberdeenshire. Part I. Rhynia gwynne-vaughani,
Kidston and Lang. Trans. R.Soc. Edin. 51, 761–784 (1917).
31. Trewin, N.H., Fayers, S.R. & Kelman, R. Subaqueous silicication of the
contents of small ponds in an Early Devonian hot-spring complex, Rhynie,
Scotland. Can. J.Earth Sci. 40, 1697–1712 (2003).
32. Dawson, J.W. On the fossil plants from the Devonian rocks of Canada.
Q.J.Geol. Soc. Lond. 15, 477–488 (1859).
FOCUS | REVIEW ARTICLE
NATURE GEOSCIENCE DOI: 10.1038/NGEO1376
© 2012 Macmillan Publishers Limited. All rights reserved

104 NATURE GEOSCIENCE | VOL 5 | FEBRUARY 2012 | www.nature.com/naturegeoscience
33. Boyce, C.K. etal. Devonian landscape heterogeneity recorded by a giant
fungus. Geology 35, 399–402 (2007).
34. Małkowski, K. & Racki, G. A global biogeochemical perturbation across
the Silurian-Devonian boundary: Ocean-continent-biosphere feedbacks.
Palaeogeogr. Palaeoclimatol. Palaeoecol. 276, 244–254 (2009).
35. Glasspool, I.J., Edwards, D. & Axe, L. Charcoal in the Early Devonian:
Awildre-derived Konservat-Lagerstatte. Rev. Palaeobot. Palyno.
142, 131–136 (2006).
36. Gerrienne, P. etal. A simple type of wood in two Early Devonian plants.
Science 333, 837 (2011).
37. Davies, N.S. & Gibling, M.R. Paleozoic vegetation and the Siluro-Devonian
rise of uvial lateral accretion sets. Geology 38, 51–54 (2010).
38. Kennedy, K. & Gibling, M.R. e Campbellton Formation of New Brunswick,
Canada: paleoenvironments in an important Early Devonian terrestrial
locality. Can. J.Earth Sci. 48, 48, 1561–1580 (2011).
39. Stein, W.E., Mannolini, F., Hernick, L.V., Landing, E. & Berry, C.M. Giant
cladoxylopsid trees resolve the enigma of the Earth’s earliest forest stumps at
Gilboa. Nature 446, 904–907 (2007).
40. Mintz, J.S., Driese, S.G. & White, J.D. Environmental and ecological
variability of Middle Devonian (Givetian) forests in Appalachian Basin
paleosols, New York, United States. Palaios 25, 85–96 (2010).
41. Godderis, Y. & Joachimski, M.M. Global change in the Late Devonian:
modelling the Frasnian-Famennian short-term carbon isotope excursions.
Palaeogeogr. Palaeoclimatol. Palaeoecol. 202, 309–329 (2004).
42. Decombeix, A.-L., Meyer-Berthaud, B. & Galtier, J. Transitional changes in
arborescent lignophytes at the Devonian - Carboniferous boundary. J.Geol.
Soc. Lond. 168, 547–557 (2011).
43. Falcon-Lang, H.J. & Galtier, J. Anatomically-preserved tree-trunks in late
Mississippian (Serpukhovian, late Pendleian-Arnsbergian) braided uvial
channel facies, near Searston, southwest Newfoundland, Canada. Rev.
Palaeobot. Palynol. 160, 154–162 (2010).
44. Falcon-Lang, H.J. & Bashforth, A.R. Morphology, anatomy, and upland
ecology of large cordaitalean trees from the Middle Pennsylvanian of
Newfoundland. Rev. Palaeobot. Palynol. 135, 223–243 (2005).
45. DiMichele, W.A., Cecil, C.B., Montañez, I.P. & Falcon-Lang, H.J. Cyclic
changes in Pennsylvanian paleoclimate and eects on oristic dynamics in
tropical Pangaea. Int. J.Coal Geol. 83, 329–344 (2010).
46. Falcon-Lang, H.J. etal. Incised channel lls containing conifers indicate that
seasonally dry vegetation dominated Pennsylvanian tropical lowlands. Geology
37, 923–926 (2009).
47. Falcon-Lang, H.J. etal. Pennsylvanian coniferopsid forests in sabkha facies
reveal the nature of seasonal tropical biome. Geology 39, 371–374 (2011).
48. Fielding, C.R., Allen, J.P., Alexander, J. & Gibling, M.R. A facies model
foruvial systems in the seasonal tropics and subtropics. Geology
37, 623–626 (2009).
49. Hacke, U.G., Sperry, J.S., Pockman, W.T., Davis, S.D. & McCulloh, K.A.
Trends in wood density and structure are linked to prevention of xylem
implosion by negative pressure. Oecologia 126, 457–461 (2001).
50. Canadell, J. etal. Maximum rooting depth of vegetation types at the global
scale. Oecologia 108, 583–595 (1996).
51. Gastaldo, R.A. & Degges, C.W. Sedimentology and paleontology of a
Carboniferous log jam. Int. J.Coal Geol. 69, 103–118 (2007).
52. Gibling, M.R., Bashforth, A.R., Falcon-Lang, H.J., Allen, J.P. &
Fielding,C.R. Log jams and ood sediment buildup caused channel
abandonment and avulsion in the Pennsylvanian of Atlantic Canada.
J.Sedim. Res 80, 268–287 (2010).
53. Nichols, G.J. & Jones, T.P. Fusain in Carboniferous shallow marine sediments,
Donegal, Ireland: the sedimentological eects of wildre. Sedimentology
39, 487–502 (1992).
54. Marriott, S.B., Wright, V.P. & Williams, B.P. J. in Fluvial Sedimentology VII
(eds Blum, M.D., Marriott, S.B. & Leclair, S.F.) 517–529 (Blackwell, 2005).
55. Davies, N.S. & Gibling, M.R. Evolution of xed-channel alluvial
plainsinresponse to Carboniferous vegetation. Nature Geosci.
4, 629–633 (2011).
56. Gurnell, A.M. etal. Wood storage within the active zone of a large European
gravel-bed river. Geomorphology 34, 55–72 (2000).
57. Rygel, M.C., Gibling, M.R. & Calder, J.H. Vegetation-induced sedimentary
structures from fossil forests in the Pennsylvanian Joggins Formation, Nova
Scotia. Sedimentology 51, 531–552 (2004).
58. Bashforth, A.R., Drabkova, J., Oplustil, S., Gibling, M.R. & Falcon-Lang,H.J.
Landscape gradients and patchiness in riparian vegetation on a Middle
Pennsylvanian braided river plain prone to ood disturbance (Nyrany
Member, Central and Western Bohemian Basin, Czech Republic). Re v.
Palaeobot. Palynol. 163, 153–189 (2010).
59. Clarke, J.T., Warnock, R.C. M. & Donoghue, P.C. J. Establishing a time-scale
for plant evolution. New Phytol. 192, 266–301 (2011).
60. Gibling, M.R., Nanson, G.G. & Maroulis, J.C. Anastomosing river
sedimentation in the Channel Country of central Australia. Sedimentology
45, 595–619 (1998).
61. Tooth, S. & Nanson, G.C. e role of vegetation in the formation of
anabranching channels in an ephemeral river, Northern plains, arid central
Australia. Hydrol. Process. 14, 3099–3117 (2000).
62. Tooth, S., Jansen, J.D., Nanson, G.C., Coulthard, T.J. & Pietsch,T.
Riparianvegetation and the late Holocene development of an
anabranchingriver: Magela Creek, northern Australia. Geol. Soc. Am. Bull.
120, 1021–1035 (2008).
63. Harwood, K. & Brown, A.G. Changing in-channel and overbank ood
velocity distributions and the morphology of forested multiple channel
(anastomosing) systems. Earth Surf. Proc. Land. 18, 741–748 (1993).
64. Rodrigues, S., Bréhéret, J.-G., Macaire, J.-J., Greulich, S. & Villar, M.
In-channel woody vegetation controls on sedimentary processes and the
sedimentary record within alluvial environments: a modern example of an
anabranch of the River Loire, France. Sedimentology 54, 223–242 (2007).
65. Abernethy, B. & Rutherfurd, I.D. e eect of riparian tree roots on the mass-
stability of riverbanks. Earth Surf. Proc. Land. 25, 921–937 (2000).
66. Dupuy, L., Fourcaud, T. & Stokes, A. A numerical investigation into the
inuence of soil type and root architecture on tree anchorage. Plant Soil
278, 119–134 (2005).
67. Pollen, N. Temporal and spatial variability in root reinforcement of
streambanks: Accounting for soil shear strength and moisture. Catena
69, 197–205 (2007).
68. Hales, T.C., Ford, C.R., Hwang, T., Vose, J.M. & Band, L.E. Topographic and
ecologic controls on root reinforcement. J.Geophys. Res. 114, F03013 (2009).
69. Abbe, T.B. & Montgomery, D.R. Patterns and processes of wood debris
accumulation in the Queets river basin, Washington. Geomorphology
51, 81–107 (2003).
70. Webb, A.A. & Erskine, W.D. Distribution, recruitment, and geomorphic
signicance of large woody debris in an alluvial forest stream: Tonghi Creek,
southeastern Australia. Geomorphology 51, 109–126 (2003).
71. Francis, R.A., Petts, G.E. & Gurnell, A.M. Wood as a driver of landscape
change along river corridors. Earth Surf. Proc. Land. 33, 1622–1626 (2008).
72. Francis, R.A., Corenblit, D. & Edwards, P.J. Perspectives on
biogeomorphology, ecosystem engineering and self-organisation in island-
braided uvial ecosystems. Aquat. Sci. 71, 290–304 (2009).
73. Nanson, G.C., Barbetti, M. & Taylor, G. River stabilisation due to changing
climate and vegetation during the late Quaternary in western Tasmania,
Australia. Geomorphology 13, 145–158 (1995).
74. Brooks, A.P., Brierley, G.J. & Millar, R.G. e long-term control of
vegetation and woody debris on channel and ood-plain evolution: insights
from a paired catchment study in southeastern Australia. Geomorphology
51, 7–29 (2003).
75. Brown, A.G. Learning from the past: palaeohydrology and palaeoecology.
Freshwater Biol. 47, 817–829 (2002).
76. Davies, N.S. & Sambrook Smith, G. Signatures of quaternary uvial response,
Upper River Trent, Staordshire, UK: A synthesis of outcrop, documentary,
and GPR data. Z.Geomorphol. 50, 347–374 (2006).
77. Gran, K. & Paola, C. Riparian vegetation controls on braided stream
dynamics. Water Resour. Res. 37, 3275–3283 (2001).
78. Murray, A.B. & Paola, C. Modelling the eect of vegetation on channel
pattern in bedload rivers. Earth Surf. Proc. Land. 28, 131–143 (2003).
79. Coulthard, T.J. Eects of vegetation on braided stream pattern and dynamics.
Water Resour. Res. 41, W04003 (2005).
80. Tal, M. & Paola, C. Dynamic single-thread channels maintained by the
interaction of ow and vegetation. Geology 35, 347–350 (2007).
81. Braudrick, C.A., Dietrich, W.E., Leverich, G.T. & Sklar, L.S. Experimental
evidence for the conditions necessary to sustain meandering in coarse-bedded
rivers. Proc. Natl Acad. Sci. USA 106, 16936–16941 (2009).
82. Perona, P. etal. Biomass selection by oods and related timescales: Part 1.
Experimental observations. Adv. Water Resour. http://dx.doi.org/10.1016/j.
advwatres.2011.09.016 (2011).
83. Edmaier, K., Burlando, P. & Perona, P. Mechanisms of vegetation
uprootingby ow in alluvial non-cohesive sediment. Hydrol. Earth Syst. Sci.
15, 1615–1627 (2011).
84. Corenblit, D. & Steiger, J. Vegetation as a major conductor of geomorphic
changes on the Earth surface: toward evolutionary geomorphology. Earth Surf.
Proc. Land. 34, 891–896 (2009).
85. Buatois, L.A. etal. Colonization of brackish-water systems through time:
evidence from the trace-fossil record. Palaios 20, 321–347 (2005).
86. Brasier, A.T. Searching for travertines, calcretes and speleothems in deep
time: Processes, appearances, predictions and the impact of plants. Earth Sci.
Re v. 104, 213–239 (2011).
REVIEW ARTICLE | FOCUS NATURE GEOSCIENCE DOI: 10.1038/NGEO1376
© 2012 Macmillan Publishers Limited. All rights reserved

NATURE GEOSCIENCE | VOL 5 | FEBRUARY 2012 | www.nature.com/naturegeoscience 105
87. Naiman, R.J., Bilby, R.E. & Bisson, P.A. Riparian ecology and management in
the Pacic coastal rain forest. BioSciences 50, 996–1011 (2000).
88. MacNaughton, R.B. etal. First steps on land: Arthropod trackways in
Cambrian-Ordovician eolian sandstone, southeastern Ontario, Canada.
Geology 30, 391–394 (2002).
89. Davies, N.S., Gibling, M.R. & Rygel, M.C. Marine inuence in the Juniata
Formation (Upper Ordovician, Potters Mills, Pennsylvania): Implications for
the history of life on land. Palaios 25, 527–539 (2011).
90. Buatois, L.A., Mangano, M.G., Genise, J.F. & Taylor, T.N. e ichnologic
record of the continental invertebrate invasion: Evolutionary trends in
environmental expansion, ecospace utilization, and behavioral complexity.
Palaios 13, 217–240 (1998).
91. Buatois, L.A. & Mangano, M.G. in Trace Fossils Concepts, Problems, Prospects
(ed. Miller, W.I.) 285–323 (Elsevier, 2007).
92. Labandeira, C. e origin of herbivory on land: Initial patterns of plant tissue
consumption by arthropods. Insect Sci. 14, 259–275 (2007).
93. Boucot, A.J. & Janis, C. Environment of the early Paleozoic vertebrates.
Palaeogeogr. Palaeoclimatol. Palaeoecol. 41, 251–287 (1983).
94. Niedzwiedzki, G., Szrek, P., Narkiewicz, K., Narkiewicz, M. & Ahlberg, P.E.
Tetrapod trackways from the early Middle Devonian period from Poland.
Nature 463, 43–48 (2010).
95. Sues, H.-D. & Reisz, R.R. Origins and early evolution of herbivory in
tetrapods. Trends Ecol. Evol. 13, 141–145 (1998).
96. Cascales-Miñana, B. New insights into the reading of Paleozoic plant fossil
record discontinuities. Hist. Biol. 23, 115–130 (2011).
97. Cascales-Miñana, B. & Cleal, C.J. Plant fossil record and survival analyses.
Lethaia 45, 71–82 (2011).
98. Labandeira, C.C., Beall, B.S. & Hueber, F.M. Early insect diversication:
Evidence from a Lower Devonian bristletail from Quebec. Science
242, 913–916 (1988).
99. Labandeira, C.C. Invasion of the continents: cyanobacterial crusts to tree-
inhabiting arthropods. Trends Ecol. Evol. 20, 253–262 (2005).
100. Gradstein, F.M., Ogg, J.G., Smith, A.G., Bleeker, W. & Lourens, L.J. A new
geologic time scale with special reference to Precambrian and Neogene.
Episodes 27, 83–100 (2004).
Acknowledgements
We thank many colleagues for discussion and assistance, especially A. Bashforth, W.
DiMichele, R. Dott, H. Falcon-Lang, R. Gastaldo, P. Gensel, M. Rygel, W. Stein and P.
Perona. Funding was provided from a Discovery Grant to M.R.G. from the Natural
Sciences and Engineering Research Council of Canada.
Author Contributions
M.R.G. and N.S.D. jointly conceived and undertook the study and eldwork
involved.Both authors contributed to the writing of the manuscript and
gureconstruction.
Additional Information
e authors declare no competing nancial interests. Reprints and permissions
information is available online at http://www.nature.com/reprints. Correspondence
should be addressed to M.R.G.
FOCUS | REVIEW ARTICLE
NATURE GEOSCIENCE DOI: 10.1038/NGEO1376
© 2012 Macmillan Publishers Limited. All rights reserved